Practical Data Communications Roger L Freeman

Practical Data

Communications

Practical Data
Communications
Second Edition

Roger L. Freeman
New York
Chichester
Weinheim
A Wiley-Interscience Publication
JOHN WILEY & SONS, INC.
Brisbane
Singapore
Toronto
Designations used by companies to distinguish their products are often claimed as trademarks.
In all instances where John Wiley & Sons, Inc., is aware of a claim, the product names
appear in initial capital or ALL CAPITAL LETTERS. Readers, however, should contact the
appropriate companies for more complete information regarding trademarks and registration.
Copyright 1995, 2001 by Roger L. Freeman. All rights reserved.
Published by John Wiley & Sons, Inc.
No part of this publication may be reproduced, stored in a retrieval system or transmitted
in any form or by any means, electronic or mechanical, including uploading, downloading,
printing, decompiling, recording or otherwise, except as permitted under Sections 107 or
108 of the 1976 United States Copyright Act, without the prior written permission of the
Publisher. Requests to the Publisher for permission should be addressed to the
Permissions Department, John Wiley & Sons, Inc., 605 Third Avenue, New York, NY
10158-0012, 212. 850-6011, fax 212. 850-6008, E-Mail: PERMREQ & WILEY.COM.
This publication is designed to provide accurate and authoritative information in regard to the
subject matter covered. It is sold with the understanding that the publisher is not engaged in
rendering professional services. If professional advice or other expert assistance is required, the
services of a competent professional person should be sought.
ISBN 0-471-22422-7
This title is also available in print as ISBN 0-471-39302-9 cloth. and 0-471-39273-1 paper..
For more information about Wiley products, visit our web site at www.Wiley.com.
For my son, Bob
CONTENTS
Preface
Chapter 1
Chapter 2
xxix
The
1.1
1.2
1.3
1.4
1.5
Enterprise Network Environment

Enterprise Networks 1
Types of Network Topology 3
Network Access: An Overview 6
Initial Network Design Considerations 7
Connection-Oriented and Connectionless
Service 8
1.6 Data Protocols: Key to Network Operation
1.6.1 Introduction 8
1.6.2 Basic Protocol Functions 9
References 11
The OSI Model and the Data-Link Layer

2.1 Introduction 13
2.2 Layering 14
2.2.1 Notation 14
2.2.2 Basic Structuring Technique 14
2.3 Type and Instance 15
2.4 Possible Sublayers 16
2.5 Data Units 17
2.6 Specific Layers of the OSI Reference Model 18
2.7 Layer Descriptions 19
2.8 Specific Comments on OSI 33
2.8.1 General 33
2.9 Discussion of OSI Layers 14 34
2.9.1 The Physical Layer Layer 1. 34
2.9.2 The Data-Link Layer Layer 2. 35
2.9.3 The Network Layer Layer 3. 37
13
vii
viii
CONTENTS
2.9.4
Internet Protocol IP. and Gateway Layer

Layer 3.5. 37
2.9.5 The Transport Layer Layer 4. 38
2.10 Procedural Versus Electrical 39
2.10.1 Narrative 39
References 39
Chapter 3
Chapter 4
High-Level Data-Link Control (HDLC) Typical

Data-Link Layer Protocol
3.1 Introduction 41
3.2 Stations and Configurations 42
3.3 Modes of Operation Used with HDLC 43
3.4 HDLC Frame Structure 43
3.4.1 Sequence Numbering in HDLC 45
3.4.2 The PollrFinal Bit 47
3.4.3 Supervisory Frames 47
3.4.4 Unnumbered U. Frames 48
3.4.5 Information Field 48
3.4.6 Frame Check Sequence FCS. Field 48
3.5 Commands and Responses 49
3.5.1 Mode-Setting Commands 49
3.5.2 Miscellaneous Commands 51
3.5.3 Responses to Unnumbered Commands 51
3.6 Frame Operation 52
3.7 Error Recovery 53
3.8 Other Station Modes 54
3.9 SDLC Variations with HDLC 55
References 55
Data Network Operations
4.1 Chapter Objective 57
4.2 General Requirements for the Interchange of
Data 57
4.3 Discussion of Issues and Requirements 58
4.3.1 Pathway s. Exists. 58
4.3.2 Access to the Pathway 59
4.3.3 Directing the Data Messages to Its
Intended Users. 59
4.3.4 Data Urgency 59
4.3.5 Maintenance of Data Integrity 60
4.4 Error Detection Schemes 61
4.4.1 Parity Checks 61
4.4.2 Cyclic Redundancy Check CRC. 62
41
57
CONTENTS
ix
4.5
Error Correction Schemes 64

4.5.1 Automatic Repeat Request ARQ. 64
4.5.2 Forward Error Correction FEC. 65
4.6 Data Switching 67
4.6.1 Philosophy of Data Signaling Versus
Telephone Signaling 69
4.6.2 Smart Bridges, Routers, and Switching
Hubs 71
References 71
Chapter 5
Data Transmission I
5.1 Electrical Communication of Information 73
5.2 The Bit and Binary Transmission of
Information 73
5.3 Binary Codes for Data Communication 76
5.4 Electrical Bit Decisions 79
5.5 Electrical Representation of Binary Data 82
5.5.1 Neutral and Polar Waveforms 82
5.5.2 Waveforms and Line Codes 83
5.6 Binary Conventions 84
5.7 Bit-Parallel and Bit-Serial 85
5.8 Baseband 86
5.9 Data Rate 87
5.9.1 Error Rate 87
References 88
73
Chapter 6
Data Transmission II
6.1 Interpreting a Serial Stream of Bits 89
6.1.1 The Problem 89
6.1.2 StartStop Transmission 89
6.1.3 Synchronous Transmission 91
6.2 Timing Distortion in a Serial Bit Stream 93
6.3 The Transmission of Digital Data 94
6.3.1 Baseband Transmission 94
6.3.2 Transmission of Data over Longer
Distances 95
6.3.2.1 The Voice Channel 95
6.3.2.2 Modems 95
6.3.2.3 Voice Channel Impairments to
Data Transmission 97
6.3.3 Data Circuit Performance 102
6.3.3.1 Introduction to Error
Performance 102
89
CONTENTS
6.3.3.2
Error Performance from an ITU-T

Perspective 103
6.3.3.3 Error Performance from a
Bellcore Perspective 107
6.4 Interface at the Physical Layer 107
6.4.2 The DTEDCE Interface 108
6.4.2.1 EIArTIA-232 Electrical Interface
110
6.4.2.2 Functional Description of Selected
Interchange Circuits 112
6.4.2.3 EIArTIA Interface Standards:
EIArTIA-530, EIArTIA-422, and
EIArTIA-423 114
6.5 The Question of Bandwidth 116
6.5.1 Bandwidth Versus Bit Rate 116
6.5.2 Bandwidth Defined 116
6.5.3 First and Second Bandwidth
Approximations 116
6.5.4 Shannon 117
6.5.5 Summary 118
References 118
Chapter 7
The Telecommunications Network as a Vehicle for

Data Transport
7.1 The Public Switched Telecommunication
Network 121
7.1.2 Access to the PSTN: The Subscriber
Network 122
7.2 Introduction to Digital Networks 123
7.2.1 Rationale 123
7.2.2 Development of a PCM Signal 123
7.2.3 The Concept of Frame 131
7.2.4 The Line Code 134
7.2.5 Regenerative Repeaters 135
7.2.6 Higher-Order PCM Multiplex
Systems 136
7.2.7 Line Rates and Codes 139
7.3 Brief Overview of Digital Switching 139
7.3.1 Advantages and Issues of PCM
Switching 139
7.3.2 Approaches to PCM Switching 140
121
CONTENTS
xi
7.3.3 Time Switch 140

7.3.4 Space Switch 142
7.3.5 TimeSpaceTime Switch 144
7.3.6 SpaceTimeSpace Switch 146
7.4 Digital Network Structure 146
7.5 Digital Network Impairments and Performance
Requirements 147
7.5.1 Error Performance 147
7.5.2 Slips 147
7.5.2.1 Definition 147
7.5.2.2 North American Slip
Objectives 148
7.5.2.3 CCITT Slip Objectives 148
7.6 Data Transmission on the Digital Network 149
7.6.1 The Problem 149
7.6.2 Some Solutions 149
7.6.2.1 The North American Digital Data
System DDS. 150
7.6.2.2 Bit Integrity in a Time Slot 150
7.7 Interconnects and Bypass 151
7.8 Bypass in Economically Evolving Nations 152
References 152
Chapter 8
The Transmission of Data over the Analog

Voice Channel
8.1 Background 155
8.2 Two-Wire Versus Four-Wire Operation 156
8.2.1 What Is Two-Wire and Four-Wire
Operation? 156
8.2.2 Two-Wire Transmission 156
8.2.3 Four-Wire Transmission 156
8.3 Echo and Singing: Telecommunication Network
Impairments 159
8.4 Amplitude Distortion and Phase Distortion 159
8.4.2 Amplitude Distortion 160
8.4.3 Phase Distortion 160
8.4.4 Conditioning and Equalization 161
8.5 Data Modems 165
8.5.1 Where We Are and Where We Are Going
165
8.5.2 Getting More Bits per Hertz 165
8.5.3 Specific High-Speed Modems 168
155
xii
CONTENTS
8.5.4
8.5.5
Scrambling and Its Rationale 174

Introduction to Trellis-Coded Modulation
TCM. 176
8.5.6 The V.34 Modem: 28,8000 bps and 33,600
bps 178
8.5.6.1 General 178
8.5.6.2 Selected Definitions 179
8.5.6.3 An Overview of Selected Key
Areas of Modem Operation 179
8.5.6.4 Framing 183
8.5.6.5 Encoder 186
8.5.7 The V.90 ModemMaximum Data Rate:
56 kbps 187
8.5.7.1 Principal Characteristics 187
8.5.7.2 Definitions 188
8.5.7.3 Overview of V.90 Operation 191
8.5.7.4 V.90 Encoder 191
8.5.7.5 Mapping Parameters 192
8.5.7.6 Input Bit Parsing 192
8.5.7.7 Modulus Encoder 192
8.5.7.8 Mapper 194
8.5.7.9 Spectral Shaping 194
8.5.7.10 Spectral Shaper 196
8.5.7.11 Spectral Shape Filter 198
8.5.7.12 Sign Assignment 198
8.5.7.13 MUX 198
8.5.7.14 Analog Modem 198
References 200
Chapter 9
Data Communications in the Office Environment,

Part 1
9.1 Introduction 203
9.2 Distinguishing Characteristics of LANs 204
9.3 How LAN Protocols Relate to OSI 206
9.4 Logical Link Control LLC. 206
9.4.1 LLC Sublayer Services and Primitives 208
9.4.1.1 Unacknowledged ConnectionlessMode Service 209
9.4.1.2 Connection-Mode Service 209
9.4.1.3 Type 1 and Type 2 Operation 210
9.4.2 LLC PDU Structure 211
9.4.2.1 Overview 211
9.4.2.2 Control Field Formats 213
203
CONTENTS
9.5
9.6
9.4.2.3 Control Field Parameters 214

9.4.1.4 Commands and Responses 216
Medium Access Control MAC. 223
9.5.2 Carrier Sense Multiple Access with
Collision Detection CSMArCD. 223
9.5.2.1 Introduction to CSMArCD 223
9.5.2.2 MAC Frame Structure for
CSMArCD 223
9.5.2.3 CSMArCD Basic Operation 229
CSMArCDCurrent Status and Advanced
Operation 232
9.6.1 General 232
9.6.2 Half-Duplex and Full-Duplex 232
9.6.3 Compatability Interfaces Physical Layer
232
9.6.4 Layer Interface Requirements 236
9.6.4.1 Functional Capabilities of the
MAC Sublayer 236
9.6.5 Allowable
ImplementationsParameterized Values
238
9.6.5.1 Definition of Slot Time 239
9.6.5.2 CSMArCD System
Designation 239
9.6.6 Physical Signaling PLS. and Attachment
Unit Interface AUI. Selected
Discussion 240
9.6.6.1 Frame Structure Across the
AUI 240
9.6.6.2 Signal Characteristics 241
9.6.6.3 Medium Attachment Unit MAU.
10BASE5 244
9.6.6.4 Broadband CSMArCD,
10BROAD36Overview 244
9.6.7 System Configurations for Multisegment
10-Mbps Baseband Networks 249
9.6.7.2 Shrinkage of the Interpacket Gap
IPG. 251
9.6.7.3 Transmission System Model 1
251
xiii
xiv
CONTENTS
9.6.7.4
9.7
9.8
Typical Topologies Applicable to

Transmission System Model 1
252
100-Mbps CSMArCD Baseband Networks 252
9.7.1 OverviewKey Points 252
9.7.1.1 Reconciliation Sublayer RS. and
Media Independent Interface
MII. Introduction 254
9.7.1.2 Physical Layer Signaling
Systems 255
9.7.1.3 Repeater 256
9.7.2 Reconciliation Sublayer RS. and MediaIndependent Interface MII. Description
Details 256
9.7.2.1 MII Frame Structure 260
9.7.3 100BASE-T4, Its PMA, and PCS 261
9.7.3.1 Introduction 261
9.7.3.2 Objectives 261
9.7.3.3 Description 262
9.7.3.4 Summary of Physical Medium
Attachment PMA. Specification
264
9.7.4 Physical Coding Sublayer PCS. and
Physical Medium Attachment PMA.
Sublayer, Type 100BASE-X 264
9.7.4.1 Objectives 264
9.7.4.2 Functional Requirement 264
9.7.5 System Considerations for Multisegment
100BASE-T Networks 271
9.7.5.1 Physical Size of the Network 271
9.7.5.2 Transmission Models 272
9.7.5.3 Round-Trip Collision Delay
Key to Network Operation and
Maximum Size 273
1000-Mbps CSMArCD Networks 274
9.8.1 Overview 274
9.8.2 Reconciliation Sublayer RS. and Gigabit
Media-Independent Interface GMII. 276
9.8.2.2 GMII Data Stream 276
9.8.3 Physical Coding Sublayer PCS. and
Physical Medium Attachment PMA.
Sublayer, 1000BASE-X 277
CONTENTS
xv
9.8.3.1
9.8.3.2
9.8.3.3
Overview 277
Objectives 278
Summary of 1000BASE-X
Sublayers 278
9.8.3.4 Physical Coding Sublayers PCS.
279
1000-Mbps Networks 283
9.8.4.2 Single-Collision-Domain
Multisegment Networks 284
9.8.4.3 Transmission System
Model 1 285
9.8.4.4 Transmission System
Model 2 285
References 288
Chapter 10
Data Communications in the Office Environment,

Part 2
10.1 Medium Access ControlToken-Passing
Schemes 289
10.1.2 Token-Passing Bus 290
10.1.2.1 Token-Bus Frame
Formats 291
10.1.2.2 Operational Aspects 292
10.1.3 Token-Passing Ring 293
10.1.3.1 Overview 293
10.1.3.2 Token-Ring Frame
Formats 295
10.1.3.3 Token-Ring Symbol
Transmission Characteristics
301
10.1.3.4 Symbol Decoder 303
10.1.3.5 Latency 303
10.1.3.6 Station Organization and
Basic PMC Requirements
305
10.1.4 Fiber Distributed Data Interface
FDDI. 308
10.1.4.1 Introduction and Overview
308
10.1.4.2 Ring Transmission 310
289
xvi
CONTENTS
10.1.4.3
10.1.4.4
Protocol Data Units 312

FDDI Timers and Timing
Budget 314
10.1.4.5 FDDI Ring Operation 314
10.1.5 Wireless LANs WLANs. 316
10.1.5.3 Frame Formats 318
10.1.5.4 Overview of the WEP 323
10.1.5.5 MAC Access 323
10.1.5.6 Overview of the Air
Interface 332
10.2 Repeaters, Bridges, Routers, and Hubs 335
10.2.1 Definitions and Interactions 335
10.3 LAN BridgesOverview 335
10.3.1 The IEEE 802.1 Bridge 338
10.3.1.1 Applications of the 802.1
Bridge 338
10.3.1.2 Principles of Operation 338
10.3.1.3 The Spanning Tree Algorithm
344
10.3.1.4 Bridge Protocol Data Unit
BPDU. 350
10.3.2 Source Routing Bridges 352
10.3.3 Remote Bridges 354
10.4 Hubs and Switching Hubs 355
10.4.1 Rationale and Function 355
10.4.2 Improvements in Hub Technology 355
10.5 Routers 356
10.5.1 Addressing with Hierarchical
Significance 356
10.5.2 Subnet Masks 358
10.6 Virtual Local Area Networks VLANs. 358
10.7 Servers and Intranets 359
10.7.1 Servers 359
10.7.2 Intranet and Extranet 360
References 360
Chapter 11
Wide Area Networks (WANs)

11.1 Background and Scope 363
11.2 Basic Approaches 364
11.2.1 Point-to-Point Links 364
363
CONTENTS
11.2.1.1
Economy Versus Data

Perishability 364
11.2.1.2 Dedicated Circuits 365
11.2.2. Data Multiplexers and Statmultiplexers
366
11.2.2.1 Time Division Multiplexing
366
11.2.2.2 Statistical Time Division
Multiplexing 366
11.3 Packet Networks for Data Communication 367
11.3.1 Introduction to Packet Data 367
11.3.2 Packet-Switched and Virtual
Connections Based on ITU-T Rec. X.25
369
11.3.2.2 X.25 Packets Are Contained
in LAPB Frames 370
11.3.2.3 Single-Link and Multilink
Procedures 372
11.3.2.4 Packet-Level DTErDCE
Interface 373
11.3.2.5 A Typical X.25 Packet 377
11.3.2.6 Typical X.25 Transmission
Exchange of Packets 381
11.3.2.7 Addressing in X.25 382
11.3.2.8 Assignment of Logical
Channels 385
11.4 Transmission Control ProtocolrInternet
Protocol TCPrIP. 386
11.4.1 Background and Application 386
11.4.2 TCPrIP and Data-Link Layers 387
11.4.3 The IP Routing Function 390
11.4.4 Detailed IP Operation 391
11.4.4.1 Frame Format of an IP
Datagram; Description of
Functions 391
11.4.4.2 IP Routing 397
11.4.4.3 Internet Control Message
Protocol ICMP. 399
11.4.5 The Transmission Control Protocol
402
11.4.5.1 TCP Defined 402
11.4.5.2 TCP Mechanisms 403
xvii
xviii
CONTENTS
11.4.5.3
11.4.5.4
11.4.5.5
11.4.5.6
11.4.5.7
ULP Synchronization 404

ULP Modes 404
Scenario 405
TCP Header Format 408
TCP Entity State Diagram
410
11.4.6 Internet Protocol Version 6 IPv6. 410
11.4.6.1 IPv6 Routing 410
11.4.6.2 IPv6 Header 411
11.5 User Datagram Protocol UDP. 414
11.5.1 UDP Header Format and Fields 414
11.6 The CLNP Protocol Based on ISO 8473 415
11.6.1 CLNP Frame Formats 416
11.6.2 Frame Field DescriptionFixed Part
416
11.6.3 Error Report PDU ER. 423
11.6.3.1 Reason for Discard 423
11.7 Networking via VSATs 424
11.7.1 Rationale of VSAT Networks 424
11.7.2 Basic Description of a VSAT Network
425
11.7.3 Disadvantages of a VSAT Network
425
11.7.4 Important Design Variables 427
11.7.4.2 Optimizing Bandwidth Usage
428
11.8 Hypothetical Reference Connections for Public
Synchronous Data Networks 430
References 431
Chapter 12
Frame Relay
12.1 How Can Networks Be Speeded Up? 433
12.1.1 The Genesis of Frame Relay 435
12.2 Introduction to Frame Relay 436
12.2.1 The Frame Structure 437
12.2.2 Address Field Discussion 441
12.2.2.1 Address Field Variables 441
12.3 DL-CORE Parameters As Defined by ANSI.
444
12.3.1 Procedures 444
12.4 Traffic and Billing on Frame Relay 445
433
CONTENTS
12.5 PVCs and SVCs 446

12.6 Two Types of Interfaces: UNI and NNI 446
12.7 Congestion Control: A Discussion 447
12.7.1 Network Response to Congestion 448
12.7.2 User Response to Congestion 448
12.7.3 Consolidated Link Layer Management
CLLM. Messages 449
12.7.4 Action of a Congested Node 452
12.8 Flow Control and Possible Applications of
FECN and BECN Bits 453
12.8.1 FECN Usage 453
12.8.2 BECN Usage 454
12.9 Policing a Frame Relay Network 455
12.9.2 Definitions 455
12.9.3 Relationship Among Parameters 456
12.10 Quality of Service Parameters 458
12.11 Network Responsibilities 459
12.12 Frame Relay Signalling Functions 460
12.12.1 PVC Management Procedures 461
12.12.1.1 Message Definition 461
12.12.1.2 Description of Operation:
Health and Welfare, PVC
Link 463
12.12.2 Signaling Required for SVCs 468
12.13 Compatibility Issues 474
References 480
Chapter 13 Integrated Services Digital Networks (ISDNs) 483
13.1 Introduction 483
13.2 ISDN Structures 483
13.2.1 ISDN User Channels 483
13.2.2 Basic and Primary Rate Interfaces 484
13.3 User Access and Interface 485
13.3.1 General 485
13.4 ISDN Protocols and Protocol Issues 488
13.5 ISDN Networks 490
13.6 ISDN Protocol Structures 493
13.6.1 ISDN and OSI 493
13.6.2 Layer 1 Interface, Basic Rate
CCITT. 494
13.6.2.1 Primitives Between Layer 1
and Other Entities 495
xix
xx
CONTENTS
13.6.2.2
13.6.2.3
13.6.2.4
13.6.2.5
13.6.2.6
Interface Functions 495

Frame Structure 496
Line Code 498
Timing Considerations 498
BRI Differences in the
United States 498
13.6.3 Layer 1 Interface, Primary Rate 500
13.6.3.1 Interface at 1.544 Mbps 501
13.6.3.2 Interface at 2.048 Mbps 504
13.7 Overview of Layer 2 Interface: Link Access
Procedure for the D Channel LAPD. 505
13.7.1 Layer 2 Frame Structure for Peer-toPeer Communication 509
13.8 Overview of Layer 3 514
13.8.1 Layer 3 Specification 516
13.8.1.1 General Message Format and
Information Elements Coding
518
13.9 ISDN Packet Mode Review 520
13.9.2 Case A: Configuration When Accessing
PSPDN Services 520
13.9.3 Case B: Configuration for the ISDN
Virtual Circuit Service 521
13.9.4 Service Aspects 523
13.9.4.1 Access to PSPDN
ServicesCase A 523
13.9.4.2 Access to the ISDN Virtual
Circuit ServiceCase B 523
References 524
Chapter 14
Building and Campus Wiring and Cabling for

Data Communications
14.1 Background and Objective 527
14.2 Major Elements of Telecommunication Building
Layout 528
14.3 Horizontal Cabling 528
14.3.1 Selection of Media 529
14.4 Backbone Cabling 530
14.4.1 Topology of Backbone Cabling 530
14.4.2 Selecting Cable Media 530
14.5 Telecommunication Closets 532
14.5.1 Cabling Practices 532
527
CONTENTS
14.6 Entrance Facilities 533

14.7 100- Unshielded Twisted-Pair UTP. Cabling
Systems 533
14.7.1 UTP Category Definitions 533
14.7.2 Horizontal UTP Cable 534
14.7.2.1 Specific Transmission
Characteristics 534
14.8 Backbone UTP Cabling 535
14.8.1 Transmission Performance
Requirements 535
14.8.2 Connecting Hardware for UTP
Cable 536
14.9 150- Shielded Twisted-Pair Cabling
Systems 538
14.9.1 Transmission Performance
Requirements 538
14.10 Optical Fiber Cabling Systems 539
14.10.1 Backbone Optical Fiber Cable 540
14.11 Grounding and Bonding Infrastructure for
Commercial Buildings 541
14.11.1 Rationale 541
14.11.2 The Telecommunications Bonding
Backbone TBB. 542
14.11.3 The Telecommunications Main
Grounding Busbar TMGB. 543
14.11.4 The Telecommunications Grounding
Busbar TGB. 544
14.11.5 Bonding to the Metal Frame of a
Building 544
14.11.6 Telecommunications Entrance Facility
TEF. 545
14.11.6.1 Placement of the
TMGB 545
14.11.7 Telecommunication Closets and
Equipment Rooms 546
14.12 Customer-Owned Outside Plant OSP.
Infrastructure 546
14.12.1 OSP Cabling Infrastructure Defined
546
14.12.2 OSP Topology 546
14.12.3 Recognized Cable Media and Media
Selection 548
References 549
xxi
xxii
CONTENTS
Chapter 15
Broadband Data Transport Techniques

15.1 Chapter Objective 551
15.2 Introduction to Fiber-Optics Transmission 552
15.2.1 What Is So Good About Fiber-Optic
Transmission? 552
15.2.2 Advantages of Optical Fiber 553
15.2.3 Overview of an Optical Fiber Link 554
15.2.4 Optical Fiber Transmission 555
15.2.4.1 Optical Fiber Composition
555
15.2.4.2 Distortion and Dispersion
558
15.2.4.3 Attenuation 561
15.2.5 Types of Optical Fiber 562
15.2.5.1 Multimode Fibers 562
15.2.5.2 Single-Mode Fiber 562
15.2.6 Fiber-Optic Cable 563
15.2.6.1 Cable Composition 563
15.2.7 Fiber-Optic Transmitters 564
15.2.7.1 Light-Emitting Diodes 565
15.2.7.2 Laser Diodes 566
15.2.8 Receivers 566
15.2.8.1 PIN Photodiodes 566
15.2.8.2 Avalanche Photodiodes 567
15.2.8.3 Noise 567
15.2.9 Repeaters 568
15.2.10 Fiber-Optic Amplifiers 568
15.2.11 The Joining of Fibers: Splices and
Connectors 569
15.2.11.1 Splices 569
15.2.11.2 Connectors 569
15.2.12 Modulation and Coding 569
15.2.12.1 Digital Signals 570
15.2.12.2 Optical Line Coding and
Transmission Rate 571
15.3 Higher-Order Multiplexing Formats: SONET
and SDH 571
15.3.1 Synchronous Optical Network
SONET. 571
15.3.1.2 SONET Rates and Formats
571
551
CONTENTS
xxiii
15.3.1.3
15.3.1.4
15.3.1.5
STS-1 Frame Structure 572

STS-N Frame Structure 576
The Structure of Virtual
Tributary 577
15.3.1.6 SONET Layers and Transport
Functions 581
15.3.2 Synchronous Digital Hierarchy 583
15.3.2.1 SDH Standard Bit Rates
584
15.3.2.2 Interface and Frame
Structure 584
15.3.2.3 Frame Structure 588
15.3.2.4 Interconnection of STM-1s
590
15.3.2.5 Basic SDH Multiplexing
Structure 591
15.4 Line-of-Sight Microwave and Millimeter-Wave
Transmission Systems 591
15.4.1 Broadband Radio Systems 591
15.4.2 An Overview of Line-of-Sight
Microwave 592
15.4.2.1 Line of Sight 592
15.4.2.2 Typical Line-of-Sight
Microwave Installation 593
15.4.3 Satellite Communications 596
15.4.3.1 Concept and Applications
596
15.4.3.2 Uplinks and Downlinks 597
15.4.3.3 Satellite Orbits 597
15.4.3.4 Digital Satellite System
Operation and Access 598
References 599
Chapter 16
Broadband ISDN (B-ISDN) and the Asynchronous

Transfer Mode (ATM)
16.1 Where Are We Going? 601
16.2 Introduction to ATM 602
16.3 UserNetwork Interface UNI., Configuration,
and Architecture 605
16.4 The ATM Cell: Key to Operation 606
16.4.1 ATM Cell Structure 606
16.4.2 Header Error Control Sequence
Generation 611
601
xxiv
CONTENTS
16.4.3 Idle Cells 613

16.5 Cell Delineation and Scrambling 613
16.5.1 Delineation and Scrambling Objectives
613
16.5.2 Cell Delineation Algorithm 614
16.6 ATM Layering and B-ISDN 615
16.6.1 Functions of Individual ATMrB-ISDN
Layers 616
16.6.1.1 Physical Layer 616
16.6.1.2 The ATM Layer 618
16.6.2 The ATM Adaptation Layer AAL.
620
16.6.2.1 Sublayering of the AAL 620
16.6.2.2 Service Classification of the
AAL 620
16.7 Services: Connection-Oriented and
Connectionless 633
16.7.1 Functional Architecture 633
16.7.2 CLNAP Protocol Data Unit PDU.
Structure and Encoding 635
16.8 Aspects of a B-ISDNrATM Network 637
16.8.1 ATM Routing and Switching 637
16.8.1.1 Virtual Channel Level 637
16.8.1.2 Virtual Path Level 638
16.9 Signaling Requirements 640
16.9.1 Setup and Release of VCCs 640
16.9.2 Signaling Virtual Channels 640
16.9.2.1 Requirements for Signaling
Virtual Channels 640
16.9.3 Metasignaling 641
16.9.3.1 Metasignaling Functions at
the User Access 641
16.9.3.2 Relationship Between
Metasignaling and User
Access Signaling
Configuration 641
16.9.3.3 Metasignaling Functions in
the Network 642
16.3.3.4 Metasignaling Requirements
642
16.9.4 Practical Signaling Considerations 643
16.9.4.1 ATM Addressing 644
CONTENTS
16.9.4.2
Circuit Setup Using ATM

Signaling 645
16.10 Quality of Service QoS. 647
16.10.1 ATM Service Quality Review 647
16.10.2 Definitions 647
16.10.3 Cell Transfer Delay 648
16.10.4 Cell Delay Variation 648
16.10.5 Cell Loss Ratio 649
16.10.6 Mean Cell Transfer Delay 649
16.10.7 Cell Error Ratio 649
16.10.8 Severely Errored Cell Block Ratio
650
16.10.9 Cell Misinsertion Rate 650
16.11 Traffic Control and Congestion Control 650
16.11.1 Generic Functions 650
16.11.2 Events, Actions, Time Scales, and
Response Times 651
16.11.3 Quality of Service, Network
Performance, and Cell Loss Priority
652
16.11.4 Traffic Descriptors and Parameters
653
16.11.5 UserNetwork Traffic Contract 654
16.11.5.1 Operable Conditions
654
16.11.5.2 Source Traffic Descriptor,
Quality of Service, and
Cell Loss Priority 652
16.11.5.3 Impact of Cell Variation
on UPCrNPC and
Resource Allocation 654
16.11.5.4 Cell Conformance and
Connection Compliance
656
16.11.5.5 Generic Cell Rate
Algorithm GCRA. 657
16.11.5.6 Traffic Contract
Parameter Specification
659
16.12 Transporting ATM Cells 659
16.12.1 In the DS3 Frame 659
16.12.2 DS1 Mapping 660
16.12.3 E1 Mapping 662
xxv
xxvi
CONTENTS
16.12.4
Mapping ATM Cells into SDH 665

16.12.4.1 At STM-1 155.520 Mbps.
665
16.12.4.2 At STM-4 622.080 Mbps.
665
16.12.5 Mapping ATM Cells into SONET
666
16.12.6 Cell Rates for Various Typical
Connections 666
16.13 ATM Operation, Administration, and
Maintenance OAM. 667
16.13.1 OAM Levels in the B-ISDN 667
16.13.2 Mechanisms to Provide OAM Flows
668
16.13.2.1 Physical Layer
Mechanisms 668
16.13.3 ATM Layer Mechanism: F4 and F5
Flows 669
16.13.3.1 F4 Flow Mechanism 669
16.13.3.2 F5 Flow Mechanism 670
16.13.4 OAM Functions of the Physical Layer
671
16.13.4.1 OAM Functions 671
16.13.5 OAM Functions of the ATM Layer
671
16.13.5.1 VP-AIS and VP-FERF
Alarms 671
16.13.5.2 OAM Functions for the
VCC F5 Flows. 677
16.13.6 ATM Layer OAM Cell Format 677
16.13.6.1 Common OAM Cell
Fields 677
References 678
Chapter 17
Last-Mile Data Distribution Systems

17.1 New Approaches to Outside Plant Distribution
681
17.2 Introducing Last-Mile Distribution Systems
682
17.3 Introduction to CATV 687
17.3.1
The Essentials of TV Signal
Distribution on a CATV System 688
681
CONTENTS
xxvii
17.3.2
Extending CATV Coverage Area

690
17.3.3
The DOCSIS Specification 692
17.3.3.1
RF Channel Assumptions
694
17.3.3.2
Communication Protocols
696
17.3.3.3
Logical Link Control 700
17.3.3.4
Link-Layer Security
Sublayer 700
17.3.3.5
MAC Sublayer 700
17.3.3.6
Downstream Transmission
Convergence Sublayer
701
17.3.3.7
PMD Sublayer Overview
701
17.3.3.8
Interface Points 702
17.3.3.9
Physical-MediaDependent Sublayer
Specification 702
17.3.3.10 Modulation Formats and
Rates 704
17.3.3.11 Symbol Mapping 704
17.3.3.12 Burst Profiles 709
17.4 Local Multipoint Distribution System LMDS.
716
17.4.1
Frequency Reuse 718
17.5 Other Methods to Breach the Last Mile 724
17.5.1
Remote Wireless Radio. Bridges
724
17.5.2
In-Building Wireless Telephone
Systems 725
References 725
Chapter 18
Network Management for Enterprise Networks

18.1 What is Network Management? 727
18.2 The Bigger Picture 727
18.3 Traditional Breakout by Tasks 728
18.3.1
Fault Management 728
18.3.2
Configuration Management 728
18.3.3
Performance Management 729
18.3.4
Security Management 729
18.3.5
Accounting Management 729
727
xxviii
CONTENTS
18.4
SurvivabilityWhere Network Management

Really Pays 729
18.5 Availability Enhancement Rapid
Troubleshooting 730
18.5.1
Troubleshooting 731
18.6 System DepthA Network Management
Problem 734
18.6.1
Aids in Network Management
Provisioning 735
18.6.2
Communication Channels for the
Network Management System 739
18.7 An Introduction to Network Management
Protocols 739
18.7.1
Two Network Management
Protocols 739
18.7.2
An Overview of SNMP 740
18.7.3
SNMP Version 2 746
18.7.4
Common Management Information
Protocol CMIP. 747
18.8 Telecommunications Management Network
TMN. 749
18.8.1
Objective and Scope 749
18.8.2
Network Management Functions
Carried Out in Upper OSI Layers
751
References 755
Appendix I
Appendix II
Index
Addressing Conventions
I.1
Transforming Decimal Numbers to Binary
Numbers and Transforming Binary Numbers
to Decimal Numbering 757
I.2
Decimal Digit Representation in IP Addresses
759
I.3
Binary, Octal, and Hexadecimal Numbers
761
Acronyms and Abbreviations
757
763
779
PREFACE TO
THE SECOND EDITION
This revised second edition of Practical Data Communications has been
prepared with the IT technologist in mind, rather than the electrical engineer. I did this simply to reach a larger readership. The revision required a
completely different perspective on my part. I found this experience refreshing.
But which came firstthe chicken or the egg? In this case, where should
the emphasis be placed? On the electrical data signal itself? Or the software
that makes a data circuit work? I opted for the latter in this edition.
Still and yet, I have not lost sight of my principal purposeto share with
others my long learning experience in the field of data communications. It
started in 1955 with datas predecessor, the automatic telegraph aboard a
ship where I served as Radio Officer, using the old Teletype Corporation
Model 15 clacking along at 60 wordsrminute wpm.. When ashore, I marveled at the Navys 100 wpm circuits which operated at 75 bps. When I was
station engineer at the USAF 2049 th Communications Group, one of the first
LogComNet circuits was established between McClellan AFB and Norton
AFB. It was truly high speed at 150 bps. The circuit was asynchronous, better
called a startstop operation. That system became DoDs AutoDin and was
indeed a data circuit, but based on telegraph technology.
My learning was based on setting up real circuits; the experience was
hands-on. Thus my goal with this text is to facilitate an understanding of data
communications that emphasizes practical application rather than theory. It
is meant to serve as a basic reference for practitioners at all levels of this vital
industry. It is tutorial as well, and will bring along the neophyte.
I have gone to great lengths to use careful language. The term bandwidth
is one of which I am particularly conscious. It is measured in Hz, not in bits
per second, which seems to be the prevailing belief. I will admit I have been
forced to slip on the misuse of words when I quote somebody.
This text deals with generic systems. I have kept clear of any description of
proprietary techniques. Ethernet and CSMArCD are used synonymously. If
some find fault with equating the twoso be it. When I prepared the first
xxix
xxx
PREFACE TO THE SECOND EDITION
edition, there were three subsets of IEEE 802.3; today there are 22 quoted in
the IEEE green tome, some 1200 pages thick. I have devoted one complete
chapter to this subject because it pervades the marketplace. There are two
reasons for this: cost and simplicity.
Some people refer to telecommunications as a technology dealing exclusively with voice communications. Data communication is a separate entity
with its own rules and laws of nature; never the twain shall meet. From my
perspective, they both fall under electrical communication, which is synonymous with telecommunication. In fact, later chapters of the text show the
melding of voice, data and video into the larger field of telecommunication.
For example, the underlying objective of the Integrated Services Digital
Networks ISDN. is the integration of all media into a single telecommunication system. ATM is based on the same premise, and in my opinion, truly
accomplishes the goal. I did prepare one section on the philosophy of
signaling showing where, in that one arena, data and traditional voice
telecommunications differ.
I also have broken with tradition and included chapters on the description
of the present underlying digital network, including SONET and SDH. Other
texts on data do not do this. I firmly believe that to design a complete data
network, more often than not, we have to take advantage of services available
from the PSTN. This is common practice with WANs. In Chapter 17,
Last-Mile Data Distribution Systems, a customer can access any one or
several of the media at hisrher premises site, which are voice, data, and
video. On the entry side, these circuits somewhere, somehow, will be transported on the PSTN. Knowledge of how the PSTN can either hurt or help, is
vital for a complete understanding of data networks.
The first edition chapter sequence started with rudimentary electrical
signals and later on introduced protocols. This edition does the reverse,
starting with a description of a data-link layer protocol after introducing the
layering concept and OSI. There are three chapters dealing with data
transmission, which serve as a tutorial basis for the remainder of the book.
LANs and WANs are treated in an area which I choose to call Enterprise
Networks. I emphasize data networks and their protocols eliminating discussion of specialized networks, such as DQDB, SMDS, and HIPPI. IPv4 is
stressed, but I do include a short overview of IPv6. An entire chapter is
devoted to ATM, highlighting data application.
I am of the opinion that three generic system types, CATV, DSL and
LMDS, will vie for last-mile distribution. One or two of these systems will
continue the battle for leadership for broadband delivery during the lifetime
of this edition. Internet connectivity is the driving force. A brief description
of each system can be found in Chapter 17.
During the preparation phase of the second edition, the Internet was
discovered to be an excellent source of information. This held particularly
true for IP. There is a data bank available for all RFCs including the latest.
xxxi
ACKNOWLEDGMENTS
Dr. Ernie Woodward gave me a new perspective on how to present practical
data communications to a general readership. Data communications may be
broken down into two basic elements: procedural and electrical. I was
brought up more on the electrical engineering side of the fence. Ernie
showed me that there was another way of doing it by following a procedural
approach from the beginning, namely, the application of data protocols. The
objective was to make the book more attractive to IT personnel. Ernie was
previously a section manager at Motorola Chandler and now is with INTEL
here in the Valley. While at Motorola, Ernie was my boss for some eight
months. In that period, I prepared the Frame Relay Interface Control
Document for the Celestri terminal segment. He was a great boss and is a
fine human being.
Many other associates offered positive suggestions to improve this work.
Dr. Ron Brown, an independent consultant, gave me a lot of encouragement
for the new approach while the book was in the outline stage. I have gone out
on a limb with several statements and Ron and John Lawlor supported me.
John is another associate and independent consultant. Our son, Bob, who
works in the field, gave me much encouragement, particularly in layer 3
switching, or is it routing?.. My wife, Paquita, kept the home fires burning.
She had great patience and gave me warm love and support.
I also received support and encouragement from the crew I worked with at
Motorola on the Celestri program which eventually became Teledesic .. In
particular, the wife-and-husband engineering team, Jill and Dave Wheeler,
gave me insight into network security. Doug White, Mr. TCPrIP, willingly
answered my many questions on such topics as time to live. Thank you all.
You are a great crew.
ROGER L. FREEMAN
Scottsdale, Arizona
December 2000
Practical Data
Communications
30
31
1
THE ENTERPRISE NETWORK
ENVIRONMENT
1.1 ENTERPRISE NETWORKS

An enterprise network is a highway that transports data from one user station
to another. An enterprise is an entity, which is either government or industry
in the broadest interpretation. Examples of government may be the USAF
Rome Laboratory or the Scottsdale, AZ, police department. Examples of
industry may be Motorola in Schaumberg, IL, the American Red Cross in
Washington, DC, or Lens auto repair shop in Warner, NH. Enterprise
networks pervade the world.
There are two types of enterprise networks: local area networks LANs.
and wide area networks WANs.. The two types differ in coverage area. The
local area network, as its name implies, covers a small, restricted area.
Commonly, it covers a portion of one floor, the entire floor, many floors, or
the whole building; in some cases, there may be workstations in several
buildings. A WAN covers a much larger area, from across town to worldwide.
We can liken a LAN to a wide highway but with just one lane, in most
cases. It accommodates large vehicles that move very fast. The WAN, on the
other hand, is a comparatively narrow-lane highway, often with one lane
coming and another going. Its car-handling capacity is notably smaller than
that of a LAN.
Another major difference between a LAN and a WAN is that an entire
LAN has single ownership. It traverses an area that is under one owner, and
the LAN is operated for the exclusive benefit of that owner. A WAN
traverses comparatively long distances where right-of-way is required, or it
traverses over the network of another entity that is in the telecommunication
transport business that will carry the connectivity* to the desired distant end.
The connectivity could be carried out over a facility owned by the enterprise.
*For an introductory discussion of connectivities, consult Ref. 1.
THE ENTERPRISE NETWORK ENVIRONMENT
Figure 1.1. An example of a LAN. Single ownership is assumed over its entire route.
Figure 1.2. An example of a WAN.
This would be a private network. With the familiar arrangement, the connectivity could be facilitated by the owners local telephone company or by a
long-distance carrier such as Sprint, MCI, France Telecom, or AT & T.
An example of a LAN is illustrated in Figure 1.1, and an example of a
WAN is shown in Figure 1.2.
The bit rate capacity of a LAN varies from 4 Mbps to 1000 Mbps; for a
WAN, from 56r64 kbps to 1.5r2.0 Mbps up to 45 Mbps or greater. The LAN
ordinarily is owned by the enterprise. Therefore there are no recurring costs
besides maintenance. The WAN customarily involves leasing, commonly from
the PSTN or other common carrier. Leasing is a recurring cost, and it
increases exponentially with bit rate and more or less linearly with distance.
Maintenance is customarily borne by the carrier leasing the circuits to the
enterprise. The exception, of course, is when the enterprise can justify a
private network. Power companies have private networks.
TYPES OF NETWORK TOPOLOGY
Often an enterprise will have many LANs. In some cases they may be
interconnected by bridges that contain routing software. Some types of LANs
have very severe distance limitations, typically Ethernet* LANs. Even with
the 2.5-km maximum extension, such a LAN can accommodate hundreds of
accesses. As the number of accesses increases, the efficiency of a LAN,
especially an Ethernet LAN, decreases. Transactions slow down; and in the
case of Ethernet, collisions and backoffs increase. Somewhere between an
18% to 35% activity factor, transactions will stop and the entire time is spent
with collision resolution.
One method to alleviate the situation is to segment the LAN into areas of
common community of interest. A bridge connects segments together. For
example, the entire accounting department is on one segment; the operations
department on another segment; and engineering on a third segment. Traffic
intensity inside a segment is high; among segments it is low.
1.2 TYPES OF NETWORK TOPOLOGY

LAN topologies may differ in many respects with PSTN topologies. Two
common topologies dominate in the LAN arena: the bus network and the
ring network. We hasten to explain that a tree network is just an extension of
a bus network.
The most elementary network topology is a point-to-point connectivity,
which is illustrated in Figure 1.3. An extension of this scheme is the
point-to-multipoint or just multipoint connectivity, illustrated in Figure 1.4.
Point-to-point connectivity may be used either in the local area or in the
Figure 1.3. Point-to-point connectivity.
Figure 1.4. Multipoint connectivity (sometimes called multidrop).

*In this text we use Ethernet and CSMArCD carrier sense multiplex access with collision
detection. interchangeably.
sense of a WAN. Figure 1.1 shows a simple bus network. A tree network is
illustrated in Figure 1.5, and a ring network is shown in Figure 1.6. Figure 1.7
is a simple star network, and Figure 1.8 shows a higher-order star network. In
some situations in the enterprise environment, star networks are used e.g.,
where a PABX acts as a data switch .. In the PSTN, the hierarchical network
is nearly universal. A hierarchical network derives from the higher-order star
network.
Figure 1.5. An example of a tree network. It is an extension of the bus network.
Figure 1.6. A ring network.
TYPES OF NETWORK TOPOLOGY
Figure 1.8. A higher-order star network. Such a network

easily evolves into a hierarchical network, typical of the
PSTN.
The two most common topologies found in the local area enterprise
environment are the busrtree configuration and the ring network. Ethernet
utilizes a tree topology; token ring, of course, uses a ring; and FDDI is based
on the ring configuration.
A LAN will have three or four types of terminals attached. These are
workstations PCs., one or more servers, and often a bridge and hubs. A
bridge connects the LAN to other LANs.rWANs. or to another segment of
the same LAN.
A serer can be broadly defined as a device that performs a service for a
user or for the network. For example, on a network running NT, a network
serer polices the network with logins and firewalls; it may also carry out
complex processing functions for a user. A print serer allows users to share a
high-speed printer. File serers contain large core storage and high-speed
processing capabilities.
Figure 1.9. A simple LAN connected in a ring topology showing the use of a wiring closet.
A hub is a centralized connecting point for a LAN. It is a multiport

device, where each LAN user is connected to a port. Physical rings or buses
are formed by internally configuring the hub ports. A typical hub may have 8
or 16 ports. Hubs can be stacked stackables . using one hub port for each
interconnection. Hubs can also have a certain amount of intelligence such as
the incorporation of a network management function. Repeaters may also be
mounted in a hub. A repeater is a simple regenerator device.
The telecommunications wiring closet performs several important functions. Figure 1.9 graphically illustrates one of the functions. First, it is the
centralized repository of common telecommunication data communication.
equipment. Among such common equipment are servers, hubs, bridges, and
so on. Second, it is the centralized location on a floor for data and voice
signal leads. These leads terminate on a distribution frame allowing easy
access by technicians and other crafts persons.
1.3 NETWORK ACCESS: AN OVERVIEW

A LAN has one single pathway connecting all users. The pathway may be a
wire pair, a length of coaxial cable, or a fiber-optic strand mounted in a
cable. In the case of FDDI there will be two strands, one passing a light
signal in one direction and the other passing light in the other direction.
Either of these three is called a transmission medium.
The transmission medium is shared by all users. It should be shared
equitably. There are two types of access to the medium: random and
controlled. The random access may also be called contention because a
station contends for use of the medium. Once a station achieves successful
access with a random access technique, it may hold that access until all its
transactions are complete. At first blush it would seem unfair and does not
meet a criterion of equitable, which we stated at the outset. Further
examination will show that it is more equitable than it first appears. The
secret lies in the average duration of a transaction. Ethernet is a typical
example of random access. A typical bit rate is 10 Mbps. Simple arithmetic
tells us that in 100 ms we have completed 1,000,000 bits of transaction or we
have passed to the distant end 125,000 bytes* of data. The average duration
of a transaction is generally not this long. Consider now that we have higher
bit rate versions of Ethernet, namely, 100 Mbps and 1000 Mbps gigabit
Ethernet ..
Token passing schemes offer controlled access. A token is a short data
packet that is passed from user to user. A user holding the token, if there is
traffic to transmit, has access to the medium for a fixed length of time. At the
*A byte in this text is assumed to have 8 bits. We prefer the term octet rather than byte because
it does not need definition.
INITIAL NETWORK DESIGN CONSIDERATIONS
end of that time duration, the user must relinquish the token to the next user
in line. If a user cannot finish transmitting the packets the station has in
queue, the user must await the next turn on the token.
We discuss these access techniques in Chapters 9 and 10.
1.4 INITIAL NETWORK DESIGN CONSIDERATIONS

The network design selection will depend largely on cost, installation simplicity, traffic load, number of accesses, distances to be covered, and
latencyrturn-around time. Consider Table 1.1 in this regard.
Some 75% of installed LANs in the United States and Canada are of the
Ethernet type. The driving factors influencing the decision to select an
Ethernet-type LAN are cost and simplicity. Network topology is dictated by
the LAN access technique. Of course, ring topologies are used with token
ring and FDDI; and busrtree topology is dictated for Ethernet. Once the
access method has been selected, the topology type is dictated by default.
WAN network designs are usually based on cost and latency. X.25 and
TCPrIP have high latency; frame relay and ATM have low latency. Latency
will also depend on distance and number of switching nodes the circuit
must pass through. The service level agreement SLA. with the carrier
should also include backup and survivability provisions. True packet networks
should involve multiple routes. Leased circuits take on the aspect of pointto-point connectivity. See Figure 1.2. Frame relay circuits from the users
viewpoint when based on permanent virtual circuits PVCs. can be seen as
point-to-point circuits.
TABLE 1.1 Initial Selection Tradeoff Analysis: Random Access and Token Passing
Item r Parameter
Cost r user
Installation simplicity
Traffic load
Number of accesses
Maximum distance
Latency r turn-around time
Random Access
(i.e., Ethernet)
Low
More simple
Breaks down at 18%
to 38% loading
Up to 1000 (in theory)
Up to 2500 m
Low when no collisions;

collisions notably increase
latency
Controlled Access
(i.e., Token Passing)
Mediumhigh
More complicated
Handles up to 100% loading
Up to 250 (token ring)
Up to 16-dB loss at 16 MHz
or 640 ft (195 m) with
CAT5 UTP. FDDI 100 km
Token passing time;
function of number of users:
10 users, low; 200 users,
mediumhigh
1.5 CONNECTION-ORIENTED AND CONNECTIONLESS SERVICE

There are two generic types of service used in data communications: connection-oriented service and connectionless service. In the OSI literature
Chapter 2. the terms are connection mode and connectionless mode of
service. The IEEE Ref. 1. defines the terms as follows:
Connection-Oriented Serice. A kind of delivery service where different
virtual circuit configurations are used to transmit messages.
Connectionless Serice. A kind of delivery service that treats each packet as
a separate entity. Each packet contains all protocol layers and destination address at each intermediate node in the network. Note: The order
of arrival of packets is not necessarily the same as the order of
transmission.
With connection-oriented service a real or virtual circuit can be traced
from origin to destination throughout the duration of the call or interchange
of packets. The analogy is the conventional telephone call.
Connectionless service can be likened to postal service. When a letter is
placed in a mailbox for delivery, we have no idea of the letters routing nor do
we care. We believe there will be a true and honest effort by the postal
service to deliver the letter to its destination. Most LANs use connectionless
service; one variation of ATM* offers connectionless service.
There are two variants of these services. They deal with acknowledgment
of receipt of packets or frames. There is connectionless service with or
without acknowledgment, and there is connection mode service with or
without acknowledgment. A common form of acknowledgment is by the use
of sequence numbers for packets or frames. This subject is introduced in
Chapter 3.
1.6 DATA PROTOCOLS: KEY TO NETWORK OPERATION
1.6.1 Introduction
To get the most out of a network, certain operational rules, procedures, and
interfaces have to be established. A data network is a big investment, and our
desire is to get the best return on that investment. One way to optimize
return on a data network is by the selection of operational protocols. We
must argue that there are multiple tradeoffs involved, all interacting one with
another. Among these are data needs to be satisfied, network topology and
architecture, selection of transmission media, switching and network management hardware and software, and operational protocols. In this section we
focus on the protocols.
*ATM stands for asynchronous transfer mode. See Chapter 14.
DATA PROTOCOLS: KEY TO NETWORK OPERATION
In the IEEE dictionary wRef. 2x, one definition of a protocol is a set of

rules that govern functional units to achieve communication. We would add
interfaces to the definition to make it more all-encompassing. In this section
we trace some of the evolution of protocols up through the International
Standards Organization ISO. OSI and its seven layers. Emphasis will be
placed on the first three layers because they are more directly involved in
communication.
This section serves as an introduction to Chapter 2, the OSI Model, and
Chapter 3, High-Level Data Link Control.
Before we delve into protocols, let us make some assumptions. First, a
data message consists of a series of bits. A bit is the lowest denominator of
information. It can be one of two possible states: a 1 or a 0. This is neat
because we can apply binary arithmetic in such a situation. Bits can be
expressed electrically fairly easily by many possible two-state arrangements.
The current is on or it is off; the voltage is positive going or negative going
just for starters.
In nearly all situations bits are taken eight at a time or in groups of eight
sequential bits. Such a grouping is called an octet preferred nomenclature.
or byte deprecated nomenclature.. A frame consists of a number of octets
or bytes. in purposeful order. A data message may consist of one or more
frames. With these assumptions, we should be able to make it through the
first several chapters of the book without any problem.
Further discussion of elementary telecommunications may be found in
Ref. 1.
1.6.2 Basic Protocol Functions
Typical protocol functions are:
Segmentation and reassembly SAR.

Encapsulation
Connection control
Ordered delivery
Flow control
Error control
A short description of each of these functions is given below.

Segmentation and Reassembly. Segmentation refers to breaking up the data
message or file into packets, frames, or blocks with some bounded size.
In asynchronous transfer mode ATMChapter 15., a message is
broken up into cells. Reassembly is the reverse of segmentation, because it involves putting the packets, frames, or blocks back into their
original order. The device that carries out segmentation and reassembly
in a packet network is called a PAD packet assemblerrdisassembler ..
10
Encapsulation. Encapsulation is the adding of header and control information to a frame in front of the text or info field and the adding of
parity-check information. The parity-check field is generally carried
after the text or info field of a frame. Use of the parity check is
described in detail in Chapter 4 in the section on data integrity Section
4.3.5..
Connection Control There are three stages of connection control:
1. Connection establishment
2. Data transfer
3. Connection termination
Of course we are talking about connection-oriented service. There is no
connection control in connectionless service.
Some of the more sophisticated protocols also provide connection interrupt and recovery capabilities to cope with errors and other sorts of interruptions.
Ordered Deliery. Packets, frames, or blocks are often assigned sequence
numbers to ensure ordered delivery of data at the destination. In a
large network with many noses and possible routes to a destination,
especially when operated in the packet mode, the packets can arrive at
the destination out of order. With a unique packet or frame numbering
plan using a simple numbered sequence, it is a rather straightforward
task for a long file to be reassembled at the destination in its original
order. Packet and frame numbering, as discussed in Chapter 4, is also a
tool used to maintain data integrity.
Flow Control. Flow control refers to the management of the data flow
from source to destination such that buffer memories do not overflow,
but maintain full capacity of all facility components involved in the data
transfer. Flow control must operate at several peer layers of protocols,
as will be discussed later.
Error Control. Error control is a technique that permits recovery of lost or
errored packets frames or blocks.. There are four possible functions
involved in error control:
1. Numbering of packets frames, blocks. to maintain a consistent
accounting system. This will indicate at the receive end if there
are packets missing. It is also a handy tool used in a data integrity
acknowledgment scheme.
2. Incomplete octets. By definition an octet must have 8 bits.
3. Incomplete frames. When key elements of a frame are missing or
are redundant. When the octet maximum in an information field
is exceeded or when the octet minimum is not complied with.
4. Bit-error detection. Error detection and error correction are
discussed in Sections 4.4 and 4.5.
REFERENCES
11
REFERENCES
1. Roger L. Freeman, Fundamentals of Telecommunications, John Wiley & Sons, New
York, 1999.
2. The IEEE Standard Dictionary of Electrical and Electronic Terms, IEEE Std. 1001996, IEEE, New York, 1996.
2
THE OSI MODEL AND THE
DATA-LINK LAYER
2.1 INTRODUCTION
The concept of Open System Interconnection OSI. and its reference model
were developed by the International Standards Organization ISO; Geneva.
in the early 1970s. The ISO impetus to develop OSI stemmed from the
proliferation of proprietary systems typically IBM 360 and UNIX, others.
that could not interoperate. The concept was that if an operating system or
protocol was developed around an open system rather than a proprietary
system, interoperation of systems became more practicable.
The IEEE Ref. 1. defines an open system as follows:
A system that implements sufficient specifications for interfaces, and supporting
formats to enable properly engineered applications software to be ported across
a wide range of systems with minimal changes, to interoperate with other
applications on local and remote systems, and to interact with users in a style
which facilitates user portability.
A key element of this definition is the use of the term open specification,
which is defined Ref. 1. as follows:
A public specification that is maintained by an open, public consensus process
to accommodate new technology over time and that is consistent with standards.
CCITT Rec.ommendation. X.200 Ref. 2. states that OSI is a reference

model used to ideally permit a full range of data communications among
disparate data equipment and networks. In other words, OSI allows standardized procedures to be defined, enabling the interconnection and subsequent effective exchange of information between users. Such users are
systems i.e., a set of one or more computers, associated software, servers,
13
14
THE OSI MODEL AND THE DATA-LINK LAYER
peripherals, terminals, human operators, physical processes, information

transfer means, etc.. that for an autonomous whole are capable of performing information processing andror information transfer. The reference model,
in particular, permits interworking between different networks, of the same
or different types, to be defined such that communication may be achieved as
easily over a combination of networks as over a single network.
2.2 LAYERING
2.2.1 Notation
In the layered OSI model the following notation is used:
N .-layer: any specific layer
N q 1.-layer: the next higher layer
N y 1.-layer: the next lower layer
2.2.2 Basic Structuring Technique

The basic structuring technique in the reference model of OSI is layering.
According to this technique, each open system is viewed as logically composed of an ordered set of subsystems, represented for convenience in the
vertical sequence shown in Figure 2.1. Adjacent systems communicate through
their common boundary. Subsystems of the same range N . collectively form
the N .-layer of the reference model of OSI. An N .-subsystem consists of
Figure 2.1. Layering in cooperating open systems.
TYPE AND INSTANCE
15
one or several N .-entities. Entities exist in each layer. Entities in the same
layer are termed peer entities. Note that the highest layer does not have an
N q 1.-layer above it and the lowest layer does not have an N y 1.-layer
below it.
Not all peer N .-entities need, nor have the ability, to communicate.
There may be conditions that prevent this communication e.g., they are not
in interconnected open systems, or they do not support the same protocol
subsets ..
2.3 TYPE AND INSTANCE

The distinction between the type of some object and an instance is a
distinction of significance for OSI. A type is a description of a class of
objects. An instance of this type is any object that conforms to this description. The instances of the same type constitute a class. A type, and any
instance of this type, can be referred to by an individual name. Each
nameable instance and the type to which this instance belongs carry distinguishable names.
For example, given that a programmer has written a computer program,
that programmer has generated a type of something, where instances of that
type are created every time the particular program is invoked into execution
by a computer. Thus a FORTRAN compiler is a type; and for each occasion
where a copy of that program is invoked in a data processing machine, it
displays an instance of that program.
Consider now an N .-entity in the OSI context. It, too, has two aspects: a
type and a collection of instances. The type of an N .-entity is defined by the
specific set of N .-layer functions it is able to perform. An instance of that
type of N .-entity is a specific invocation of whatever it is when the relevant
open system that provides the N .-layer functions is called for by its type for
a particular occasion of communication. It follows from these observations
that N .-entity types refer only to the properties of an association between
peer N .-entities, whereas an N .-entity instance refers to the specific,
dynamic occasions of actual information exchange.
It is important to note that actual communication occurs only between
N .-entity instances at all layers. It is only at connection establishment time
or its logical equivalent during a recovery process. that N .-entity types are
explicitly relevant. Actual connections are always made to specific N .-entity
instances, although a request for a connection may well be made for arbitrary
N .-entity instances of a specified type. Nothing in the OSI standard as
reflected in CCITT Rec. X.200 Ref. 2., however, precludes the request for a
connection with a specific named instance of a peer N .-entity. If an
N .-entity instance is aware of the name of its peer N .-entity instances, it is
able to request another connection to that N .-entity instance.
16
2.4 POSSIBLE SUBLAYERS

It may be necessary to further divide a layer into smaller substructures called
sublayers* and to extend the technique of layering to cover other dimensions
of OSI. A sublayer is described as a grouping of functions in a layer which
may be bypassed. The bypassing of all sublayers of a layer is not allowed.
A sublayer uses the entities and connections of its layer.
Except for the highest layer of OSI, each N .-layer provides N q 1.entities in the N q 1.-layer with N .-services. The highest layer is assumed
to represent all possible uses of the services which are provided by the lower
layers.
Note 1. Note that all open systems provide the initial source or final
destination of data; such open systems need not contain the higher layer of
the architecture as shown in Figure 2.6.
Each service provided by an N .-layer may be tailored by the selection of
one or more N .-facilities which determine the attributes of that service.
When a single N .-entity cannot by itself support a service requested by an
N q 1.-entity, it calls upon the cooperation of other N .-entities to help
complete the service requested. In order to cooperate, N .-entities in any
layer, other than those in the lowest layer, communicate by means of a set of
services provided by the N y 1.-layer see Figure 2.2.. The entities in the
lowest layer are assumed to communicate directly via the physical media
which connect them.
The services of an N .-layer are provided to the N q 1.-layer using the
N .-functions performed within the N .-layer and, as necessary, the services
available from the N y 1.-layer.
An N .-entity may provide services to one or more N q 1.-entities and
use the services of one or more N y 1.-entities. An N .-service access point
SAP. is the point at which a pair of entities in adjacent layers use or provide
services. Cooperation between N .-entities is governed by one or more
N .-protocols. The entities and protocols within a layer are shown in
Figure 2.3.
Figure 2.2. ( N q 1)-entities in the ( N q 1)-layer communicate through the ( N )-layer.

*We will note that this is done with IEEE 802-type LANs where the data link layer is divided
into two sublayers: logical link control sublayer and the medium access control sublayer. This is
described in Chapter 9.
DATA UNITS
17
Figure 2.3. ( N )-protocols between ( N )-entities.
2.5 DATA UNITS

Information is transferred in various types of data units between peer entities
and between entities attached to a specific SAP. The following are several
definitions dealing with data units:
A protocol data unit PDU. is a unit of data specified in an N .-protocol
and consisting of N .-protocol control information and possibly N .user data.
N .-interface data is information transferred from N q 1.-entity to an
N .-entity for transmission to a correspondent N q 1.-entity over an
N .-connection, or, conversely, information transferred from an N .entity to an N q 1.-entity after being received over an N .-connection
from a correspondent N q 1.-entity.
An N .-interface data unit is a unit of information transferred across the
N .-service access point between an N y 1.-entity and an N .-entity
in a single interaction. Each N .-interface data unit contains N .-interface control information and may also contain the whole or part of an
N .-service data unit.
The relationship among data units is shown in Figures 2.4 and 2.5.
Figure 2.4. Relationship among data units.
18
Figure 2.5. An illustration of mapping between data units in adjacent layers. PCI, protocol
control information; PDU, protocol data unit; SDU, service data unit. Note 1. This figure
assumes that neither segmenting nor blocking of ( N )-service data units is performed. Note 2.
This figure does not imply any positional relationship between protocol control information
and user data in protocol data units. Note 3. An ( N )-protocol data unit may be mapped
one-to-one into an ( N y 1)-service data unit, but other relationships are possible. (From
Figure 9, page 16, ITU-T Rec. X.200, Ref. 2.)
Except for the relationships defined in Figures 2.4 and 2.5, there is no
overall architectural limit to the size of data units. There may be other size
limitations at specific layers. The size of N .-interface data units is not
necessarily the same at each end of the connection. Data may be held within
a connection until a complete data unit is put into the connection.
2.6 SPECIFIC LAYERS OF THE OSI REFERENCE MODEL
The OSI reference model contains seven layers as shown in Figure 2.6. The
highest layer, the layer at the top of the figure, is the application layer, which
consists of application entities that cooperate in the OSI environment OSIE..
The lower layers provide services through which the application entities
cooperate.
Layers 16 together with the physical media for OSI provide a step-by-step
enhancement of communication services at which an OSI service recommendation is defined, while the functioning of the layers is governed by OSI
protocol recommendations.
Not all open systems provide initial source or final destination of data.
When the physical media for OSI do not link all open systems directly, some
open systems act only as relay open systems,* passing data to other open
systems. The functions and protocols that support the forwarding of data are
then provided in the lower layers as shown in Figure 2.7. Note in the figure
that the relay function only requires the first three OSI layers.
*A node that carries out a relay function.
LAYER DESCRIPTIONS
19
Figure 2.6. The seven-layer OSI reference model with peer protocols. (From Figure 11, page
28, ITU-T Rec. X.200, Ref. 2.)
Figure 2.7. Communication involving relay open systems. (From Figure 12, page 29, ITU-T
Rec. X.200, Ref. 2.)
2.7 LAYER DESCRIPTIONS

Application Layer. The purpose of the application layer is to serve as the
window between correspondent application processes actual user data. that
are using the OSI to exchange meaningful information. The application layer
necessarily provides all OSI services directly usable by application processes.
20
The application entity contains one user element and a set of application
service elements. The user element represents that part of the application
process which uses those application service elements needed to accomplish
the communication objectives of that application process. Application service
elements may call upon each other andror upon presentation services layer
6. to perform their function.
The only means by which user elements in different systems may communicate is through the exchange of application protocol data units. These
application protocol data units are generated by application service elements.
Other services that may be included in the application layer are a.
identification of intended communication partners e.g., by name and address, by definite description, and by generic description ., b. determination
of the current availability of intended communication partners, c. establishment of the authority to communicate, d. authentication of intended communication partners, e. determination of cost allocation methodology, f.
determination of adequacy of resources, g. determination of acceptable
quality of service such as response time and error rate., h. synchronization
of cooperating applications, i. selection of the dialogue discipline including
the initiation and release procedures, j. agreement on the responsibility for
error recovery, k. agreement on procedures for control of data integrity, and
l. identification of constraints on data syntax character sets, data structure ..
When a specific instance of an application process wishes to communicate
with an instance of an application process in some other open system, it must
invoke an instance of an application entity in the application layer of its own
open system. It then becomes the responsibility of this instance of the
application entity to establish an association with an instance of an appropriate application entity in the destination open system. This process occurs by
invocation of instances of entities in the lower layers. When the association
between the two application entities has been established, the application
processes can communicate.
Presentation Layer. Two definitions need to be given here:
Concrete Syntax. Those aspects of the rules used in the formal specification
of data which embody a specific representation.
Transfer Syntax. The concrete syntax used in the transfer of data between
open systems.
The presentation layer provides for the representation of information that
application entities either communicate or refer to in their communication.
The presentation layer covers two complementary aspects of this representation of information: i. the representation of data to be transferred between
application entities and ii. the representation of the data structure which
application entities refer to in their communication, along with the representation of the set of actions which may be performed on this data structure.
LAYER DESCRIPTIONS
21
The complementary aspects of the representation of information outlined

above refer to the general concept of transfer syntax. The presentation layer
is concerned with the syntax meaning the representation of data. and not
with its semantics its meaning to the application layer., which is known only
by the application entities.
The presentation layer provides for a common representation to be used
between application entities. This relieves application entities of any concern
with the problem of common representation of information, and thus it
provides them with syntax independence. The syntax independence can be
described in two ways:
1. The presentation layer provides common syntactical elements that are
used by application entities.
2. The application entities can use any syntax, and the presentation layer
provides the transformation between these syntaxes and the common
syntax needed for communication between application entities. This
transformation is performed inside the open systems. It is not used by
other open systems and therefore has no impact on standardization of
presentation protocols.
The approach in 2 above is the one commonly used.
The presentation layer provides session services layer 5. and the following
facilities: a. transformation of syntax and b. selection of syntax. Transformation of syntax is concerned with code and character set conversions, the
modification of the layout of the data, and the adaptation of actions on the
data structures. Selection of syntax provides the means of initially selecting a
syntax and subsequently modifying the selection.
Session services are provided to application entities in the form of presentation services. The presentation layer performs the following functions to
assist in accomplishing the presentation services: a. session establishment
request, b. data transfer, c. negotiation and renegotiation of syntax, d.
transformation of syntax including data transformation, formatting, and special-purpose transformations e.g., compression., and e. session termination
request.
There are three syntactic versions of the data: the syntax used by the
originating application entity, the syntax used by the receiving application
entity, and the syntax used between presentation entities the transfer syntax..
Of course, any two or all three of these syntaxes may be identical. The
presentation layer contains the functions necessary to transform between the
transfer syntax and each of the other syntaxes as required.
To meet the service requirements specified by the application entities
during the initiation phase, the presentation layer may use any transfer
syntax available to it. To accomplish other service objectives e.g., data
volume reduction to reduce data transfer cost., syntax transformation may be
22
performed either as a specific syntax-matching service provided to the application entities or as a function internal to the presentation layer.
Negotiation of syntax is carried out by communication between presentation entities on behalf of the application entities to determine the form that
data will have while in the OSI environment. The negotiations will determine
what transformations are needed if any. and where they will be performed.
Negotiations may be limited to the initiation phase, or they may occur any
time during a session.
Session Layer. Two definitions need to be given here:
Interaction Management. A facility of the session service which allows correspondent presentation entities to control explicitly whose turn it is to
exercise certain control functions.
Session-Connection Synchronization. A facility of the session service which
allows presentation entities to define and identify synchronization points
and to reset a session connection to a predefined state and to agree on a
resynchronization point.
The purpose of the session layer is to provide the means necessary for
cooperating presentation entities to organize and synchronize their dialogue
and to manage their data interchange. To do this, the session layer provides
services to a. establish a session connection between two presentation
entities and b. support orderly data exchange interactions.
To implement the transfer of data between presentation entities, the
session connection is mapped onto and uses a transport connection transport layer, layer 4..
A session connection is created when requested by a presentation entity at
a session-service access point. During the lifetime of the session connection,
session services are used by the presentation entities to regulate their
dialogue and to ensure orderly message exchange on the session connection.
The session connection exists until it is released by either of the presentation
entities or the session entities. While the session connection exists, session
services maintain the state of the dialogue even over data loss by the
transport layer.
A presentation entity can access another presentation entity only by
initiating or accepting a session connection. A presentation entity may be
associated with several session connections simultaneously. Both concurrent
and consecutive session connections are possible between two presentation
entities.
The initiating presentation entity designates the destination presentation
entity by a session address. In many systems, a transport address may be used
as the session address i.e., there is a one-to-one correspondence between the
session address and the transport address .. In general, however, there is a
LAYER DESCRIPTIONS
23
many-to-one correspondence between the session address and the transport

address. This does not imply multiplexing of session connections onto transport connections, but it does imply that at session-connection establishment
time, more than one presentation entity is a potential target of a session-connection establishment request arriving on a given transport connection.
The following services are provided by the session layer to the presentation layer: a. session-connection establishment, b. session-connection release, c. normal data exchange, d. quarantine service, e. expedited data
exchange, f. interaction management, g. session-connection synchronization, and h. exception reporting.
Transport Layer. The transport service provides transparent transfer of data
between session entities and relieves them from any concern with the
detailed way in which reliable and cost-effective transfer of data is achieved.
The transport layer optimizes the use of the available network service to
provide the performance required by each session entity at minimum cost.
This optimization is achieved within the constraints imposed by the overall
demands of all concurrent session entities and the overall quality and
capacity of the network service available to the transport layer.
All protocols defined in the transport layer have end-to-end significance,
where the ends are defined as correspondent transport entities. Therefore
the transport layer is OSI end open system oriented, and transport protocols
operate only between OSI end open systems.
The transport layer is relieved of any concern with routing and relaying
because the network service provides network connections from any transport entity to any other including the case of tandem subnetworks see
section entitled Network Layer on page 24..
The transport functions invoked in the transport layer to provide a
requested service quality depend on the quality of the network service. The
quality of the network service depends on the way the network service is
achieved see section entitled Network Layer..
The following are services provided to the session layer by the transport
layer. The transport layer uniquely identifies each session entity by its
transport address. The transport service provides the means to establish,
maintain, and release transport connections. Transport connections provide
duplex transmission between a pair of transport addresses.
More than one transport connection can be established between the same
pair of transport addresses. A session entity uses transport-connection endpoint identifiers provided by the transport layer to distinguish between
transport-connection endpoints. The operation of one transport connection is
independent of the operation of all others except for the limitations imposed
by the finite resources available to the transport layer.
The quality of service provided on a transport connection depends on the
service class requested by the session entities when establishing the transport
connection. The selected quality of service is maintained throughout the
24
lifetime of the transport connection. The session entity is notified of any

failure to maintain the selected quality of service on a given transport
connection.
The following services are provided by the transport layer: a. transportconnection establishment, b. data transfer, and c. transport-connection
release.
The following are functions performed within the transport layer: a.
mapping transport address onto network address, b. multiplexing end-toend. transport connections onto network connections, c. establishment and
release of transport connections, d. end-to-end sequence control on individual connections, e. end-to-end error detection and any necessary monitoring
of the quality of service, f. end-to-end error recovery, g. end-to-end segmenting, blocking, and concatenation, h. end-to-end flow control on individual connections, i. supervisory functions, and j. expedited transport-service
data unit transfer.
Network Layer. The following definition needs to be given here:
Subnetwork. A set of one or more intermediate open systems which provide
relaying and through which end systems may establish network connections.
The network layer provides a. the means to establish, maintain, and
terminate network connections between open systems containing communicating applications entities and b. the functional and procedural means to
exchange network-service data units between transport entities over network
connections. It provides to the transport entities independence from routing
and relay considerations associated with the establishment and operations of
a given network connection. This includes the case where several subnetworks are used in tandem or in parallel. It makes invisible to transport
entities how underlying resources such as data-link connections are used to
provide network connections.
Any relay functions and hop-by-hop service enhancement protocols used
to support the network service between the OSI end open systems are
operating below the transport layer i.e., within the network layer or below..
The basic service that the network layer provides to the transport layer is
to provide the transparent transfer of data between transport entities. This
service allows the structure and detailed content of submitted data to be
determined exclusively by layers above the network layer.
The network layer contains functions necessary to provide the transport
layer with a firm networktransport-layer boundary which is independent of
the underlying communications media in all things other than quality of
service. Thus the network layer contains functions necessary to mask the
differences in the characteristics of different transmission and subnetwork
technologies into a consistent network service. The service provided at each
LAYER DESCRIPTIONS
25
end of a network connection is the same even when a network connection

spans several subnetworks, each offering dissimilar services.
The quality of service is negotiated between the transport entities and the
network service at the time of establishment of a network connection. While
this quality of service may vary from one network connection to another, it
will be agreed upon for a given network connection and be the same at both
network-connection endpoints.
The following are services provided by the network layer: a. network
addresses, b. network connections, c. network-connection endpoints identifiers, d. network-service data unit transfer, e. quality of service parameters,
f. error notification, g. sequencing, h. flow control, i. expedited networkservice data unit transfer, j. reset, k. release, and l. receipt of confirmation.
Network Addresses. Transport entities are known to the network layer by
means of network addresses. Network addresses are provided by the network
layer and can be used by transport entities to identify uniquely other
transport entities i.e., network addresses are necessary for transport entities
to communicate using the network service.. The network layer uniquely
identifies each of the end open systems, represented by transport entities and
by their network addresses. This may be independent of the addressing
needed by the underlying layers.
A network connection provides the means of transferring data between
transport entities identified by network addresses. The network layer provides the means to establish, maintain, and release network connections. A
network connection is point-to-point. More than one network connection
may exist between the same pair of network addresses.
The network layer provides to the transport entity a network-connection
endpoint identifier that identifies the network-connection endpoint uniquely
with the associated network address. The network layer may provide sequenced delivery of network-service data units over a given network connection when requested by the transport entities.
The following are functions performed by the network layer: a. routing
and relaying, b. network connections, c. network-connection multiplexing,
d. segmenting and blocking, e. error detection, f. error recovery, g.
sequencing, h. flow control, i. expedited data transfer, j. reset, k. service
selection, and l. network layer management.
Routing and Relaying. Network connections are provided by network entities in end open systems but may involve intermediate open systems which
provide relaying. These intermediate open systems may interconnect subnetwork connections, data-link connections, and data circuits. Routing functions
determine an appropriate route between network addresses. In order to set
up the resulting communication, it may be necessary for the network layer to
use the services of the data-link layer to control the interconnection of data
circuits.
26
The control of interconnection of data circuits which are in the physical

layer. from the network layer requires interaction between a network entity
and a physical entity in the same open system. Because the OSI reference
model permits direct interaction only between adjacent layers, the network
entity cannot interact directly with the physical entity. This interaction is thus
described through the data-link layer which intervenes transparently to
convey the interaction between the network layer and the physical layer.
Data-Link Layer. The data-link layer provides functional and procedural
means to establish, maintain, and release data-link connections among network entities and to transfer data-link service data units. A data-link connection is built upon one or several physical connections. The data-link layer
detects and possibly corrects errors that may occur in the physical layer. In
addition, the data-link layer enables the network layer to control the interconnection of data circuits within the physical layer.
SERVICES PROVIDED TO THE NETWORK LAYER. In the connection-mode, the
facilities provided by the data-link layer are as follows:
a.
b.
c.
d.
e.
f.
g.
Data-link addresses
Data-link connection
Data-link service data units SDUs.
Data-link connection endpoint identifiers
Error notification
Quality of service QoS. parameters
Reset
In the connectionless mode, the facilities provided by the data-link layer

are as follows:
a. Data-link addresses
b. Transmission of data-link service data units of a defined maximum size
c. Quality of service QoS. parameters
Data-Link Addresses. Network entities are known to the data-link layer by
means of data-link addresses. Data-link addresses are provided by the
data-link layer and can be used by network entities to identify other network
entities which communicate using the data-link service. A data-link address is
unique within the scope of the set of open systems attached to a common
data-link layer.
Data-Link Connection. A data-link connection provides the means of
transferring data between network entities identified by data-link addresses.
A data-link connection is established and released dynamically.
LAYER DESCRIPTIONS
27
Data-Link Serice Data Units. The data-link layer allows for exchange of
data-link service data units over a data-link connection or exchange of
data-link-service data units that bear no relation to any other data-link
service data units. using the connectionless mode data-link service. The size
of the data-link service data units may be limited by the relationship between
the physical-connection error rate and the data-link layer error detection
capability.
Data-Link Connection Endpoint Identifiers. If needed, the data-link layer
provides data-link connection endpoint identifiers that can be used by a
network entity to identify a correspondent network entity.
Error Notification. Notification is provided to the network-entity when any
unrecoverable error is detected by the data-link layer.
Quality of Serice Parameters. Quality of service parameters may be optionally selectable. The data-link layer establishes and maintains a selected
quality of service for the duration of the data-link connection. The quality of
service parameters include mean time between detected but unrecoverable
errors, residual error rate where errors may arise from alteration, loss,
duplication, misordering, misdelivery of data-link service data units, and
other causes., service availability, transit delay, and throughput.
Reset. The network-entity can force the data-link entity invocation into a
known state by invoking the reset facility.
FUNCTIONS WITHIN THE DATA-LINK LAYER. In the connection mode and connectionless mode, the functions performed by the data-link layer are as
follows:
a.
b.
c.
d.
e.
f.
Data-link service data unit mapping

Identification and parameter exchange
Control of data-circuit interconnection
Error detection
Routing and relaying
Data-link layer management
In the connection mode, the following functions are also performed by the
data-link layer:
a. Data-link connection establishment and release
b. Connection mode data-link data transmission
c. Data-link connection splitting
28
d.
e.
f.
g.
h.
Sequence control
Delimiting and synchronization
Flow control
Error recovery
Reset
Data-Link Connection Establishment and Release. These functions establish

and release data-link connections on activated physical connection. When a
physical connection has multiple endpoints for example multipoint connection., a specific function is needed with the data-link layer to identify the
data-link service access points without establishing a data-link connection.
Connectionless Mode Data-Link Data Transmission. The connectionless
mode data-link transmission provides the means for transmission of data-link
service data units between data-link service access points without establishing
a data-link connection.
Data-Link Serice Data Unit Mapping. This function maps data-link service
data units into data-link protocol data units on a one-to-one basis.
Data-Link Connection Splitting. This function performs splitting of one
data-link connection onto several physical connections.
Delimiting and Synchronization. These functions provide recognition of a
sequence of physical service data units i.e., bitssee Physical Layer. transmitted over the physical connection as a data-link protocol data unit. Note:
These functions are often called framing..
Sequence Control. This function maintains the sequential order of data-link
service data units across a data-link connection.
Error Detection and Recoery. Error detection detects transmission, format,
and operational errors occurring either on the physical connection or as a
result of a malfunction of the correspondent data-link entity. Error recovery
attempts to recover from detection transmission, format, and operational
errors and notifies the network entities of errors which are unrecoverable.
Flow Control. In the connection mode, each network entity can dynamically control up to the agreed maximum. the rate at which it receives
data-link service data units from a data-link connection. This control may be
reflected in the rate at which the data-link layer accepts data-link service
data units at the correspondent data-link connection endpoint. In the connectionless mode, there is service boundary flow control, but no peer flow
control.
LAYER DESCRIPTIONS
29
Identification and Parameter Exchange. This function performs data-linkentity identification and parameter exchange.
Reset. This function performs a data-link reset forcing the data-link-entity
invocation to a known state.
Control of Data-Circuit Interconnection. This function conveys to network
entities the capability of controlling the interconnection of data circuits
within the physical layer.
Note: This function is used in particular when a physical connection is
establishedrreleased across a circuit-switched subnetwork by relaying within
an intermediate system between data circuits. These data circuits are elements of the end-to-end path. A network entity in the intermediate system
makes the appropriate routing decisions as a function of the path requirements derived from the network signaling protocols.
Routing and Relaying. Some subnetworks, and particularly some configurations of LANs, require that routing and relaying between individual local
networks be performed in the data-link layer.
Data-Link Layer Management. The data-link layer protocols deal with
some management activities of the layer such as activation and error
control.. See Section 8 of ITU-T Rec. X.200 Ref. 2. and ITU-T Rec. X.700
Ref. 3. for the relationship with other management aspects.
Physical Layer. The following definition needs to be given here:
Data Circuit. A communication path in the physical media for OSI between
two physical entities together with the facilities necessary in the physical
layer for the transmission of bits on it.
The physical layer provides mechanical, electrical, functional, and procedural means to activate, maintain, and deactivate physical connections for bit
transmission between data-link entities. A physical connection may involve
intermediate open systems, each relaying bit transmission within the physical
layer. Physical layer entities are interconnected by means of a physical
medium.
SERVICES PROVIDED TO THE DATA-LINK LAYER. The services or elements of
services provided by the physical layer are as follows:
a. Physical connection
b. Physical-service data units
c. Physical connection endpoints
30
Figure 2.8. Interconnection of data circuits within the physical layer. (From Figure 17, page
50, ITU-T Rec. X.200, Ref. 2.)
d.
e.
f.
g.
Data-circuit identification
Sequencing
Fault condition notification
Quality of service parameters
Physical Connections. The physical layer provides for the transparent

transmission of bit streams between data-link-entities across physical connections. A data circuit is a communication path in the physical media for OSI
among two or more physical entities together with the facilities necessary in
the physical layer for the transmission of bits on it.
A physical connection may be provided by the interconnection of data
circuits using relay functions in the physical layer. The provision of a physical
connection by such an assembly of data circuits is illustrated in Figure 2.8.
The control of the interconnection of data circuit is offered as a service to
data-link entities.
Physical Serice Data Units. A physical service data unit consists of one bit
or a string of bits.
Note: Serial or parallel transmission can be accommodated by the design
of the protocol within the physical layer.
A physical connection may allow full. duplex or half-duplex transmission
of bit streams.
Physical Connection Endpoints. The physical layer provides physical connection endpoint identifiers that may be used by a data-link entity to identify
physical connection endpoints.
Data Circuit Identification. The physical layer provides identifiers that
uniquely specify the data circuits between two adjacent open systems.
LAYER DESCRIPTIONS
31
Note: This identifier is used by network entities in adjacent open systems

to refer to data circuits in their dialogue.
Sequencing. The physical layer delivers bits in the same order in which
they were submitted.
Fault Condition Notification. Data-link entities are notified of fault conditions detected within the physical layer.
Quality of Serice Parameters. The quality of service of a physical connection is derived from the data circuits forming it. The quality of service can be
characterized by the following:
a. Error rate, where errors may arise from alteration, loss, creation, and
other causes
b. Service availability
c. Transmission rate
d. Transit delay
FUNCTIONS WITHIN THE PHYSICAL LAYER. The functions of the physical layer
are determined by the characteristics of the underlying medium and are too
diverse to allow categorization into connection mode and connectionless
mode.
The functions provided by the physical layer are as follows:
a.
b.
c.
d.
Physical connection activation and deactivation

Physical service data unit transmission
Multiplexing
Physical layer management
Physical Connection Actiation and Deactiation. These functions provide

for the activation and deactivation of physical connections between two
data-link entities upon request from the data-link layer. These include a relay
function that provides for interconnection of data circuits.
Physical Serice Data Unit Transmission. The transmission of physical service data units i.e., bits. may be synchronous or asynchronous. Optionally,
the function of physical service data unit transmission provides recognition of
the protocol data unit corresponding to a mutually agreed sequence of
physical service data units that are being transmitted.
Multiplexing. This function provides for two or more physical connections
to be carried on a single data circuit. This function provides the recognition
of the framing required to enable identification of the physical layer PDUs
32
Figure 2.9. Examples of interconnection. (From Figure 19, page 52, ITU-T Rec. X.200,
Ref. 2.)
conveyed by the individual physical connections over the single data circuit.
The multiplexing function is optional.
Note: A particular example of the use of multiplexing is offered when a
transmission media is divided into data circuits in support of the different
data link protocols used in the signaling phase and in the data transfer phase
when using circuit-switched subnetworks. In such usage of multiplexing, flows
of different nature are permanently assigned to different elements of the
multiplex group.
Physical Layer Management. The physical layer protocols deal with some
management activities of the layer such as activation and error control.. See
Section 8 of Ref. 2 and ITU-T Rec. X.700 Ref. 3. for the relationship with
other management aspects.
Note: The above text deals with interconnection between open systems as
shown in Figures 2.6 and 2.7. For open systems to communicate in the real
environment, real physical connections should be made as, for example, in
Figure 2.9a. Their logical representation is as shown in Figure 2.9b and is
called the physical media connection. The mechanical, electromagnetic, and
other media-dependent characteristics of physical media connections are
defined at the boundary between the physical layer and the physical media.
Definitions of such characteristics are specified in other standards which
shall be discussed below.
SPECIFIC COMMENTS ON OSI
33
Figure 2.10. The OSI reference model, conventional drawing.
2.8 SPECIFIC COMMENTS ON OSI

2.8.1 General
The following are comments regarding Section 2.7, which was extracted from
ITU-T Rec. X.200 Ref. 2.. This section consists of the authors interpretations of OSI layers 1 through 4, where one might consider layer 4 as the end
of the data communicators responsibility.
Figure 2.10, the OSI reference model, repeats Figure 2.6, but is drawn in
the more conventional manner. The purpose of the model is to facilitate
communication data entities. It takes at least two to communicate. Thus we
consider the model in twos: one entity on the left side of the figure and one
on the right. The terms peer and peers are used. Peers are corresponding
entities on each side of Figure 2.10 e.g., the left-side link layer is peer to the
right-side link layer., and they communicate by means of a common protocol.
We wish to add that there is no direct communication between peers except
at the physical layer layer 1.. That is, above the physical layer, each protocol
entity sends data down to the next lower layer to get data across to its peer
entity on the other side.
We can consider each layer of the OSI model as a functional module. It
depends on the module below it layer N y 1. for service, and it provides
service to the module above it layer N q 1..
As was pointed out in Section 2.7, each entity communicates above it and
below it across an interface. The interface is at a service access point SAP.,
which is shown in Figure 2.11. The N .-connection endpoint identifier
distinguishes an N .-connection when an N q 1.-entity establishes an N .connection with another N q 1.-entity as illustrated in Figure 2.11.
34
Figure 2.11. Entities, service access points (SAPs), and identifiers. (From Figure 7, page 14,
ITU-T Rec. X.200, Ref. 2.)
The actual communication between entities is by means of primities.*

A primitive specifies a function to be performed and is used to pass data and
control information. ITU-T Rec. X.210 Ref. 4. describes four types of
primitives used to define the interaction between adjacent layers of the OSI
architecture. A brief description of each of these primitives follows:
Request. A primitive issued by a service user to invoke some procedure and
to pass parameters needed to fully specify the service.
Indication. A primitive issued by a service provider either to invoke some
procedure or to indicate that a procedure has been invoked by a service
user at the peer service access point.
Response. A primitive issued by a service user to complete at a particular
SAP some procedure invoked by an indication at the SAP.
Confirm. A primitive issued by a service provider to complete, at a particular
SAP, some procedure previously invoked by a request at that SAP.
The ITU-T reference adds that confirms and responses can be positive or
negative, depending on the circumstances.
2.9 DISCUSSION OF OSI LAYERS 14
2.9.1 The Physical Layer (Layer 1)
Layer 1 is the lowest layer of the OSI model. It provides the physical
connectivity between two data end users who wish to communicate. The
services it provides to the data-link layer layer 2. are those required to
*We call these generic primitives. They are widely used in the world of protocols, which we
consider to be the heart of data communications.
DISCUSSION OF OSI LAYERS 14
35
connect, maintain connected, and disconnect the physical circuits that form
the physical connectivity. The physical layer describes the traditional interfaces between data terminal equipment DTE. and data communication
equipment DCE.. These interfaces are described in Section 6.4 Chapter 6..
The physical layer has four important characteristics:
Mechanical
Electrical
Functional
Procedural
The mechanical aspects include the actual cabling and connectors necessary
to connect the communications equipment to the transmission medium.
Electrical characteristics include voltage, signal sense,* impedance, and balanced or unbalanced circuits for signal and control. Functional characteristics
cover connector pin assignments at the interface and the precise meaning
and interpretation of the various interface signals and data set controls.
Procedures cover sequencing rules that govern the control functions necessary to provide higher-layer services such as establishing a connectivity across
a switched network.
Some of the applicable standards for the physical layer are as follows:
EIArTIA-232, EIArTIA-422, EIArTIA-423, and EIArTIA-530

ITU-T CCITT. Recs. V.10, V.11, V.24, V.28, V.35, X.20, and X.21 bis
ISO 2110, 2593, 4902, and 4903
US Fed. Stds 1020, 1030, and 1031
US Mil-STD-188-114B
The reader is reminded that often standards from one organization

reflect the restatement of standards from another organization. For
example, MIL-STD-114B restates much of EIArTIA-232, EIArTIA-422,
and EIArTIA-423.
2.9.2 The Data-Link Layer (Layer 2)
The data-link layer provides services for the reliable interchange of data
across a data link established by the physical layer. Link layer protocols
manage the establishment, maintenance, and release of data-link connections. Data-link protocols also control the flow of data and supervise data
error recovery. A most important function of this layer is recovery from
abnormal conditions. The data-link layer services the network layer layer 3.
*Signal sense: associating voltage polarity with a binary 1 or binary 0.
EIA stands for Electronics Industries Alliance; TIA stands for Telecommunications Industry
Association.
36
Figure 2.12. Generalized data-link layer frame. The address and control fields taken together
are often referred to as the header.
or logical link control LLC in the case of LANs. and inserts a payload data
unit into the INFO portion of the link layer. A generic data-link layer frame
is illustrated in Figure 2.12.
There is a fair universality of this type of frame for the data-link layer
across numerous data protocol families. Chapter 3 contains a detailed discussion of one application of a typical data-link layer frame.
Moving from left to right in Figure 2.12, the first field is the flag field. In
nearly all data-link frames of this type, this field is the bit sequence 01111110.
It tells the receiving processor of the frame to start counting bitsroctets
bytes. at the end of the 8th bit of the flag sequence. The receiving processor
knows a priori the length of the address and control fields so it can delineate
these two. It also knows by counting octets. that the end of the control field
is the beginning of the information field.
The information field, depending on the protocol, may be of fixed length
or variable length. If it is of fixed length, the processor will have this length
information a priori. For the variable-length case, the control field will carry
that length information which will tell the receiving processor what length to
expect.
The processor continues its counting of bytes octets. and will know the
end boundary of the information field. It will now know that the next two
octets bytes. carry the frame check sequence FCS. information for the case
of a WAN, and it will know that the next four octets carry the FCS in the
case of a LAN.
The functions of the fields are described as follows. The flag provides the
marker for the beginning of operations. It delimits the frame at both ends. If
only one frame is to be transmitted, the flag bit sequence not only starts the
frame but it is also appended at the end of the frame to indicate termination.
If another frame follows directly after the first frame, the ending flag
sequence of the first frame is the beginning flag sequence of the second
frame.
In some types of data-link layer protocols the address field only contains
the destination address. However in most cases, it contains both the destination address es. and the source address.* In other words, the address field
tells the receiving processor who is the intended recipient s. of the data
message or frame. In some cases a second address field follows the destination address. This is the source address field that tells the receiving processor
who originated the data message or frame.
*The source is the originator of the frame or packet.
DISCUSSION OF OSI LAYERS 14
37
The control field covers a number of functions such as sequence numbering, a commandresponse bit, and type of frame such as information frame,
supervisory frame, and unnumbered frame. The control frame will often have
a flow control responsibility, usually incorporated in the sequence numbering. In the cases of a variable-length frame, it will give some indication to the
receiving processor the length of the information field.
The information field contains a payload service data unit inserted by the
network layer. For fixed-length frames, the information field will contain
padding if there are insufficient bytes octets. inserted by the network layer
to fulfill minimum byte octet. requirements. There must be some sort of
indicator to tell the processor that there is padding and how much to expect.
The frame check sequence FCS. is the primary tool used to detect
transmission errors. See Section 4.4 for a detailed discussion of the operation
of an FCS.
2.9.3 The Network Layer (Layer 3)
The network layer moves data through the network. At relay or switching
node along the traffic route, layering concatenates. In other words, the
higher layers above layer 3. are not required, and they are used only at user
endpoints. The user endpoints and relay function are shown in Figure 2.7.
The network layer carries out the functions of switching and routing,
sequencing, and logical channel control. We note the duplication of error
recovery with the data-link layer. Link layer error recovery is only on the link
it is responsible for. Network layer error recovery is network-wide.
The network layer also provides and manages logical channel connections
between points in the network such as virtual circuits across the PSTN. On
simpler data connectivities, such as point-to-point, point-to-multipoint, and
LANs, the network layer is not required and can be eliminated. A packetswitched network, on the other hand, is a typical application where the
network layer is most necessarily needed.
The best-known network layer standard is ITU-T Rec. X.25. Other protocols are ISO 8348 and the Internet Protocol IP of the TCPrIP family of
protocols.. However, in this chapter we treat IP in its specialized function of
interfacing between disparate networks in an artificial layer called layer 3.5.
2.9.4 Internet Protocol (IP) and Gateway Sublayer (Layer 3.5)
William Stallings in his Handbook of Computer-Communication Standards
Ref. 5., has introduced a sublayer called layer 3.5. This sublayer provides
functions for interfacing different networks. When one of the networks is a
LAN and the other a WAN, Stallings uses the term gateway as the device
that performs the interface function. Commonly, this function is performed
by a router or even a bridge.
A popular protocol for interfacing disparate networks is the Internet
Protocol IP., which is usually joined with a related protocol called Transmis-
38
Figure 2.13. Sublayering of OSI layer 3 to achieve internetting.
sion Control Protocol TCP..* Thus we have TCPrIP, which was developed
for the U.S. Department of Defense DoD.. It is now one of the most widely
used protocols in the commercial sector for OSI layers 3 and 4.
One applicable gateway protocol is ITU-T Rec. X.75 Ref. 6., which is a
companion to X.25 Ref. 7.. Figure 2.13 shows the sublayering of OSI layer 3
for internetworking or internetting as suggested by Stallings Ref. 5..
2.9.5 The Transport Layer (Layer 4)
The transport layer layer 4. is the highest layer of the services associated
with the telecommunications provider. From our perspective, OSI layers 14
are the responsibility of the telecommunications system engineer, whereas
layers 5, 6, and 7 are the responsibility of the data end user.
The transport layer has the ultimate responsibility of providing a reliable
end-to-end data delivery service for higher-layer users. It is defined as an
end-system function, located in the equipment using network service or
services. In this way its operations are independent of the characteristics of
all the networks that are involved. Services that a transport layer provide are
as follows:
Connection Management. This includes establishing and terminating connections between transport users. It identifies each connection and
negotiates values of all needed parameters.
Data Transfer. This involves the reliable delivery of transparent data between users. All data are delivered in sequence with no duplication or
missing parts.
Flow Control. This is provided on a connection basis to ensure that data
are not delivered at a rate faster than the users resources can accommodate.
*In fact TCPrIP is part of a rather large and popular family of protocols discussed in Section
11.4.
REFERENCES
39
Transmission Control Protocol TCP, Ref. 8. was the first working version
of a transport protocol and was created, as was the rest of the TCPrIP family
of protocols, by DARPA Defense Advanced Research Projects Agency. for
its DARPANET. All the features of TCP have been adopted in the ISO
version. The applicable ISO references are ISO 8073, called Transport
Protocol Specification, and ISO 8072, called Transport Serice Definition.
Many readers will recognize the TCPrIP family of protocols for the part
they play in the popular, and nearly ubiquitous, INTERNET.
2.10
2.10.1
PROCEDURAL VERSUS ELECTRICAL

Narrative
Up to this point we have discussed how to connect data facilities together

forming a network based on a selected topology Chapter 1.. More emphasis
has been placed on procedural issues in Chapters 2, 3, and 4. The objective is
to transmit a serial bit stream from an origin to an intended destination. This
involves both procedural issues and electrical circuit standards. The protocols
which we will be dealing with mostly concentrate on standardized procedures. In the LAN arena, the Medium Access Control MAC. is so closely
allied to its electrical signal that the procedural and electrical cannot be
separated and yet still present a reasonable argument.
To meet these requirements, Chapters 5 and 6 present information on
basic data transmission, Chapter 7 introduces the reader to voice-channel
modems, and Chapter 8 gives an overview of the public switched telecommunications network PSTN.. The PSTN is a major provider of long-distance
data circuits. Chapters 9 and 10 pick up on the issues of LAN protocols;
Chapters 10 through 13 deal with long-distance data communication including WANs.
REFERENCES
1. James Isaak, Kevin Lewis, Kate Thompson, and Richard Straub, Open Systems
Handbook, IEEE Standards Press, IEEE, New York, 1994.
2. Information TechnologyOpen Systems InterconnectionBasic Reference Model:
The Basic Model, ITU-T Rec. X.200, ITU Geneva, July 1994.
3. Management Framework for Open Systems Interconnection (OSI) for CCITT Applications, CCIT Rec. X.700, ITU Geneva, September 1992.
4. Information TechnologyOpen Systems InterconnectionBasic Reference Model:
Conentions for the Definition of OSI Serices, ITU-T Rec. X.210, ITU Geneva,
November 1993.
40
5. William Stallings, Handbook of Computer Communications, Vol. 1, Macmillan, New

York, 1987.
6. Packet-Switched Signaling System Between Public Networks Proiding Data Transmission Serices, ITU-T Rec. X.75, ITU Geneva, October 1996.
7. Interface Between Data Terminal Equipment (DTE) and Data Circuit-Terminating
Equipment (DCE) for Terminals Operating in the Packet Mode and Connected to
Public Data Networks by Dedicated Circuit, ITU-T Rec. X.75, ITU Geneva, October
1996.
8. Transmission Control Protocol (TCO), RFC 793, DDN Network Information Center, SRI International, Menlo Park, CA, September 1981.
3
HIGH-LEVEL DATA-LINK CONTROL
(HDLC) TYPICAL DATA-LINK
LAYER PROTOCOL
3.1 INTRODUCTION
The majority of the data-link layer protocols in use today derive from HDLC
Ref. 1.. It was developed by the International Standards Organization ISO.,
and the relevant standard is ISO 3309. The following is a partial list of
data-link layer standards spawned by HDLC:
1. Advanced Data Communications Control Procedures ADCCP., developed by ANSI. It has been adopted by the U.S. government for official
use. See FIPS Pub. 71-1 and FED-STD-1003A. It is identical to
HDLC.
2. Link Access Procedure Balanced LAPB. operates in the HDLC balanced mode only, otherwise it is almost identical to HDLC. It is
supported by the ITU-T Organization and is incorporated as part of the
ITU-T Rec. X.25 document Ref. 2.. See Section 11.3.2.
3. Link Access Procedure D-Channel LAPD.. Originally designed for the
ISDN D-channel, Basic Rate Service. Variants of LAPD are used in
the frame relay service see Section 13.7..
4. Synchronous Data-Link Control SDLC. was developed by IBM, particularly for SNA, and has wide use in that arena.
5. Logical Link Control LLC. resides in the upper sublayer 2 of all IEEE
802 and FDDI LANs. See Section 9.4.
In the next section, we define some terms used in HDLC.
41
42
HIGH-LEVEL DATA-LINK CONTROL (HDLC) TYPICAL DATA-LINK LAYER PROTOCOL
3.2 STATIONS AND CONFIGURATIONS

Primary Station. A logical primary station is an entity that has primary link
control responsibility. It assumes responsibility for organization of data
flow and for link level error recovery. Frames issued by a primary station
are called commands.
Secondary Station. A logical secondary station operates under control of a
primary station. It has no direct responsibility for control of the link but
instead responds to primary station control. Frames issued by a secondary
station are called responses.
Combined Station. A combined station combines the features of primary and
secondary stations. It may issue both commands and responses.
Unbalanced Configuration. An unbalanced configuration consists of a primary station and one or more secondary stations. It supports both fullduplex and half-duplex operation and point-to-point and multipoint
circuits.
Balanced Configuration. A balanced configuration consists of two combined
stations in which each station has equal and complementary responsibility
of the data link. A balanced configuration operates only in the point-topoint mode and supports full-duplex and half-duplex operation. Figure 3.1
shows HDLC stations and configurations.
There is also a symmetric configuration that connects two unbalanced
stations multiplexed through a single point-to-point connection. Transmission
may be in the half- or full-duplex mode over nonswitched or switched
facilities. There are logically two separate streams of command and response
frames. A station may be configurable in that it can operate as a different
type in a different environment.
Figure 3.1. HDLC stations and configurations.
HDLC FRAME STRUCTURE
43
3.3 MODES OF OPERATION USED WITH HDLC

In order for a data-link mechanism to provide the needed services, different
relationships between the involved stations must be established. HDLC provides for a variety of relationships as well as the means to change them in the
most appropriate ways.
Response modes establish a relationship between stations that provides
for information transfer. The response modes must be agreed upon before
information can be transferred between stations.
Three data transfer modes are described:
Normal Response Mode (NRM). In this mode a primary station initiates data
transfer to a secondary station. A secondary station transmits data only in
response to a poll from a primary station. This mode of operation applies
to an unbalanced configuration.
Asynchronous Response Mode (ARM). In this mode a secondary station may
initiate transmission without receiving a poll from a primary station.
It is useful on a circuit where there is only one active secondary
station. The overhead of continuous polling is thus eliminated.
Asynchronous Balanced Mode (ABM). This is a balanced mode that provides
symmetric data transfer capability between combined stations Figure 3.1b..
Each station operates as if it were a primary station, can initiate data
transfer, and is responsible for error recovery. One application of this mode
is hub polling, where a secondary station needs to initiate transmission.
3.4 HDLC FRAME STRUCTURE

Figure 3.2 shows the format of an HDLC frame. Moving from left to right in
the figure in the order of transmission, we have the flag field F., which
delimits the frame at both ends with the unique bit pattern 01111110. If
frames are sent sequentially, the closing flag of the first frame is the opening
flag of the next frame see Section 2.9.2..
The flag field is a special bit pattern, and receiving stations are constantly
searching for the flag bit sequence to mark the beginning and end of a frame.
There is no reason why this same sequence cannot appear somewhere in
midframe. This, then, would incorrectly tell a receiving station that the frame
has ended and that a new frame has begun. Of course, the misinterpretation
corrupts the entire frame and probably some subsequent frames. To avoid
this problem, no sequence of six consecutive 1s bracketed by 0s is permitted
in other than the flag field locations.
Lets consider this a little more carefully. To ensure transparent transmission of everything between the delimiting flags, a zero insertion scheme is
employed. This prevents a flag sequence from appearing inside a frame
44
Figure 3.2. The HDLC frame format.
where it would be erroneously interpreted as an end of frame delimiter.

There are three cases that must be dealt with: a. a valid flag pattern, b. a
frame abort pattern seven or more consecutive 1s., and c. data that can
look like either. The transmitter will insert a zero after 5 consecutive 1s in
the data stream. The receiver will monitor the data stream and will look for a
pattern of a binary 0 and 5 binary 1s. If the next bit is a binary 0, it is
discarded and the binary 1s are passed as data. If the sixth bit was a binary 1,
the next bit is tested. If it is a binary 0, a flag is received. If it was a binary 1,
the frame reception is aborted. Transmission of a frame is aborted by sending
between 7 and 14 consecutive binary 1s in the data stream. An abort
sequence will cause the receiver to discard any accumulated data and ready
itself for new frame reception.
By the use of zero insertion, HDLC is transparent to any 8-bit sequence.
The address field A. immediately follows the opening flag of a frame and
precedes the control field C.. Each station in a network normally has an
individual address and a group global. address. A group address identifies a
family of stations. It is used when data messages must be accepted from or
destined to more than one user. Normally, the address is 8 bits long,
providing 256 bit combinations or addresses. In HDLC and ADCCP the
address field can be extended in increments of 8 bits. When this is implemented, the least significant bit is used as an extension indicator. When that
bit is 0, the following octet is an extension of the address field. The address is
terminated when the least significant bit of an octet is 1. It thus can be seen
that the address field can be extended indefinitely.
Commands always carry the address of the destination, and responses
always carry the address of the source.
The control field C. immediately follows the address field A. and precedes the information field I.. The control field conveys commands, responses, and sequence numbers to control the data link. The basic control
45
Figure 3.3. The three control field formats of HDLC.
field is 8 bits long and uses modulo-8 sequence numbering. There are three
types of control field: i. I frame information frame., ii. S frame supervisory frame., and iii. U frame unnumbered frame.. The three control field
formats are shown in Figure 3.3.
The control field may be extended by the addition of a second contiguous
octet immediately following the basic field. The extension increases the
modulus count to 128. The extended control field is often used on long
propagation delay circuits such as geostationary satellite circuits, basically to
extend the sequence modulus number.
Consider the basic 8-bit form of the control field as shown in Figure 3.3.
The information follows from left to right. If the first bit in the control field
is a binary 0, the frame is an I frame Figure 3.3a.. If the bit is a 1, the frame
is an S frame or a U frame Figures 3.3b and 3.3c.. If the first bit is a 1
followed by a 0, it is an S frame; and if the bit is a 1 followed by a 1, it is a U
frame. These bits are called format identifiers.
3.4.1 Sequence Numbering in HDLC
Turning now to the I figure Figure 3.3a., its purpose is to carry user data.
Here bits 2, 3, and 4 of the control field in this case carry the send sequence
of transmitted messages I frames.. Sequence numbers are used to detect lost
or out-of-sequence frames. This provides one check on the integrity of the
received data. Sequence numbers are controlled by send and receive state
variables and are represented by send and receive sequence numbers that
appear in the control field.
46
Each information frame is sequentially numbered and may have the value
0 through MODULUS minus 1, where MODULUS is the MODULUS of the
sequence number. Modulus equals 8 for the basic control field and 128 for
the extended control field. The sequence numbers cycle through the entire
range.
The maximum number of sequentially numbered information format
frames that a station may have outstanding i.e., unacknowledged . at any
given time may never exceed one less than the modulus of the sequence
numbers. This restriction is to prevent any ambiguity in the association of
transmission frames with sequence numbers during normal operation andror
error recovery actions.
Conceptually, sequence number control is often represented as a window
that is open to the value equal to the number of unacknowledged frames
allowed before transmission is halted flow control.. The two variables and
two numbers involved are as follows:
Send State Variable V ( S ). This is the sequence number of the next insequence information frame to be transmitted.
Send Sequence Number N( S ). The value of V S . is inserted in the control
field of a transmitted information frame as N S ..
Receive State Variable V ( R ). This denotes the sequence number of the next
expected in-sequence information frame to be received.
Receive Sequence Number N( R ). The value of V R . is inserted into the
control field of a transmitted frame to indicate that frames numbered
N R . y 1 and lower have been received.
To describe window, consider data link X to Y. The receiver Y has seven
buffers the number was selected arbitrarily .. Thus Y can accept seven
frames or messages. and X is allowed to send seven frames without acknowledgment. To keep track of which frames have been acknowledged, each is
labeled with a sequence number 07 modulo 8.. Station Y acknowledges a
frame by sending the next sequence number expected. For instance, if Y
sends a sequence number 3, this acknowledges frame number 2 and is
awaiting frame number 3. Such a scheme can be used to acknowledge
multiple frames. As an example, Y could receive frames 2, 3, and 4 and
withhold all acknowledgments until frame 4 arrives. By sending sequence
number 5, it acknowledges the receipt of frames 2, 3, and 4 all at once.
Station X maintains a list of sequence numbers that it is allowed to send, and
Y maintains a list of sequence numbers it is prepared to receive. These lists
are thought of as a window of frames.
HDLC allows a maximum window size of 7, or 127 frames. In other words,
a maximum number of 7, or 127 unacknowledged frames, can be sent, or one
less than the modulus 8 or 128. N S . is the sequence number of the next
frame to be transmitted, and N R . is the sequence number of the frame to
be received.
47
3.4.2 The Pollr

r Final Bit
Each frame carries a pollrfinal PrF. bit. It is bit 5 in each of the three
different types of basic. control fields, and it is bit 9 in the extended control
field. The PrF bit serves a function in both command and response frames.
In a command frame it is referred to as a poll P. bit, and in a response
frame it is referred to as a final F. bit. In both cases the bit is set to 1. The P
bit is used to solicit a response or sequence of responses from a secondary or
a balanced station. On a data link only one frame with a P bit set to 1 can be
outstanding at any given time. Before a primary or a balanced station can
issue another frame with a P bit set to 1, it must receive a response from a
secondary or a balanced station with the F bit set to 1. In the NRM mode,
the P bit is set to 1 in command frames to solicit response frames from the
secondary station. In this mode of operation the secondary station may not
transmit until it receives a command frame with the P bit set to 1.
The F bit is used to acknowledge an incoming P bit. A station may not
send a final frame without prior receipt of a poll frame. As can be seen, P
and F bits are exchanged on a one-for-one basis. Thus only one P bit can be
outstanding at a time. As a result, the N R . count of a frame containing a P
or F bit set to 1 can be used to detect sequence errors. This capability is
called check pointing. It can be used not only to detect sequence errors but to
indicate the frame sequence number to begin retransmission when required.
3.4.3 Supervisory Frames
Supervisory S. frames, shown in Figure 3.3b, are used for flow and error
control. Both go-back-n and selective ARQ can be accommodated.* There
are four types of supervisory or S. frames:
S Frame
1.
2.
3.
4.
Receive ready RR.

Receive not ready RNR.
Reject REJ.
Selective reject SREJ.
Bits 1 to 8
1000 PrF
1001 PrF
1010 PrF
1011 PrF
N R.
N R.
N R.
N R.
The RR frame is used by a station to indicate that it is ready to receive

information and acknowledge frames up to and including N R . y 1. Also a
primary station may use the RR frame as a command with the poll F. bit set
to 1.
The RNR frame tells a transmitting station that it is not ready to receive
additional incoming I frames. It does acknowledge receipt of frames up to
*Go-back-n and selectie ARQ are forms of error recovery. Errors and error recover are
described in Section 4.4.
48
and including sequence number N R . y 1. I frames with sequence number

N R . and subsequent frames, if any, are not acknowledged. The REJ frame
is used with go-back-n ARQ to request retransmission of I frames with frame
sequence number N R ., and N R . y 1 frames and below are acknowledged.
3.4.4 Unnumbered (U) Frames
The unnumbered U. frames are used for a variety of control functions. They
do not carry sequence numbers, as the name indicates, and do not alter the
flow of sequencing I frames. Unnumbered U. frames can be grouped into
the following four categories:
1.
2.
3.
4.
Mode-setting commands and responses

Information transfer commands and responses
Recovery commands and responses
Miscellaneous commands and responses
There are also unnumbered information UI. frames that furnish a means of
providing connectionless service at the data-link level. The UI frames are
transmitted without disturbing the N S . and N R . values used in normal I
frame transmissions. Because of this, the sequencing and verification facilities provided by I, RR receive ready command., SREJ selective reject
command., and so on, are not applicable. The UI frame may be lost or
duplicated without any notification to the sender. If a more reliable environment for UI is needed, similar mechanisms to the normal I frames must be
constructed at an OSI layer higher than the data-link layer. The UI frames
can be used for higher layer procedures or for interruption of the normal
data transfer mechanisms without disturbing them. Prioritized or real-time
transmissions can be accommodated by the UI frame.
3.4.5 Information Field
The information field I. follows the control field, as illustrated in Figure 3.2.
The information field is present only in information I. frames and some
unnumbered U. frames. The I field may carry any number of bits in any
code, related to character set structure or not. Its length is not specified in
the ISO underlying standard wISO 3309 Ref. 1.x. Specific system implementations, however, usually place an upper limit on I field size. Some implementations of HDLC require that the I field contain an integral number of octets.
3.4.6 Frame Check Sequence (FCS) Field
The frame check sequence field FCS. follows the information field or the C
field if there is no I field. and is carried in each frame. The FCS field detects
COMMANDS AND RESPONSES
49
errors due to transmission. The FCS field contains 16 bits, in most implementations, which are a result of a mathematical computation on the digital value
of all bits excluding inserted 0s zero insertion. in the frame and including
the address, control, and information fields.
The CRC used in HDLC is based on the CRC-16 FCS developed by
CCITT in Rec. V.41 Ref. 7. using generating polynomial X 16 q X 12 q
X 5 q 1. It should be noted, however, that in some situations that require
more stringent undetected error conditions andror because of frame length,
a 32-bit CRC FCS. may be used. This is similar to the 32-bit CRC used with
LANs, described in Chapter 9.
Error recovery and the operation of the FCSrCRC are described in
Section 4.4.
3.5 COMMANDS AND RESPONSES

Table 3.1 gives a listing of commands and responses used in the HDLC
family of protocols.
3.5.1 Mode-Setting Commands

A set of eight mode-setting commands are described below. These are used
to establish the mode of the connected receiving station. When an E is
appended to the end of the abbreviation, it indicates an extended mode.
These commands, of course, are used in the U format, and the mode setting
information is carried in the M or modifier bits shown in Figure 3.3c, bits 3,
4, 6, 7, and 8.
Set normal response mode SNRM . command is used to place the addressed secondary station in the normal response mode NRM. where all
control fields are one octet in length. This command places the secondary
receiving station under control of the transmitting primary station. When in
the normal response mode, no unsolicited transmissions are allowed from a
secondary station. Bit sequence 1100P001.
Set asynchronous response mode SARM . command places the addressed
secondary station in asynchronous response mode ARM.. Bit sequence
1111P000.
Set asynchronous balanced mode SABM . places the addressed balanced
station in the ABM. Bit sequence 1111P100.
There are three extended modes. SNRME, SARME, and SABME are the
previous three mode commands extending control fields to two octets in
length. Bit sequences: SNRME 1111P011; SARME 1111P010; and SABME
1111P110.
SIM or set initialization mode command starts up system-specified link-level
initialization procedures at a remote station.
50
COMMANDS AND RESPONSES
51
Disconnect DISC . command performs a logical disconnect. It informs the

receiving station that the transmitting station is suspending operation with
the receiving secondary or balanced station. The DISC command may also be
used to initiate a physical disconnect in a switched network. Secondary
stations remain disconnected until receipt of an SNRM or SIM command.
3.5.2 Miscellaneous Commands

There are four miscellaneous commands, which are described below.
Unnumbered poll UP. command solicits response frames from one or a
group of secondary stations to establish a logical operation condition. Bit
sequence: 1100P100.
TEST command is used to test the data-link layer. It causes the addressed
station to reply with a TEST response. Bit sequence: 1100P111.
RSET or reset command is used by a combined station to reset the receive
state variable and the FRMR condition. Bit sequence: 1111P001.
The exchange identification XID. command causes the addressed station
to report its station identification. It can be used optionally to provide the
station identification of the transmitting station to the remote station. With
this command an information I. field is optional. If it is used, it will contain
the identification ID. of the transmitting station. If the unique address of
the secondary station is unknown, the XID command may use the global
group. address. Bit sequence: 1111P101.
3.5.3 Responses to Unnumbered Commands

The request disconnect RD. is used to indicate to the remote station that
the transmitting station wishes to be placed in a disconnected mode, or it is
used to request a physical disconnect operation on switched networks. Bit
sequence: 1100F010.
The unnumbered acknowledge UA. command acknowledges receipt and
acceptance of numbered commands. Bit sequence: 1100F110.
The TEST response replies to a TEST command in any mode. Bit
sequence: 1100F111.
The request for initialization RIM. is used to inform another station of
the need of a SIM command. No command transmissions are accepted until
the RIM condition is reset by receipt of a RIM or DISC. The receipt of any
command except a RIM or DISC will cause the station to repeat the RIM.
Bit sequence: 1110F000.
When there is an errored condition that is not recoverable by retransmission of the errored frame, the frame reject response FRMR. is used. These
conditions are as follows: a command that is not implemented or is invalid,
an information field which exceeds maximum length, and an invalid receive
sequence number. Bit sequence: 1110F001.
52
The disconnected mode DM. response reports a nonoperational status

where the station is logically disconnected from the link. The DM response
requests the remote station to issue a set mode command or, if sent in
response to the reception of a set mode command, to inform the addressed
primary station that the transmitting station still cannot respond to the set
mode command. Bit sequence: 1111F000.
The exchange identification XID. response replies to an XID command.
An information field containing the ID of the transmitting secondary station
is optional. Bit sequence: 1111F101.
Note: In Sections 3.5.1 through 3.5.3, where a control frame bit sequence
was given, the bit position 5 was a letter, either a P or an F. The incisive
reader would have deduced correctly that bit 5 was the PrF bit. If a P were
given, then the PrF bit would be in the poll state; where it was F, of course
it would be in the final state.
3.6 FRAME OPERATION

An HDLC link is in an actie state when a station is transmitting frames, an
abort sequence, or interframe time fill. When in an active state, the transmitting station reserves the right to continue transmission.
An HDLC link is in an idle state when it is not in an active state. An idle
state is identified by the detection of a continuous transmission of 1s that
persists for 15 or more bit times. This is called the idle sequence.
When a transmitting station wishes to maintain a link in the active state
but without traffic, it uses the interframe time fill to maintain link synchronization. Time fill may also be used to avoid timeouts and to hold the
authority to transmit. Interframe time fill consists of a series of contiguous
flag sequences. They immediately follow the closing flag of the last frame
transmitted and follow the opening flag when the next frame is to be
transmitted.
Abort may be carried out during the transmission of a frame when the
transmitting side decides, before the end of a frame, to terminate that frame
and to force the receiver to discard the frame in question. The aborting is
accomplished by transmitting at least 7 but less than 15 consecutive 1s with
no zero insertion. The receipt of 7 contiguous 1s is interpreted as an abort.
An inalid frame is one that is not properly bounded by an opening and
closing flag or one that is too short in length i.e., less than 32 bits between
flags.. An aborted frame is an invalid frame. A receiving station ignores
invalid frames.
In HDLC, bits are transmitted low-order bit first a bit with a weighting of
.
2 . This order of bit transmission is valid for addresses, commands, responses,
and sequence numbers. The order of bit transmission for data contained
within the information field is application-dependent and is not specified.
The order of bit transmission for the FCS is most significant bit first.
ERROR RECOVERY
53
3.7 ERROR RECOVERY

Checkpoint Recovery. This is the basic error recovery procedure in HDLC
and in related bit-oriented protocols. Checkpoint error recovery uses poll and
final frames to effect recovery. This recovery procedure is used when neither
REJ nor SREJ reconvert is implemented. It also serves as a fall-back method
for both REJ and SREJ because it detects unsuccessful attempts at
REJrSREJ recovery such as an errored REJ frame.
A checkpoint can be in information I., supervisory S., or unnumbered
U. format frame with the P or F bit set to 1. Each poll frame forces a
final frame response. Another poll may be issued only when an error-free
final frame is received in response. In the case of a lost poll or final frame, a
poll may be issued after a system-defined timeout.
A checkpoint exchange is a pair of consecutive checkpointsthat is, a poll
frame and its responding final frame.
When a receiving station receives a frame with the PrF bit set to 1, it
starts retransmission of all unacknowledged I frames with sequence numbers
less than the send variable S. at the time the previous PrF frame was
transmitted. Retransmission of such frames starts with the lowest numbered
unacknowledged information I. frame, and the frames are retransmitted
sequentially. The term checkpoint retransmission is the retransmission of I
frames.
Duplicate transmissions are avoided during a checkpoint cycle by any of
the following conditions:
a. If an REJ frame with POLL or FINAL set in the OFF state has
already been received and acted upon
b. If a combined station receives and acts on any REJ except one with
the FINAL in the ON condition
c. Recovery is also stopped by an SREJ or any unnumbered format with
the POLLrFINAL in the ON condition
d. An acknowledgment of any frame with POLLrFINAL in the ON
condition before the next checkpoint occurs.
Retransmission of I frames can be specified by a station returning an RR
which has an N R . less than the N S . number the sending station is using.
Retransmission will begin with the indicated number.
Timeout Recovery. When the expected acknowledgment has not been received of a previously transmitted frame, timeout functions take over. Expiration of the timeout function causes initiation of timer recovery.
Suppose a receiving station, due to transmission error, does not receive or
receives and discards . a single I frame or the last I frames. in a sequence of
I frames. It will not detect an out-of-sequence exception and therefore will
not transmit SREJrREJ. Following the completion of a system specified
54
timeout period, the station that transmitted the unacknowledged I frames.

takes the appropriate timeout recovery action to determine the sequence
number at which retransmission must begin.
Generally, a recovery action results in the station detecting the timeout to
send a status request. The receiving station responds with a frame indicating
the value of its receive variable indicating the point at which retransmission
begins.
Reject Recovery. Reject recovery is used to initiate faster retransmission
response than is possible with checkpoint recovery following the detection of
a sequence error. Its use is typical on full-duplex circuits with their fast
turnaround.
Only one REJ exception condition may be established between two
stations at any given time. The REJ commandrresponse may only be transmitted once with a checkpoint cycle. The REJ exception condition is cleared
i.e., reset. upon receipt of an I frame with an N S . number equal to the
N R . number of the REJ command response.
Selective Reject Recovery. Selective reject recovery provides more efficient
error recovery by requesting the retransmission of a single information frame
following the detection of a sequence error rather than requesting go-back-n
frames and repeating all subsequent frames. It simply invokes selective ARQ
and is used where only a single I frame is missing. This is determined on
receipt of the out-of-sequence N S . number. When an I frame sequence
error is detected, the SREJ is transmitted as soon as permissible. Only one
SREJ exception condition may be outstanding at a time between two given
stations. A sent SREJ exception condition is cleared when the requested I
frame is received.
3.8 OTHER STATION MODES

The specifications of response modes were the means of activating the
data-link relationship by proposing and agreeing to the rules to be used
between stations. This relationship is deactivated by the use of the DISC
command to disconnect a secondaryrcombined station. If the connection was
over switched circuits, this can create a physical disconnection. If the station
is still physically connected, it is either in normal disconnected mode NDM.
or asynchronous disconnected mode ADM. based on agreed upon procedures for a network.
A station in disconnected mode has a limited set of possible actions.
All commands received without POLL are simply discarded. Any mode
setting command, such as SARM, is responded to in the normal manner.
Any commands that are not mode setting and have a POLL will have either
DM, RIM, or XID as the responses. These have been described in Section
3.5.3.
REFERENCES
55
The final station, the initialization state, uses the set initialization mode
SIM. command, defined in Section 3.5.1, to provide a means for a secondary
station to undergo a system-defined procedure. The procedure will vary
according to the type of equipment and network parameters in a given
implementation. It is intended to place a secondary station in a known state
where it can proceed with data-link control procedures.
There is an unnumbered POLL UP. that is an optional polling mechanism for a primary station. It can be used with a single secondary station or
as a group poll. The group poll procedure is not specified. When the POLL
bit is set, the secondary station must respond with at least one usually RR if
no I frames are ready. frame with FINAL bit set in the last frame. If the
POLL bit is off, the secondary may respond if it has UI, I, unacknowledged
frames, or a change in status to report. The final bit is off in all frames.
3.9 SDLC VARIATIONS WITH HDLC

IBMs SDLC synchronous data link control. variations with HDLC are as
follows:
1. It has no balanced mode of operation.
2. The SDLC loop operation is not an HDLC option.
3. It uses different names for its command and response modes.
Chapter 3 is based on Refs. 16.
REFERENCES
1. High Leel Data Link Control (HDLC) ProceduresFrame Structure, ISO 3309,
International Standards Organization, Geneva, 1976.
2. High Leel Data Link Control ProceduresConsolidation of Elements of Procedures,
ISO 4335, International Standards Organization, Geneva, 1977.
3. High Leel Data Link Control ProceduresConsolidation of Classes of Procedures,
ISO 7809, International Standards Organization, Geneva, 1978.
4. John McConnell, Internetworking & Adanced Protocols, Network Technologies
Group, Inc., Boulder, CO, 1985.
5. Roger L. Freeman, Reference Manual for Telecommunication Engineering, 3rd
edition, John Wiley & Sons, New York, 2001.
6. William Stallings, Handbook of Computer Communication Standards, Vol. 1,
Macmillan, New York, 1987.
7. Code Independent Error Control System, CCITT Rec. V.41, Fascicle VIII.I, IXth
Plenary Assembly, Melbourne, 1988.
4
DATA NETWORK OPERATIONS
4.1 CHAPTER OBJECTIVE
Data communications means to communicate data. The objective of this
chapter is to discuss how we can carry out this function effectively. At the
outset, we assume that there are two terminals in a data connectivity, a
near end of the circuit and a far end. We could call this one-to-one pairing.
There is also another possibility, a one-to-many connectivity, sometimes
called broadcast. For this discussion, data connectivities are on a one-to-one
basis.
The data to be interchanged consists of 8-bit sequences* whose meanings
are understood at both ends of the connection.
This chapter, then, contains a general overview of data operations or the
effective interchange of data between two stations. The word effectie implies
that data messages will be delivered to the destination users. error-free. The
reader should appreciate that no data circuit is completely error-free. Thus
the subject of data integrity is treated with considerable detail.
The final section gives an overview of data switching, along with the
switching alternatives available to the network designer.
4.2 GENERAL REQUIREMENTS FOR THE INTERCHANGE OF DATA

To efficiently deliver a data message from source to destination, a number of
conditions must be met or requirements must be satisfied:
1. A pathway needs to exist between the two ends. The pathway may or
may not pass through one or more relay nodes. The pathway may or
may not be shared by other users. We assume that only one pathway
exists.
*Called octets or bytes.
57
58
2. A method to access the pathway must exist.

3. A method must be incorporated in the data message to direct the
message to the intended far-end user. The format of the data message
must be such that its content is useful to the intended recipient.
4. Data urgency is an issue.
5. There must be some means to assure data integrity at the distant end.
Data network operations involve, as a minimum, these five issues and
requirements.
4.3 DISCUSSION OF ISSUES AND REQUIREMENTS

4.3.1 Pathway(s) Exist(s)
If a data connectivity is desired to far end-user Y, then a circuit must exist
between near-end local user X and far-end user Y. The circuit may be real or
virtual. If we are dealing with a LAN, there is a circuit available. However, it
is shared by other users on the LAN. The assumption is, of course, that we
are dealing with connection-oriented service.
When dealing with a WAN, the alternatives available are many. Some of
these alternatives were covered in Chapter 1. We might lease a T1DS1. or
E1 from the local PSTN, or from some other entity providing equivalent
service. Traffic volume may be so great that a private network is warranted.
If traffic volume is not so great, a fractional T1DS1. or E1 may be
considered.
A T1DS1. carries 24 time-division circuits. Each circuit can carry a
full-duplex telephone call or 56r64 kbps of full-duplex data. This is called a
DS0. If we view that circuit from a users perspective, it only belongs to that
user 1r24th of the time. The other 23r24ths of the time, other users occupy
the circuit. Thus we call this a irtual circuit. A real circuit is one where the
physical facilities or an equivalent frequency slot are assigned for use or to a
call 100% of the time.
Another option is to lease bit rate capacity from other carriers or quasicarriers. A quasi-carrier may be the local power company that has extensive
telecommunication facilities or a cellular radio company. There are any number of other alternatives.
One such possibility is signing a service level agreement SLA. for a.
frame relay service Chapter 12. between the points of interest, b. X.25
service Chapter 11., or c. ATM service with TCPrIP Chapters 16 and 11..
Providers of such services would include a. the local telephone company
LEC, or local exchange carrier. or b. interexchange carriers such as
AT & T, BT, Telefonica Spain., Cable and Wireless, MCI, Sprint, and so
forth.
DISCUSSION OF ISSUES AND REQUIREMENTS
59
Another possibility is to set up a VSAT very small aperture terminal.

network. This is a private-line geosatellite network used for interchange of
data traffic.
4.3.2 Access to the Pathway
In nearly all situations, the selected circuit pathway . is not connected to our
data set 100% of the time. Other data stations in the facility may want access
to that same circuit. The media access control MAC. carried out this
function with LANs. In the case of a WAN, access will be determined by OSI
layers 1 and 2. If the WAN is ATM-based asynchronous transfer mode; see
Chapter 16., other access methods will be employed. In either case OSI or
ATM., the physical layer interface must be available. Physical layer interface
is discussed more in detail in Section 6.4. These access functions deal with
protocols.
4.3.3 Directing the Data Message to Its Intended User(s)
This function is carried out by addressing. Addressing information is carried
in the link layer and network layer address fields in messagerframerpacket
headers. In the case of LANs, there are MAC addresses and LLC logical
link control; see Chapter 9. addresses. The LLC address identifies the final
recipients s. of the message.
For a WAN, Section 3.4 Chapter 3. discusses the HDLC address field.
This, of course, is link layer addressing. The TCPrIP protocol family Section
11.4. covers an entire realm of network layer addressing. Frame relay and
ATM each has its own particular form of addressing. As described in
Chapter 2, we can appreciate that addressing is yet another important
function of layer 2 and layer 3 protocols.
4.3.4 Data Urgency
When we ship an item by some agency such as Federal Express, you would be
asked when you want it to arrive at the destination: tomorrow morning, twoday service, five-working-days delivery, and so forth. There is, in essence, the
same situation when dealing with the delivery of data.
The most economic means of shipping data is probably by the Postal
Service or United Parcel Service. The data could be stored on a floppy disk
or on a CD, which is very easy to ship. Delivery to the desired recipient is
almost assured. It certainly is cost-effective, if there is not much urgency.
Credit card verification is an example of high-urgency data. We may want to
consider using Federal Express for long payroll files. The difference in
urgency is dramatic.
Consider the situation from still another angle. There is a high investment
in a data network. The more data we can push through that network, the
60
better is the return on our investment. This brings up the term latency. This
is the time it takes a data message to arrive at its destination from its origin.
Based on averages, the time it takes a data message to cross the North
American continent using packet transmission based on a popular protocol is
about 200 ms; when employing frame relay for the same message, it takes
about 20 ms.
There are several components in the latency equation:
Propagation time on WANs, this will be the largest contributor.

Time in queue when switching is involved
Processing time
Circuit quality noisy circuits will require message repeats on errored
messages.
Collision resolution on Ethernet LANs
Circuit setup for virtual circuits or for physical circuits
The longer a message sits in queue or is being serviced by repeats or is

delayed for other reasons, the less efficient the circuit becomes. This is
because of the time lost when it could be used by other traffic. Both ATM
and frame relay are low-latency systems.
4.3.5 Maintenance of Data Integrity
Ideally a data messagerframerpacket should arrive at its destination s. as an
identical replica of the message transmitted. The following is a listing of
features and parameters of a data message that may be checked to assure
integrity:
The messagerframe does not have an integral number of octets bytes.

The messagerframe has one or more bit errors
Lost frames
Out-of-sequence frames
Incomplete frames e.g., fields are missing or incomplete, noncompliance with minimum number of octets in information field, header errors,
etc..
In most situations where a frame or message is corrupted, the framermessage is discarded. In the some cases the originator is informed of the nondelivery. However, in most cases, where feasible, a repeat of the corrupted
frame is requested.
The most common form of a corrupted frame is one in which one or more
errors have occurred. Errors are detected by the receive-end through the
processing the FCS field see Figure 3.2.. A certain known bit sequence is
ERROR DETECTION SCHEMES
61
expected in the FCS if the frame is error-free. If some other sequence

appears, the frame contains one or more bit errors.
Error performance of data circuits lies at the very heart of quality of
service of a data network. The reader should consider the impact of just one
bit in error in a data file. The goal is an error-free system. Remember this
point: No system is error-free. Error performance can be improved such that
we can approach an error-free condition. In the following sections, we cover
some commonly used methods of error detection and error correction.
4.4 ERROR DETECTION SCHEMES

4.4.1 Parity Checks
One technique used in early data transmission systems was a simple parity
check. A source code widely employed in data communications is the ASCII
code* see Section 5.3.. It is a 7-bit code. There are two possible regimes for
error detection: odd parity and een parity. For parity checking, we count only
the 1s in a bit sequence. In this case the sequence is 7 bits long. An eighth
bit is added as a parity bit complying with the convention of odd parity or
even parity.
Lets assume now that a particular data circuit is operating under an odd
parity rule and that the information sequence is 7 bits long. Consider the
following examples:
Example 1. 0101001. There are three 1s, and 3 is an odd number; thus we
have complied with the convention. Add a 0 for the eighth bit.
Example 2. 1101100. There are four 1s, an even number. To comply with the
odd-parity convention, the eighth bit will be a 1 to make the number of 1s an
odd number.
This is called ertical redundancy checking VRC.. Its error detection
capability is weak. In Example 1 a certain noise burst occurred where a 1
changed to a 0 and a 0 changed to a 1; the convention would be complied
with, yet there were two errors that went undetected.
The other type of parity checking is called longitudinal redundancy checking
LRC.. Generally, LRC was used in addition to VRC to provide a more
robust error detection capability. We build a matrix as shown in Figure 4.1.
The character bits go from left to right, where the 8th bit is the conventional
parity bit derived from VRC. There are seven rows for the seven 7-bit
characters, each with an 8th bit appended using odd parity for the row. Seven
characters are represented, thus seven rows. An eighth row is derived on the
*ASCII stands for American Standard Code for Information Interchange.
62
Figure 4.1. Sample matrix employing VRC and LRC together. The convention here is odd
parity.
bottom from longitudinal redundancy checking LRC., again using the oddparity checking on each column. We could transmit the 8th row to the distant
end where a similar matrix could be constructed to verify that the same 8-bit
sequence is derived from LRC. Of course the VRC would also be checked.
Using LRC and VRC together reduces the residual error rate from 1 in 100
to 1 in 10,000 Ref. 1..
It can be demonstrated that VRCrLRC used in tandem leaves a fair
amount of residual errors. The next section discusses cyclic redundancy check,
which has excellent residual error performance.
4.4.2 Cyclic Redundancy Check (CRC)
A more powerful method of error detection involves the use of CRC. It is
employed with HDLC Chapter 3.. It is used with messages that are data
blocks, frames, or packets. Such a data message can be simplistically represented as shown in Figure 4.2.
The bit sequence in a data frame can be represented by a polynomial.
Lets call this the message polynomial, G X .. G X . is multiplied by X 16 ,
which corresponds to shifting the message G X . 16 places to provide space
for the 16 bits in the FCS. We divide the resulting product w X 16 GX .x by a
known generating polynomial. The result of this division is a quotient and a
remainder. The remainder is 16 bits long, and it is placed in the FCS slot
Figure 4.2.. At the receive side, essentially the same operation is repeated
and a remainder is generated. If the remainder generated by the receiver is
the same sequence as found in the FCS slot of the received frame, the frame
is error-free. If it is not the same, there was at least one bit error in the
received frame.
ERROR DETECTION SCHEMES
63
Figure 4.2. Coverage of CRC in a data-link frame (based on HDLCrLAPB). Note that there
are 16 bits in the FCS field (Ref. 2).
Note: The description of CRC operation given above is valid for HDLC
and its related link layer protocols such as ADCCP and LAPB. In fact, for
error-free operation, the remainder is the sequence 0001110100001111. Traditional CRC systems that do not start with a preset value in the shift register
as HDLC does. have an all-zero remainder for error-free operation.
There are two generator polynomials accepted by the industry:
ANSI CRC16.:
CCITT ITU-T.:
X 16 q X 15 q X 2 q 1
X 16 q X 12 q X 5 q 1
ANSI CRC-16. is applied to synchronous systems that use 8-bit characters. The FCS accumulation is 16 bits. It provides error detection of error
bursts up to 16 bits in length. Additionally, more than 99% of error bursts
greater than 16 bits can be detected.
The CCITT ITU-T. standard generator polynomial also operates with
8-bit characters in synchronous bit streams. The FCS accumulation is 16 bits.
It provides error detection of bursts up to 16 bits in length. In addition, more
than 99% of error bursts greater than 16 bits can be detected. A typical shift
register used with this CRC is shown in Figure 4.3.
Figure 4.3. Shift register used with CRC-CCITT. MSB, most significant bit; LSB, least
significant bit. [, exclusive-OR gate modulo-2 addition.
64
To derive an algebraic polynomial from a string of bits, review the following: A polynomial is usually represented algebraically by a string of terms
in powers of X such as X n q qX 3 q X 2 q X q X 0 or 1.. In binary
form, a 1 is placed in each position that contains a term; absence of a term is
indicated by a 0. The convention used in the following presentation
is to place the X 0 bit at the right. For example, if a polynomial is given as
X 4 q X q 1, its binary representation is 10011 3rd- and 2nd-degree terms
are not present .. The number 1 in the polynomial represents the X 0 power
that was artificially added.
Suppose a message polynomial was 110011. This would be represented by
X 5 q X 4 q X q X 0 or X 5 q X 4 q X q 1..
4.5 ERROR CORRECTION SCHEMES

4.5.1 Automatic Repeat Request (ARQ)
The term ARQ automatic repeat request. derives from the days of automatic telegraphy. ARQ is widely used to request the sending end to repeat a
frame or several frames that have errors. It requires full-duplex or half-duplex
operation. There are three variants of ARQ:
Stop-and-wait ARQ
Continuous, sometimes called selective, ARQ
Go-back-n ARQ
Stop-and-Wait ARQ. Stop-and-wait ARQ is the most straightforward to implement and the least complex. However, it is inefficient. With stop-and-wait
ARQ, a frame is transmitted to the distant end. At the completion of sending
the frame, transmission ceases stop.. The receiving terminal processes the
frame and then examines the FCS field. If it is the expected sequence for
error-free operation, it sends an acknowledgment ACK. to the sending end,
which then promptly transmits the next frame. However, if the receiving end
does not encounter the expected sequence in the FCS field as a result of its
processing, it sends a negative acknowledgment NACK. to the transmitting
end. The transmitter then resends that frame.
Stop-and-wait ARQ is inefficient primarily due to the stop time. This is
the wasted period of time waiting for the receive end to process the frame
and to send the ACK or NACK. It can be particularly wasteful of circuit time
on connectivities with long propagation delays such as geostationary satellite
circuits, and there are long frames involved. It is, however, ideal for halfduplex operation.
Selective or Continuous ARQ. Selective or continuous ARQ eliminates the
quiescent nonproductive time of waiting and then sending the appropriate
ERROR CORRECTION SCHEMES
65
signal back after each frame or packet.. With this type of ARQ, the transmit
end sends a continuous string of frames or packets., each with an identifying
number in the header of each frame packet.. When the receive end detects a
frame in error, it requests the transmit end to repeat that frame. The request
that the receiver sends to the transmitter includes the identifying number
more commonly called sequence number . of the frame in error. The transmit
end retrieves that frame from its buffer storage and retransmits it, with the
appropriate identifying number, as another frame in the continuous string.
The receive end receives the repeated frame out of sequence order., reruns
the processing procedure to ensure error-free reception, and then places the
frame in proper numerical sequence for delivery to the end-user facility.
Obviously, selective or continuous ARQ is more efficient regarding circuit
usage. It requires full-duplex operation. Such operation uses bookkeeping at
both ends of the link. It requires more buffer storage at both ends, but the
cost of memory continues to erode.
Go-Back-n ARQ. Go-back-n ARQ is probably the most popular of the three
varieties of ARQ. It, too, permits continuous frame transmission. As with
continuous ARQ, when an errored frame is encountered, the receiving end
informs the far-end transmitter by the return channel. The errored frame is
identified by sequence number. The transmitter repeats the frame that was in
error and all subsequent frames, even though they have been received before.
This approach alleviates the problem at the receive end of inserting the
errored block in its proper sequence. With this type of ARQ, buffers must be
capable of storing two or three frames, depending on the propagation delay
of the circuit and the frame length.
The reader should turn back to Section 3.4.1 for a description of how
ARQ is implemented in HDLC. It will be noted that HDLC uses a go-back-n
strategy with its window of frames and method of acknowledgment.
4.5.2 Forward Error Correction (FEC)
FEC is a method of error correction, or, if you will, a method of error
reduction at the far-end receiver. No return channel is required. The concept
is based on adding systematic redundancy at the transmit end of a link such
that errors caused by the transmission medium can be corrected at the
receiver by means of a decoding algorithm. The amount of redundancy is
dependent on the type of code selected and the level of error correction
capability desired. Unless an additional processing step is included, FEC will
only correct random errors i.e., not burst errors..
Forward error correction uses channel coding whereas the coding that
many readers are familiar with Section 5.3ASCII code, EBCDIC, etc.. is
source coding. The channel coding used in FEC can be broken down into two
broad categories of codes: block codes and conolutional codes. Modulo-2
addition is widely used in both error detection and error correction. This is
66
Figure 4.4. Simplified diagram of FEC operation.
the same as binary addition but without carries. It is just an exclusive-OR

operation Figure 4.3..
Example
1111
q1010
0101
A measure of error detection and correction capability of a specific code is
given by the Hamming distance. The distance is the minimum number of
digits in which two encoded words or bit sequences differ. Consider these
binary sequences: 1010110 and 1001010. The Hamming distance is 3. If the
Hamming distance between any two bit-sequences of a code is e q 1, then it
is possible to detect the presence of e or fewer errors in a received sequence
of binary words.
If the Hamming distance between any two words of a code is equal to
2 e q 1, then it is possible to correct any e or fewer errors occurring in a
received binary sequence. A code with a minimum Hamming distance of 4
can correct a single error and detect two symbols in error Ref. 3..
Figure 4.4 illustrates a digital communication system with FEC. The binary
data source generates information bits at R S bits per second. The information bits input to an FEC encoder with a code rate R. The output of the
encoder is a serial bit stream with a symbol rate of R C symbols per second.
The code rate R is related to the encoder input and output by the following
expression:
R s R S rR C
The code rate R is the ratio of the number of information bits to the number
of encoded symbols for binary transmission.
Example 1. If the code rate were 12 and the output of the data source were
14,000 bits per second, what would be the output of the encoder in symbols
per second? The symbol and bit are synonymous.. Answer s 28,000 symbolsrsecond.
R s 14,000r28,000 s 1r2
Example 2. The output of the encoder is 36,000 symbols per second and the
input to the encoder is 24,000 bits per second, what is the code rate R .?
24,000r36,000 s 2r3
DATA SWITCHING
67
The encoder output sequence is then transmitted to the line via a modulator e.g., modem. or modulator-transmitter. At the receive end, demodulation is performed to recover the coded symbols at R C symbols per second.
These symbols are then fed to a decoder with an output of R S bits per
second, which are then delivered to the data destination.
The major advantages of an FEC system are as follows:
No feedback channel is required as with an ARQ system.

There is constant information throughput i.e., no stop-and-wait gaps..
Decoding delay is generally small and constant.
It is ideal for use on marginal transmission channels.
There are two basic disadvantages with an FEC system. To effect FEC for
a fixed information bit rate, the bandwidth must be increased to accommodate the redundant symbols. For example, a rate 12 FEC system would
require twice the bandwidth of its noncoded counterpart. However, trellis
coding, described later, ameliorates some of the bandwidth expansion requirement. There is also the cost and complexity of the added encoder and
decoder. Nevertheless, with present VLSIrVHSIC technology, an encoder
and decoder can be mounted on a chip, and they are readily available
off-the-shelf for some of the more popular encodingrdecoding configurations.
4.6 DATA SWITCHING

There are three approaches to data switching:
1. Circuit switching
2. Message switching store-and-forward .
3. Packet switching
Plain old telephone service, affectionately called POTS, uses circuit switching. A data switch can operate in a similar manner. It will set up a circuit
between origin and destination based on the message header. The circuit may
be a physical circuit if the network is analog, or it may be a virtual circuit if
the network is digital. A virtual circuit is a time slot or a series of time slots,
possibly a different time slot on each link as a irtual connection is set up
end-to-end. The virtual connection remains intact during the duration of a
call or message, in the case of data circuit switching.
Message switching, often called store-and-forward switching, accepts data
messages from originators, stores the messages, and then forwards each
message to the next node or destination when circuits become available. We
wish to make two points when comparing circuit and message switching.
68
Circuit switching provides end-to-end connectivity in near real time. Message

switching does not. There is usually some delay as the message makes its way
through the network to its destination. The second point deals with efficient
use of expensive transmission links. A well-designed message-switching system keeps a fairly uniform traffic load throughout the working day and even
into the night on each link. Circuit-switching systems, such as encountered
with the public switched telecommunication network PSTN., are designed
for busy-hour loading, and the system tends to loaf when off the busy hour.
Message switching, therefore, makes much more efficient use of expensive
transmission links than does circuit switching. This improved efficiency is at
the expense of some delay in service.
Telex is a good example of message switching. Another example is the
older torn-tape systems. Such systems were labor-intensive and rather
primitive by todays standards. With a torn-tape switching center, messages
arriving from originators or relay nodes used paper tape to copy the messages, usually with the five-element ITA 2 code Chapter 5. at somewhere
between 50 and 75 bps. Headers were printed out laterally along the tape
holes. Operators at the torn-tape center would read the header, tear off the
tape, and then insert that same tape at the tape reader of the indicated
outgoing circuit. Tape would often overflow onto the floor at either the
receive terminal, when operators were very busy, or at the read head, when
circuits were busy. Higher-priority messages moved up to the head of the
queue. Ive seen clothespins and paper clips used to piece tapes together in
such queues.
Store-and-forward systems today use magnetic tape or other electronic
storage rather than paper tape. A processor arranges messages by priority in
the correct queue. Message processors read precedence and address information from the header and automatically pass the data traffic to the
necessary outgoing circuits. In older systems, the floor was a storage bin of
almost infinite size. In modern systems, a concern remains on memory
overflow during peak traffic conditions.
Packet switching utilizes some of the advantages of message switching and
some of circuit switching. It mitigates some of the disadvantages of both.
A data packet is a comparatively short block of message data of fixed
length. Each packet, of course, has its own header and FCS at the end of the
packet. These packets may be sent over diverse routes to their eventual
destination, and each packet is governed by an ARQ-type error-correction
protocol. Because these packets are sent on diverse routes, they may not
arrive at the far-end receiving node in proper sequential order. Thus the
far-end node must have the capability of storing the packets and rearranging
them in proper sequential order. The destination node then reformats the
message as it was sent by the originator and forwards it to the final-destination end user.
A packet-switching network can show considerably greater efficiency when
compared to circuit- and message-switching networks. There is a one caveat,
DATA SWITCHING
69
however. For efficient packet-switching network operation, a multiplicity of

paths three or more. must exist from the originating local node to the
destination local switching node. Well-designed packet-switching systems can
reduce delivery delays when compared to conventional message switching.
Expensive transmission facilities tend to be more uniformly loaded when
compared with networks based on other data-switching methods. Adaptive
routing becomes possible where a path between nodes has not been selected
at the outset. Nodes are kept current on load conditions, and alternative-route
traffic is directed to more lightly loaded trucks. Delivery delay is also reduced
when compared to message switching because a packet switch starts forwarding packets before receipt of all constituent packets making up a message.
Again, path selection is a dynamic function of real-time conditions of the
network. Packets advancing through the network can bypass trunks and
nodes that are congested or that have failed.
A packet switch is a message processor. Packets are forwarded over
optimum routes based on route condition, delay, and congestion. It provides
error control and notifies the originator of packet receipt at the destination.
Packet switching approaches the intent of real-time switching. Of course,
there is no real connection from originator to destination. The connection
may be classified as one form of a irtual connection.
Table 4.1 summarizes the advantages and disadvantages of these three
methods of data switching.
4.6.1 Philosophy of Data Signaling Versus Telephone Signaling

Data are delivered to their distant-end destination either by connectionoriented service or by connectionless service. Conventional telephony operates in a connection-oriented service by circuit switching described above.
During call setup a circuit is established from the calling user to the called
user. A signaling sequence is sent just once during the circuit setup period by
means of address signaling. The circuit is held in its connected condition by
means of supervisory signaling until one or the other party goes on-hook
i.e., hangs up..
In fact, frame relay and ATM have entire separate signaling regimes for
setting up a permanent virtual circuit PVC. and to maintain health and
welfare of the circuit when it is quiescent.
When data are to be communicated, there are several ways to effect the
connectivity for the communication:
1. A PVC is set up for the duration of the data call.
2. A PVC exists between each end of the call and there is no setup or
clear-down of the call.
3. A switched virtual circuit SVC. is used.
70
TABLE 4.1 Summary of Data Switching Methods

Switching Method
Circuit switching
Advantages
Mature technology
Near real-time
connectivity
Excellent for inquiry
and response
Leased service
attractive
Message switching
Efficient trunk
utilization
Cost-effective for
low-volume leased
service
Disadvantages
High cost of switch
Lower system utilization,
particularly link
utilization
Privately owned service

can only be justified
with high traffic
volume
Delivery delay may be a
problem
Not viable for inquiry and
response
Survivability
problematical
Requires large storage
buffers
Packet switching
Efficiency
Approaches near realtime connectivity
Highly reliable,
survivable
Multiple route and node

network expensive
Processing intensive
Large traffic volume
justifies private
ownership
Low traffic volume

attractive for
leased service
4. Connectionless service is provided. Here data are communicated by

packets, called datagrams in many standards. The circuits involved are
usually SVCs.
5. A physical circuit is set up or in place.
A major difference between data communication and conventional telephony communication is that data communication has a header on each and
every frame, packet, or ATM cell see Chapter 16.. In other words, address
signaling is repeated over and over again, whereas in telephony it is only
transmitted once to set up the call.
Frame relay Chapter 7. has made an attempt to reduce the burden of the
large header, yet its data-link connection identifier DLCI. can be thought of
in the light of a conventional header, directing the frame through the network to its proper destination. Comparatively, IP uses very long headers.
REFERENCES
71
Data frame headers, along with the trailing FCS field, are overhead. The
info field is the payload. It carries the revenue-bearing portion of a frame.
Surely, some of us have gathered that even the info field carries overhead
information. This is embedded information used to service the message or
frame. The information derives from the upper OSI layers. A telephone call
carries no such overhead except for the simple supervisory signaling. From
this perspective, conventional telephony is more efficient than its data transmission equivalent.
4.6.2 Smart Bridges, Routers, and Switching Hubs
Each of these devices carries out some form of switching. The general application is with LANs. Some will say that each device carries out a routing
function. The basis of routing is by means of a routing table which may or
may not be dynamically updated. Bridges, routers, and hubs are discussed in
Sections 10.2, 10.3, and 10.4, respectively.
REFERENCES
1. John E. McNamara, Technical Aspects of Data Communication, 2nd ed., Digital
Equipment Corp., Bedford, MA, 1982.
2. High Leel Data Link Control (HDLC) Frame Structure, ISO 3309, International
Standards Organization, Geneva, 1976.
3. Reference Data for Radio Engineers, 5th ed., ITT Howard W. Sams, Indianapolis,
1975.
5
DATA TRANSMISSION I
5.1 ELECTRICAL COMMUNICATION OF INFORMATION
Telecommunication transmission involves electrical communication of information. We communicate voice, image, and data. Our concern here is the
electrical communication of data.
By electrical communication of data, we mean the formation of an electrical signal that connotes information which can be interpreted at some distant
receiver with minimal or no ambiguity. One way of doing this is to impress a
voltage on an electrical circuit. The on voltage can signal something, and
the off voltage signals something else. Think of turning a light on and
turning the light off. The on condition could represent a binary 1 and the
off condition can represent a binary 0. We now have related the very
essence of expressing information electrically.
5.2 THE BIT AND BINARY TRANSMISSION OF INFORMATION

What was just described in the paragraph above was the bit. Some say the bit
is an acronym for binary digit. One definition of the bit is a contraction of
the term binary digit; a unit of information represented by either a one or a
zero Ref. 1..
The bit, therefore, has one of two possible states at any moment in time.
In our everyday life we can think of many two-state situations. A light is on or
it is off; a telephone line is idle or it is busy; a person is home or is out, a
person is asleep or awake, just to name a very few.
On the one hand, we surmise that a two-state information system does not
tell us very much; on the other we can find it very useful. For example, in
telephone transmission and switching, it is imperative that we know whether
a line is idle or busy. Nearly all telephone systems are built around this
concept. It is called superisory signaling or line signaling. Just one of two
states can exist: idle or busy.
73
74
DATA TRANSMISSION I
In another important system used in telecommunications, all we wish to

know is whether a part, card, assembly, unit, or device is working or has
failed. Again, just one of two states. Automatic trouble shooting of very
sophisticated electronic equipment is based on this premise.
The earliest electrical representation of a bit followed the onroff concept.
A 1 was represented by a positive voltage and a 0 was represented by no
voltage 0 volts., as shown below:
For the requirements of data communications, a single bit with its two
possible states provides little information. Suppose now that two contiguous
bits are transmitted rather than just one bit. We now have a four-state system
or four possible bit combinations as follows:
00, 01, 10, and 11.
The electrical representation of these are shown below:
Let us carry this thinking forward one more step. This time each piece of
information we wish to transmit contains three consecutive bits. Each of the
bits, of course, can be a 1 or a 0. If we write down all the possibilities of 1s
and 0s in a three-bit sequence, we find there are eight as follows: 000, 001,
010, 011, 100, 101, 110, and 111. Each bit sequence could represent a letter of
the alphabet. However, we would only cover eight letters such as A, B, C, D,
THE BIT AND BINARY TRANSMISSION OF INFORMATION
75
E, F, G, and H. These bit sequences are shown with their electrical equivalents as follows:
As we said, if we were to assign letters of the alphabet to these, we would

only reach the letter H. Certainly, if we wish to develop a means of
transmitting alphanumeric information using bits, we would want to cover the
entire 26 letters of the alphabet. The digits 0 through 9 should also be
included as well as punctuation. There are also nonprinting characters such
as space bar., carriage return, and line feed word-wrap .. There are many
graphic symbols which are useful to express such as @, , $, %, &, *, and
., s , q, and _.
It follows, therefore, that we must enrich the code.
Some readers may have caught the trend. Every time a binary code is
extended 1 bit, the total number of different possibilities doubles. We then
would expect that if we moved from a 3-bit code to a 4-bit code, we would
have 16 possibilities rather than only 8. Now consider the following series:
One-bit code:
Two-bit code:
Three-bit code:
Four-bit code:
Five-bit code:
Six-bit code:
Seven-bit code:
2 1 s 2, or two distinct possibilities

2 2 s 4, or four distinct possibilities
2 3 s 8, or eight distinct possibilities
2 4 s 16, or sixteen distinct possibilities
2 5 s 32, or thirty-two distinct possibilities
2 6 s 64, or sixty-four distinct possibilities
2 7 s 128, or one hundred twenty-eight possibilities
Then if n is the number of bits per character in a binary code set, the
number of different characters, graphic symbols, and nonprinting characters
control functions. is simply 2 n.
Today we commonly use 5-, 6-, 7-, and 8-bit codes.
76
DATA TRANSMISSION I
5.3 BINARY CODES FOR DATA COMMUNICATION

One of the earliest codes was the Baudot code, which is a 5-bit code. Thus
there are 2 5, or 32, character possibilities. Now how can such a code
accommodate the 26-letter alphabet, 10 digits 0 through 9., punctuation, and
some nonprinting characters such as space bar. with a total of only 32
characters?
The approach used was the same as that of the typewriter. The code is
extended by using uppercase. Two additional nonprinting characters were
Figure 5.1. The ITA2 code.
BINARY CODES FOR DATA COMMUNICATION
77
required to do this: uppercase and lowercase. Thus the net gain was to
double the code possibilities minus the two nonprinting characters used to
achieve the code extension. This left a code that could accommodate 62 characters, which met the requirements of the day.
A somewhat modified Baudot code is still in use today. It is the ITA2
code. ITA stands for International Telegraph Alphabet. The code is employed chiefly on record traffic circuits, but has also been used for data
communications. Figure 5.1 shows the ITA2 code.
In the late 1950s, when data communications began to take hold, it was
found that the ITA2 code was not rich enough to support this emerging
industry. Several ad hoc codes were being fielded, mostly 6-bit codes. By now
we know that a 6-bit code provides 2 6 s 64 possibilities, or a 64-character
code. This was really not much better than the ITA2 code with its upperand lowercase characters.
Supported by some of the emerging giants in the data processing world,
the American National Standards Institute ANSI. developed a 7-bit code,
which today has worldwide acceptance. This is the American Standard Code
for Information Interchange, which goes by the acronym ASCII. The code
provides 128 distinct code possibilities 2 7 s 128., or 128 characters. ASCII is
illustrated in Figure 5.2. One should note that it provides a large number of
nonprinting characters, many of which may be used for data-link control
Figure 5.2. American Standard Code for Information Interchange (ASCII). See Table 5.1 for
definitions. (Basic source: Ref. 3; see also Ref. 4.)
78
DATA TRANSMISSION I
purposes, the forerunner of present-day link-layer protocols. Among these we

find EOT, end of text; SOH, start of heading; EOM, end of message; SOM,
start of message; and NACK, negative acknowledgment. Table 5.1 gives the
meanings of these control characters.
CCITT now ITU-T. followed suit with a similar 7-bit code previously
called International Alphabet No. 5 IA5.. It is now called International
Reference Alphabet, which is shown in Figure 5.3, the basic code table.
TABLE 5.1 Definitions for Figure 5.2: Control Characters and Format Effectors
NUL: All zero characters. A control character used for fill.
SOH (Start of heading): A communication control character used at the beginning of a
sequence of characters which constitute a heading. A heading contains address or routing
information. An STX character has the effect of terminating a heading.
STX (Start of text): A communication control character that terminates a heading and indicates
the start of text or the information field.
ETX (End of text): A communication control character used to terminate a sequence of
characters started with STX and transmitted as an entity.
EOT (End of transmission): A communication control character used to indicate the conclusion of transmission which may have contained one or more texts and associated headings.
ENQ (Enquiry): A communication control character used as a request for response from a
remote station. It may be used as a Who are you (WRU) to obtain identification, or may be
used to obtain station status, or both.
ACK (Acknowledge): A communication control character transmitted by a receiver as an
affirmative response to a sender.
BEL: A character for use when there is a need to call human attention. It may control alarm or
attention devices.
BS (Backspace): A format effector which controls the movement of the active position one
space backward.
HT (Horizontal tabulation): A format effector which controls the movement of the active
position forward to the next character position.
LF (Line feed): A format effector which controls the movement of the active position advancing
it to the corresponding position of the next line.
VT (Vertical tabulation): A format effector which controls the movement of the active position
to advance to the corresponding character position on the next predetermined line.
FF (Form feed): A format effector which controls the movement of the active position to its
corresponding character position on the next page or form.
CR (Carriage return): A format effector which controls the movement of the active position to
the first character position on the same line.
SO (Shift out): A control character indicating the code combination which follows shall be
interpreted as outside of the character set of the standard code table until a shift in (SI)
character(s) is reached. It is also used in conjunction with the (ESC) escape character.
SI (Shift in): A control character indicating that the code combinations which follow shall be
interpreted according to the standard code table (Figure 5.2).
DLE (Data-link escape): A communication control character which will change the meaning of
a limited number of contiguously following bit combinations. It is used exclusively to provide
supplementary control functions in data networks. DLE is usually terminated by a shift in (SI)
character(s).
DC1, DC2, DC3, DC4 (Device controls): Characters for the control of ancillary devices associated with data or telecommunication networks. They switch these devices on or off. DC4
is preferred for turning a device off.
NAK (Negative acknowledgment): A communication control character transmitted by a receiver as a negative response to a sender.
ELECTRICAL BIT DECISIONS
79
TABLE 5.1 Continued.

SYN (Synchronous idle): A communication control character used by a synchronous transmission system in the absence of any other character to provide a signal from which
synchronism may be achieved or retained.
ETB (End of transmission block): A communication control character used when block
transmission is employed. It indicates the end of a block of data.
CAN (Cancel): A control character used to indicate that the data preceding it is in error or is to
be disregarded.
EM (End of medium): A control character associated with the sent data which may be used to
identify the physical end of the medium, or the end of the used or wanted portion of
information recorded on a medium. It should be noted that the position of the character does
not necessarily correspond to the physical end of the medium.
SUB (Substitute): A character that may be substituted for a character which is determined to
be invalid or in error.
ESC (Escape): A control character intended to provide code extension (supplementary
characters) in general information interchange. The ESC character itself is a prefix affecting
the interpretation of a limited number of contiguously following characters. ESC is usually
terminated by an SI (shift in) character(s).
FS (File separator), GS (Group separator), RS (Record separator), and US (Unit separator):
These information separators may be used within data in optional fashion, except that their
hierarchical relationship shall be FS is the most inclusive, then GS, then RS, and US is the
least inclusive. (The content and length of a File, Group, Record, or Unit are not specified.)
DEL (Delete): This character is used primarily to erase or obliterate erroneous or unwanted characters in perforated tape. (In the strict sense DEL is not a control character.) DEL
characters may also be used for fill without affecting the meaning or information content of the
bit stream.
Note: SO, ESC, and DLE are all characters which can be used, at the discretion of the system
designer, to indicate the beginning of a sequence of digits having special significance.
SP (Space): A nonprinting character used to separate words or sequences. It is a format
effector that causes the active position to advance one character position.
Source: Ref. 3, also see Ref. 4.
CCITT Rec. T.50 Ref. 2. points out that positions 0r0 to 1r15 the first two
columns, which are blank in the figure. are reserved for control characters.
Note the control characters in the first two columns of Figure 5.2..
IBM went one step further and defined an 8-bit code. It is called
EBCDIC, which stands for Extended Binary Coded Decimal Interchange
Code. This code is shown in Figure 5.4. It will be noted that EBCDIC is very
rich in graphic characters. It should also be clear that some of the character
possibilities are not used i.e., left blank..
5.4 ELECTRICAL BIT DECISIONS
Electrical noise tends to corrupt a telecommunication system. Let us define
noise as unwanted disturbances superimposed upon a useful signal that
tend to obscure its information content Ref. 1.. Accept the fact that all
electrical communication systems have some sort of noise component superimposed upon a wanted signal. All other signal impairments will be disregarded for now.
80
DATA TRANSMISSION I
Figure 5.3. Basic code table for CCITT International Alphabet No. 5 (IA5), the International
Reference Alphabet (IRA). (From Table 4, page 11, CCITT Rec. T.50, Ref. 2.)
We find that binary digital receivers are very forgiving of noise. The
reason is that such a receiver is a device that must make a simple decision:
whether a pulse is a 1 or a 0.
In the case of the scenario described in Section 5.2, a positive pulse is a
1 and no pulse is a 0. Consider Figure 5.5, which shows the ideal
situation of a stream of alternating 1s and 0s. The signals have not been
corrupted by noise. Figure 5.6 shows the same signals corrupted by noise.
What is meant by corruption is really the likelihood that an incorrect decision
could be made. Conventionally, a binary digital receiver makes a decision at
ELECTRICAL BIT DECISIONS
81
Figure 5.4. The EBCDIC code.
Figure 5.5. A string of 0s and 1s uncorruptedideal signal.
mid-pulse. The decision will be made that it is a 1 if the voltage reaches at

least half-amplitude in the negative direction i.e., a negative polarity voltage.;
it is a 0 if the voltage has reached at least half-amplitude in the positive
direction i.e., a positive polarity voltage..
Suppose at the moment the decision is made, a noise spike is present that
exceeds half-amplitude in the negative direction during the time a 0 was
supposed to be received. Of course the decision circuit will confuse it for a
binary 1. The bit will be in error.
Several measures could be taken to improve the chances of making the
right decision. We could increase the negative and positive voltages of the
82
DATA TRANSMISSION I
Figure 5.6. The same signal as in Figure 5.5, but corrupted with noise, cable pair capacitance, and loss. The capacitance causes rounding, and the loss causes a reduction in
amplitude.
Figure 5.7. A simplified drawing of a binary digital regenerator.
pulses. We could set the decision level higher, say to 34 amplitude, or use
some other method of representing the 1 and the 0. There are many ways to
improve performance of binary transmission of data in the presence of
impairments such as noise. Often an apparent improvement will also have a
drawback. The fact that binary data receivers have to make a decision of one
of only two possibilities makes this type of transmission very forgiving in the
presence of impairments. However, we must remember that the forgiveness is
allowed only up to a certain point before performance suffers. That point will
vary, of course, depending on the level of performance that is set for the
system.
Another aspect of digital transmission is that the signal can be regenerated. A regenerator is simply a decision circuit and a clock. A regenerator
accepts a corrupted pulse at its input and clocks out a squared-up 1 or
0. This concept is shown in Figure 5.7.
5.5 ELECTRICAL REPRESENTATION OF BINARY DATA

5.5.1 Neutral and Polar Waveforms
In Section 5.2 we used a neutral type of electrical representation of 1s and
0s. In this case, one state is active and the other state is passive. In our
examples, a 1 was a positive voltage and a 0 was a no-voltage condition, a
ELECTRICAL REPRESENTATION OF BINARY DATA
83
Figure 5.8. Neutral and polar waveforms compared.
passive state. A passive state can be easily corrupted by a noise spike such
that a receiver would mistakenly consider it an active state if the noise level
were of sufficient amplitude. To help avoid such situations, a polar* representation of a binary 1 and 0 was devised, where both states are active.
Neutral and polar representations of binary data are compared in Figure 5.8.
In the future we will call these neutral and polar waveforms. These states are
a negative voltage and a positive voltage. To be consistent with Section 5.2,
we chose to call the positive pulse a 1 and a negative pulse a 0.
It should be noted that in modern Layer 1 standards, the states are
reversed: a binary 1 is a negative voltage, and a 0 is a positive voltage
typically as set out in EIArTIA-232E Ref. 5..
At times, while working with binary bit streams we can, either by accident
or on purpose, reverse the sense of the bits. This simply means reversing the
bits* polarity, where the negative battery and positive battery are reversed.
You might say where the 1s become 0s and the 0s become 1s. This would be
highly undesirable for operational circuits, but we may want to reverse the
sense to carry out some form of testing or troubleshooting.
5.5.2 Waveforms and Line Codes
There are yet other ways of representing 1s and 0s electrically, each waveform with its own advantages and disadvantages. Generally, binary data
waveforms can be divided into two groups: non-return-to-zero NRZ. and
return-to-zero RZ.. These are shown and compared in Figure 5.9. Among
the advantages and disadvantages are simplicity and the ability to extract
timing, but at the expense of increasing the number of transitions per average
unit time. This, in essence, increases the frequency of the waveform, which,
under certain circumstances, may be highly undesirable typically on a
fiber-optics link..
*Polar is called double-current by CCITTrITU-T.
84
DATA TRANSMISSION I
Figure 5.9. Some waveforms (line codes) used in binary digital transmission. Note that
Manchester coding is widely used on baseband LANs.
5.6 BINARY CONVENTIONS

Another aspect of binary transmission is ambiguity. If we are to set up a data
circuit, both ends of the circuit must agree on which of the two possible
states is a 1 and which is a 0. For example, there is no reason why we
cannot establish a 1 as a positive-going voltage and a 0as a negative-going
voltage see reversing the sense in Section 5.4..
CCITT ITU-T. has established some conventions to remove the ambiguity of which state is a 1 and which is a 0. This may be found in CCITT Rec.
V.1. We have embellished upon the information in Rec. V.1 in Table 5.2.
This table is called a table of mark-space conention. The terms mark and
space derive from the days of telegraph to distinguish between the active and
BIT-PARALLEL AND BIT-SERIAL
85
TABLE 5.2 Table of Mark-Space Convention

Binary 1
Mark or marking
Perforation (paper tape)
Negative voltage a
Condition Z
Tone on (amplitude modulation)
Low frequency (frequency shift keying [FSK])
Opposite to the reference phase
No phase inversion
(Differential two-phase modulation)
Binary 0
Space or spacing
No perforation
Positive voltage
Condition A
Tone off
High frequency
Reference phase (phase modulation)
Inversion of the phase
CCITT Recs. V.10 and V.11 (Refs. 7 and 8, respectively) and EIA r TIA-232 (Ref. 5 ).
Source: CCITT Rec. V.1 (Ref. 6 ). Courtesy of ITU-CCITT.
passive state. To confuse matters, the mark is associated with the 1 and the
space is associated with the 0; we call them synonymous, meaning that a
binary 1 and a mark have the same meaning. Contrary to our discussion in
Section 5.2, modern data systems use the convention that a negative-going
voltage is a 1 and a positive-going voltage is a 0.
One key standard, probably familiar to many users of this text, namely
EIArTIA-232, states that a voltage of q3 volts or greater is a 0 and that a
voltage of y3 volts or greater e.g, y4 volts. is a 1. Voltages between y3
and q3 volts are indecisive. Many other standards generally follow suit such
as EIArTIA-422, EIArTIA-423, and ITU-T Rec. V.10 and V.11, to name
a few.
5.7 BIT-PARALLEL AND BIT-SERIAL

Data may be transferred in serial format over a single line one bit at a time
contiguously or in parallel over several lines at once. In parallel transmission, bits are transmitted a character at a time, and each bit of the set of bits
that represent a character has its own wire or line. An additional wire called
the strobe or clock lead notifies the receiver unit that all the bits are present
on their respective wires so that the voltages on the wires can be sampled.
Parallel transmission is shown conceptually in Figure 5.10. Figure 5.11
illustrates a typical parallel-to-serial conversion circuit. The 1 and 0 bits at
either extreme of the register are bits that are appended for controltypically startstop operation, which is discussed in Chapter 6. A register is a
digital storage device.
Computers and other digital data processors generally operate on parallel
data, so data are transferred in parallel between these devices wherever they
are in close proximity. However, as the distance between these devices
increases, not only do the multiple wires become more costly, but the
86
DATA TRANSMISSION I
Figure 5.10. Bit-parallel transmission, 8-bit character.
Figure 5.11. Serial data transmitter with parallel data input, configured for startstop operation and an 8-bit character.
complexity of line drivers and receivers increases, due to the major difficulty
of properly driving and receiving signals on long runs of wire pairs.
Bit-serial transmission is generally used where the cost of the communications medium is high enough to justify a relatively complex transmitter and
receiver system that will serialize bits that represent a character, send over a
single line, and reassemble them in parallel form at the receiving end. In
most data transmission systems, serial transmission is preferable to parallel
transmission.
5.8 BASEBAND
The term baseband is defined by the IEEE Ref. 1. as the band of
frequencies occupied by the data. signal before it modulates a carrier or
subcarrier . frequency to form the transmitted line or radio signal. Baseband,
therefore, is the data signal as it leaves or enters a data processing or
DATA RATE
87
presentation device. The signal has frequency content extending into the
direct current region. Baseband data can be transmitted hundreds and even
thousands of feet. This is commonly done on wire pair. Figure 5.9 shows
several common baseband data waveforms. Transmission distance is limited
because of several factors. The data signal suffers loss due to the length and
characteristics of the wire pair. It can be corrupted by noise, which often can
be related to length of the wire pair. The signal itself will become distorted
due to the electrical characteristics of the medium. Distances can be extended by the use of regenerators, which were described above.
Long distances can be achieved by having the baseband signal modulate a
radio or light carrier signal. Some refer to this as carrier transmission.
5.9 DATA RATE

In this text, data rate and bit rate are synonymous. Both are defined as
the transmission rate in bits per second bps.. Bit rate is constrained by the
capability of the transmitting and receiving devices, the bandwidth of the
intervening media, and a number of electrical impairments typical of each
medium in question.* We might successfully transmit at a megabit per
second, but the distant-end receiving device may be suffering an excessive
error rate on receiving that signal. In other words, a high percentage of the
characters received are in error, and the data circuit is useless to a user.
Thus, when we use the terms bit rate or data rate, there is some implication of
satisfactory operation end-to-end.
5.9.1 Error Rate
The most common measure of a data connectivitys performance is error rate.
Often the unit of measurement is bit error rate BER., which is the probability of bit error. BER is often expressed using the powers of 10. For example,
a certain data circuit has a bit error rate of 1 in 10 5. This means that there is
a probability that one error will occur for every 100,000 bits transmitted.
More often such a BER is expressed as 1 = 10y5.
When describing a bit rate on a data circuit, it often is most useful to give
the bit rate with the accompanying BER. For example, the data circuit on the
Boston to New York leg operates at 64 kbps with a BER of 3 = 10y7. With a
statement such as this, circuit performance is implicit.
Bit error rate performance in industrialized nations has shown dramatic
improvements. CCITT specifies a bit error rate of 1 = 10y6 or better on
ISDN circuits ITU-T Rec. G.821.. On WANs a user should specify 5 = 10y10
or better see Bellcore TSGR, Ref. 9.. Bit error rate requirements and other
expected performance criteria are discussed at length in Section 6.3.3.
*These impairments are described in Chapters 6 and 8.
88
DATA TRANSMISSION I
REFERENCES
1. The IEEE Standard Dictionary of Electrical and Electronic Terms, 6th ed., IEEE
Standard 100-1996, IEEE, New York, 1996.
2. International Reference Alphabet (IRA) Information Technology7-bit Coded Character Set for Information Interchange, CCITT Rec. T.50, ITU Geneva, September
1992.
3. Common Long-Haul and Tactical Communication System Technical Standards, MILSTD-188-100, US Department of Defense, Washington DC, November 15, 1972.
4. American National Standard, Coded Character Sets: 7-bit American National Standard Code for Information Interchange (7-bit ASCII), ANSI X3.4-1986, ANSI, New
York, 1986.
5. Interface Between Data Terminal Equipment and Data Circuit Terminating Equipment
Employing Serial Binary Data Interchange, EIArTIA-232E, Telecommunications
Industry Association, Washington DC, July 1991.
6. Equialence Between Binary Notation Symbols and the Significant Conditions of a
Two-Condition Code, CCITT Rec. V.1, Fascicle VIII.1, IXth Plenary Assembly,
Melbourne, 1988.
7. Electrical Characteristics for Unbalanced Double-Current Interchange Circuits Operating at Data Signaling Rates Nominally up to 100 kbps, ITU-T Rec. V.10, ITU
Geneva, March 1993.
8. Electrical Characteristics for Balanced Double-Current Interchange Circuits Operating
at Data Signaling Rates up to 10 Mbps, ITU-T Rec. V.11, ITU Geneva, October
1996.
9. Transport Systems Generic Requirements (TSGR): Common Requirements, Telcordia
Bellcore. GR-499-CORE, Issue 1, Piscataway, NJ, December 1995.
6
DATA TRANSMISSION II
6.1 INTERPRETING A SERIAL STREAM OF BITS

6.1.1 The Problem
A serial bit stream consists of bits transmitted one at a time sequentially and
contiguously. At the very instant one bit ends, the next bit begins. Consider
the serial bit stream shown in Figure 6.1. It contains the ASCII character A
preceded by an ASCII character and followed by another ASCII character.
We presume that the receiver knows it is receiving ASCII characters at some
nominal bit rate. Each character in the serial bit stream in the figure consists
of 7 bits plus an added parity bit for a total of 8 bits per character. If we
would allow that the receiver samples at the center of each bit, then the
receiver would be in exact synchronism with the transmitter. Getting and
keeping the receiver in exact synchronism with its companion far-end transmitter is yet another problem that must be solved. We address this issue in
the next two subsections.
6.1.2 StartStop Transmission

One of the oldest methods of data circuit timing and synchronization is
startstop transmission. Many entities, including CCITT, call this type of
transmission asynchronous. The IEEE calls this type of transmission nonsynchronous. Startstop transmission dates back to the first automatic telegraphy circuits. In those days the mark or 1 was the active state and the
space or 0 was the passive state. A typical ASCII startstop character is
shown in Figure 6.2. As illustrated in the figure, the beginning and end of
each character is well-delineated.
Note: The 1s and 0s in Figure 6.2 are using todays convention: A
positive-going voltage indicates a 0, and a negative-going voltage indicates
a 1.
89
90
Figure 6.1. A serial ASCII bit stream.
Figure 6.2. Startstop character format. SS, start-space; SM, stop-mark. ASCII letter A and
the beginning of the ASCII letter S. Note: The duration of the stop-space is one unit interval
(i.e., duration of 1 bit), and the stop-mark (in the figure) has a duration of two unit intervals
(i.e., 2 bits).
In those old teleprinters, the convention was adopted that when the line is
idle, current is flowing e.g., voltage is present .. The commencement of the
transmission of a character or string of characters is indicated when this
current is interrupted and the normal current or voltage. state drops to 0,
the space condition. To start the receiver teleprinter, the line condition
changes from one of current to a no-current condition, and the 0 state
holds for the duration of one bit 1-bit time.. The space element is called
the start bit or start element. The start element is followed by eight successive
bit times or unit intervals see Figure 6.2. which represent the character
being sent. These bit times may be either a 1 active state. or 0 passive
state. in any combination thereof for 8 consecutive bits. Note that if 8
information bits are used, we could be dealing with the EBCDIC code or
ASCII with the added parity bit..
The last element transmitted is the stop element, which can have 1-,
1.5-, or 2-bit duration, depending on the convention accepted. The stop
element allows the receiver to coast back to a known position in time until
the beginning of the next start space element signifying the beginning of
the next character.
With startstop transmission it is true that precise synchronization between the far-end transmitter and the near-end receiver is not necessary.
However, the receiver must sample the income bit stream at roughly the
same rate as the transmitted bits are gated out on the transmit side so that
the receiver can sample each incoming bit at the proper time.
When a line is idle in startstop transmission, a continuous 1 or mark is
transmitted and current is flowing in the data loop. Once the receiver detects
a transition from current flow to no-current flow, it counts out 5, 6, 7, or 8
bits, depending on the code used, and then knows that the next element must
be the stop-mark or current-flow state. If another character directly follows
the first, there again is the mark-to-space transition right after the stop-mark
and the start-space begins and lasts for the duration of 1 bit, and the
receiver starts its count again. As one can see, the receiver must maintain
INTERPRETING A SERIAL STREAM OF BITS
91
some sort of synchronization with its companion far-end transmitter only

during the 5-, 6-, 7-, or 8-unit intervals or bits in this case. of one character.
The receiver catches up with the transmitter or slows down to let the
transmitter catch up with its clock during the time period of the stop-mark
element.
The advantage of asynchronous startstop. transmission is that very loose
requirements are placed upon the timing and synchronization of the data
link. We can easily delineate the beginning of a character at the receiver by
the mark-to-space transition that begins the start-space element. The end of
a character is defined by the stop-mark. Timing starts at the start-space and
ends at the stop-mark, not a very difficult feat to accomplish.
The disadvantage of this type of transmission is the overhead of the
control elements: the start-space and stop-mark. Let us say that the code
used is ASCII with 1 parity bit, for a total of 8 bits per character. Allow the
start-space and stop-mark each to be of 1-bit duration. Thus it takes 10 bits
per character to transmit 8 useful bits. There is a built-in 20% inefficiency.
6.1.3 Synchronous Transmission
Synchronous transmission avoids most of the inefficiency cited above. One
method proposed to ensure that the receiver clock is kept in exact synchronism with the gating of transmitted bits was to extend a separate clocking
lead from the transmitter to the receiver. In most situations this approach is
untenable.
If we consider for just a moment the makeup of a serial bit stream, we can
see that there is a built-in clocking system, namely the mark-to-space and
space-to-mark transitions of the information bits themselves. If we slave the
receiver clock to the incoming transitions, often accomplished by phase-locked
loop techniques, the receiver will maintain exact synchronism with the
transmitter, but will be delayed by the propagation time of the intervening
circuit. Of course the bits we are receiving are delayed by the same amount
as we would expect.
With synchronous transmission, characters are delineated by simple bit
counting. Our most common data transmission codes are based on 8-bit
characters, typically ASCII with a parity bit and EBCDIC. Some call an
8-bit character a byte, but the term octet is perhaps a better term. All
bit-oriented protocols rigidly require that data frames be composed of an
integer number of octets. One might imagine character markers at all integer
multiples of eight bits. Eight-bit characters are also popular with processors
because there are 8-bit, 16-bit, 32-bit, and now 64-bit processors. For example, a 64-bit processor can gobble up 8 characters at a time.
The incisive reader will immediately have picked up the next problem.
How do we know when a message a block, frame, packet, or cell. begins and
ends when synchronous transmission is employed? A message always starts
with a unique field or bit sequence called a flag. The majority of bit-oriented
92
protocols use the sequence 01111110 as the starting and ending flag. At the
end of the starting flag, we start counting increments of eight delineating bits.
Of course we must start counting with the very first bit appearing after the
flag sequence.
At this juncture we must define a character. The IEEE Ref. 1. defines a
character data transmission. as one of a set of elementary symbols which
normally include both alpha and numeric codes plus punctuation marks, and
any other symbol which may be read, stored or written, and are used for
organization, control, or representation of data. With a few exceptions such
as the ITA 2 code., all of our characters are 8 bits or 1 octet long. A
character can be an operational sequence performing the function of space
bar, hard return, soft return, and so on.
Before proceeding further, let us consider what a generic data frame looks
like. Depending on the type of system, some may call a data frame a packet,
a block, or a cell. A generic data frame is shown in Figure 6.3. Note that the
flag is shown as the first and last field of the frame. The address and control
fields together are often called the header. Generally packets and cells have
fixed length text or info field; frames and blocks may carry varying-length text
or info fields. The term arying should not be confused. All text or info fields
have length constraints that is, maximum length and minimum length.
Thus, it would appear we would like long frames long info fields. to
amortize overhead. However, on noisy circuits where a lot of frame repeats
are required, short frames are more desirable.
Frame structure has had detailed explanation in Chapters 2 and 3. Our
discussion here deals with synchronous transmission. With startstop transmission, the beginning and end of a character were well-delineated by the
start and stop elements. In the case of synchronous transmission, the flag
carries out this function; the receiver-processor counts bitsroctets after the
initial flag. The processor has been programmed in advance for the link-layer
protocol to be used. Thus, the processor knows a priori where each of the
header fields begins and ends and acts upon the contents.. For fixed-length
frames, the processor knows the length of the info field, where it ends and
where the FCS begins, and where the frame ends, by means of the trailing
flag.
To measure the efficiency of a certain data link, we sum all the bits in the
header, FCS, and flag fields and divide this value by the total number of bits
in a frame. We then subtract that value from 100 for a percent efficiency
value.
Figure 6.3. A generic frame or packet of data. See Chapters 2 and 3 for more detailed
discussion of frame structure.
TIMING DISTORTION IN A SERIAL BIT STREAM
93
Example. A certain data link has 16 bits for the two flags, 16 bits of address
information, 8 bits of control information, and 16 bits for the FCS field. The
information field has 1600 octets 12,800 bits..
There are 56 bits of overhead, and the total number of bits in the frame is
56 q 12,800 or 12,856 bits, 56r12,856 s 0.004356.
100 y 0.004356 s 99.995%
Without considering header overhead, startstop transmission based on
11-bit characters of which 3 bits are used for start and stop elementswill
have the following efficiency:
3r11 s 0.2727
100 y 0.2727 s 99.727%
On synchronous circuits that are idle no traffic is being passed ., fill or idle
patterns are transmitted to maintain synchronization. If a circuit is dormant
and at some moment in time traffic will be passed, a short training period
should precede the traffic to allow synchronization and to initiate other
features such as automatic equalizers.
6.2 TIMING DISTORTION IN A SERIAL BIT STREAM
We define distortion as the unwanted modification or change of signals from
their form by some characteristic of the communication line or equipment. In
data communications, distortion is usually manifested by improper timing.
One type of distortion is called bias distortion, where the duration of the
mark pulse the 1 pulse. is elongated a certain amount and the space pulse
0 pulse. is shortened a corresponding amount. Of course we can have the
reverse situation, where the space pulse is lengthened at the expense of the
mark pulse.
One cause of bias distortion is wire-pair capacitance. Another cause is that
it is likely that the resistance of the driver transmitter end. and the receiver
circuitry is different for the mark-to-space transition than for the space-tomark transition. Thus, there will be a different amount of time required to
charge the cable capacitance of the two transitions. One of the most wellknown standards involved in this interface is EIArTIA RS-232 Ref. 2.,
which states that there should be no more than 2500 picofarads of cable
capacitance on such a circuit. If the value of 2500 picofarads is exceeded, the
increase in bias distortion can be dramatic. Figure 6.4 shows a binary bit
stream with marking bias bias distortion..
Figure 6.4. A binary data serial bit stream with marking bias. Ideally the 1 bits and the 0 bits
should be of equal duration.
94
Figure 6.5. Types of timing distortion. Note: In bit stream drawings, such as this one, the
current sense convention as established in EIAr TIA-232 and in similar standards may or may
not be followed, depending on what is being expressed. This convention established the 1
as a negative-going voltage and the 0 as a positive-going voltage, as shown in Table 6.2.
The sense is reversed in the drawing above.
The timing distortion of signal elements bits. can be resolved into two
components, commonly termed systematic and fortuitous. Systematic distortion is repetitive and is the average distortion as a data signal is transmitted
through the system. The fortuitous component is the variation from the
average. The sum of the systematic and fortuitous components makes up the
total distortion.
Systematic distortion is further broken down into bias and characteristic
distortion. Bias distortion is brought about by the lack of symmetry in the
system, as was discussed above. Characteristic distortion is brought about by
intersymbol interference ISI. resulting from nonlinear phase versus frequency characteristic. Fortuitous distortion is caused by extraneous factors
such as noise, crosstalk, and carrier phase effects, all of which may occur at
random with respect to the signal bit sequence.
There are two ways of expressing the amount of distortion. It can be
stated in time units, commonly milliseconds, or it can be expressed as a
percent of a unit interval a bit.. The latter is more common in the industry.
Of course when using percentage, we must state the speed of transmission
i.e., the data rate. at which the percent of distortion applies. The speed is
almost universally given in bits per second. Figure 6.5 shows several types of
distortion that have been discussed above.
6.3 THE TRANSMISSION OF DIGITAL DATA
6.3.1 Baseband Transmission
Data are transported electrically on a copper wire pair over comparatively
short distances. For example, data may be transmitted across a room, across
a floor, throughout a building, or, in some cases, throughout several contiguous buildings. This is commonly called baseband transmission. Distances are
limited by the buildup of loss, distortion, and noise. Distance is also a
function of bit rate.
THE TRANSMISSION OF DIGITAL DATA
95
As a wire pair is extended, resistance, capacitance, and inductance of the

cable increase. As capacitance increases, it takes longer to charge the
capacitance for each bit transmitted, thus limiting bit rate andror distance.
Keep in mind that we can view a wire pair as a very long capacitor. It meets
the basic requirements because it consists of two conductors separated by an
insulator.
Hysteresis due to inductance is another limiting factor. All of these factors
limit bit rate and distance for baseband transmission.
One way to increase distance for baseband applications is to use a
balanced pair rather than an unbalanced pair. Remember that with an
unbalanced pair, one lead is grounded. In the case of balanced transmission,
neither lead is grounded, and one lead is used for voltage reference, eliminating problems with ground potential differences from one end of the circuit to
the other. Baseband distance can also be extended by the use of regenerative
repeaters. The use of larger-diameter wire can also offer some advantages
because it reduces loss and resistance per unit length. The use of low-capacitance wire pairs is yet another approach to achieve higher bit rates over
longer distances.
6.3.2 Transmission of Data Over Longer Distances
6.3.2.1 The Voice Channel. We look to our ubiquitous telephone facilities to transport data electrically over longer distances. It was cost-effective
because the telephone was everywherethat is, worldwide. To do this we
had to accept certain shortcomings of telephone facilities, particularly the
intervening network. The telephone network consists of voice channels.
ITU-T defines a voice channel as occupying the voice-frequency band
from 300 to 3400 Hz. Two important shortcomings can immediately be seen.
First, the voice channel cannot accommodate DC direct current . transmission, which we have called baseband transmission. Of course, the 0-Hz DC
component of baseband transmission cannot be coupled to the voice channel
passband. Second, its 3100-Hz bandwidth i.e., 300 Hz3400 Hz. places a
severe limitation on data rate. If we were to allow 1 bit per hertz of
bandwidth, then the maximum bit rate over a voice channel is 3100 bps. In
practice, it is really less than this value. Certainly most readers know that
considerably greater bit rates are handled by the voice channel by using
sophisticated coding and modulation schemes. We discuss these important
techniques at length in Chapter 8. Of major importance at this juncture is
that we must convert the DC baseband serial bit stream to a signal that is
compatible with the voice channel. The device that carries out this function
is the popular modem, which is an acronym for modulatordemodulator.
6.3.2.2 Modems. A modem is any device that carries out i. a modulation
function on the transmit side of a circuit and ii. a demodulation function on
the receive side. However, for our discussion we define a modem as a device
96
TABLE 6.1 Data Modulation Terminology

Conventional
Terminology
Amplitude modulation
Frequency modulation
Phase modulation
Data Modulation
Terminology
Abbreviation
Commonly Used
Amplitude shift keying

Frequency shift keying
Phase shift keying
ASK
FSK
PSK
that generates an audio tone whose frequency is compatible with the voice
channel. This tone is modulated by a baseband digital data signal in either
the frequency, amplitude, or phase domain or combinations of two of these.,
and the resultant signal is coupled to the voice channel. On the receive side,
a modem converts the modulated audio tone back to a digital data serial bit
stream.
There are three generic types of modulation: amplitude modulation AM.,
frequency modulation FM., and phase modulation PM.. As we remember, a
wave e.g., a sine wave. is defined in terms of amplitude, phase, and
frequency. Information can be transported on such a wave by varying the
value of any one of these properties in accordance with an electrical baseband signal which contains that information. This, in essence, is modulation.
Many data transmission terms derive from its predecessor, the telegraph.
When dealing with data transmission, consider Table 6.1. The term keying
derives from Morse code operation of a transmitter using the Morse key. The
key was held down i.e., contact was made. just for a moment for a dot, and
it was held down for a longer period for a dash. When the key was pressed
down, an electrical contact was made and the transmitter was turned on,
thereby emitting a carrier frequency CW. wave. When the key was released,
the transmitter was turned off and the carrier wave ceased. The terminology
further evolved through automatic telegraph, a true forerunner of data
communications.
When these automatic telegraph systems were introduced, the radiofrequency carrier remained on all the time, and the information was transmitted by means of shifting changing. the carrier frequency upward in
frequency some hundreds of hertz for a mark or 1 and downward in
frequency an equal amount for a space or 0. Thus the term shift keying
came about; that is, the carrier was shifted.
For ASK modulation, a carrier is turned on for a mark or 1 and is turned
off for a space or 0, PSK systems came later, and in this case a binary 1 is
transmitted at reference phase, and a binary 0 is transmitted with the phase
retarded, usually 180 degrees.
Data modems employ FSK modulation almost universally for all systems
operating at 1200 bps or less. Above 1200 bps, some form of PSK is used. For
higher data rate modems, above 4800 bps typically CCITT Rec. V.29; see
Ref. 3., a combination of ASK and PSK is used.
97
6.3.2.3 Voice Channel Impairments to Data Transmission. The voice

channel remains the basic building block of the public switched telecommunication network PSTN.. Figure 6.6 illustrates the concept of a voice channel
traversing the telecommunications network.
The voice channel is constrained in frequency because it will be multiplexed with other voice channels to share the expensive transmission media.
As a minimum, there is a low-pass filter at every voice channel input of the
multiplexing channel bank and another low-pass filter at the companion
demultiplexing bank at the other end of the circuit. There may be other
filters in the intervening circuit.
The transmission media can be wire pair, coaxial cable, fiber-optic cable,
line-of-sight microwave radio, and satellite links. On long circuits we would
expect a mix of media.
There will be one or more modulation stages which the voice channel will
be subjected to as it traverses the circuit end-to-end over the transmission
media. Each modulation and amplification stage generates noise, and each
will cause errors in the case of digital transmission.
The intervening equipment and media through which the voice channel
passes cause transmission impairments to the voice channel. There are three
basic impairments:
Amplitude distortion frequency response.

Phase distortion delay distortion.
Noise
In fact, these three impairments will affect any transmission medium, especially metallic media e.g., wire pair and coaxial cable.. There are two other
impairments that affect the voice user much more than the data user. These
are echo and singing.
Figure 6.6. A simplified conceptual illustration of a voice channel traversing the telecommunications network. Note that the multiplexing may be in the time domain or in the frequency
domain. If the multiplexing is in the time domain, the switching is digital using time division
techniques. M, modulator; D, demodulator.
98
Amplitude Distortion. Amplitude distortion is called frequency response by

audio enthusiasts. It is the distortion caused by the variation of transmission
loss with frequency. The frequency band of interest here is the voice band or
voice channel, that band between 300 and 3400 Hz. Suppose we were to
place a 1000-Hz signal at the near-end input of a voice channel at a
y10-dBm level. The level measured at the far-end output of the voice
channel might be y16 dBm. If the channel were free of amplitude distortion,
or we had what is called a flat channel, we would expect to get y16 dBm for
any frequency measured at the far-end output between 300 and 3400 Hz with
a near-end input of y10 dBm. A typical amplitudefrequency distortion
characteristic for a voice channel is shown in Figure 6.7. Figure 6.8 shows
amplitude distortion for a voice channel as specified crosshatch . and as
measured with equipment back-to-back. Low-pass and band-pass filters are a
leading cause of amplitude distortion, particularly when they are in tandem,
as one might expect in the PSTN.
The effect of nonuniform attenuation across the voice channel on data
communications is to distort the received spectrum and, in turn, the data
signal waveform. A reasonable amount of amplitude distortion can be tolerated on a voice channel transporting data. Provided that the loss is not
actually infinite at any important signal frequency, it is theoretically possible
to construct linear compensating networks that equalize amplitude characteristics over the frequency band of interest. However, if the variation of loss is
very great, there are practical difficulties in supplying sufficient signal power
to override noise in the high-loss portion of the voice band. The resultant
Figure 6.7. Permissible attenuation variation with frequency, reference 800 Hz. Objective for
worldwide four-wire chain of 12 circuits in tandem, terminal service. (From Ref. 4. Courtesy of
ITU-CCITT.)
99
Figure 6.8. Typical attenuation distortion curve for a voice channel. This is a North
American-defined, 200- to 3200-Hz voice channel with reference frequency at 1000 Hz. (From
Ref. 5. Reprinted with permission.)
effect is a fairly definite limit on useful bandwidth and, consequently, data

transmission rate.
Phase Distortion. Phase distortion constitutes the most limiting impairment
to data transmission, particularly over the telephone voice channel. What we
mean here is that it places an upper limit on data rate. Speech communications over a voice channel can tolerate phase distortion far in excess of that
which is severe for high data rate transmission over that same channel.
Phase distortion results from different velocities of propagation at different frequencies across the voice channel. A signal takes a finite time to
traverse the telecommunications network end-to-end. If two tones are keyed
exactly at the same time at the near end, one at 400 Hz and one at 1800 Hz,
the 1800-Hz tone will arrive at the far end before the 400-Hz tone. We can
say that the 400-Hz tone was delayed relative to the 1800-Hz tone. Thus we
have come to use the term delay distortion. Figure 6.9 illustrates typical
differential delay across a voice channel.
The measurement of enelope delay distortion EDD. is a common method
used to determine the deleterious effects of phase distortion on data transmission. One technique Ref. 6. injects a test tone at several test frequencies
in the voice channel. The test tone is amplitude-modulated with a lowfrequency signal 83.33 Hz.. At the far end of the circuits, three signals
appear: the carrier, an upper sideband, and a lower sideband. The three
signals arrive at different timesthat is, an envelope of arrival times.
Our interest here is relative envelope delaythat is, the delay at one
frequency compared to another where one of these frequencies is a reference
100
Figure 6.9. Typical differential delay across a voice channel. (From Ref. 5. Reprinted with
permission.)
frequency. It is usually selected to be at a point of minimum delay. This

occurs on voice channels between 1500 and 2000 Hz. The reference frequency is assigned a delay of 0 s, and signal energy at all other discrete
frequencies above and below the reference arrive later.
Excessive phase distortion gives rise to ISI intersymbol interference ..
That is, signal energy components of the first symbol remain, impairing the
capability of a decision circuit to determine whether a 1 or a 0 has been
received for the second contiguous symbol. ISI can be a major contributor of
errors in a received data bit stream.
Noise. There are four types of noise one must consider as impairments to
data transmission. These are:
Thermal noise
Crosstalk
Intermodulation noise
Impulse noise
Thermal noise, as defined by the IEEE Ref. 1., is the noise occurring in
electric conductors and resistors and resulting from random movement of
free electrons contained in the conducting material. The noise derives from
the fact that such random motion depends on the temperature of the
material. Thermal noise has a flat power spectrum out to extremely high
frequencies.
101
Thermal noise is always present in an electrical circuit. It can be characterized by an equivalent noise temperature measured in kelvins temperature
scale based on absolute zero. or by noise figure. Noise figure in decibels . of
a device can be defined as the difference of the signal-to-noise ratio between
the output of a device and its input.
In this text, unless otherwise specified, we will be dealing with thermal
noise. Thermal noise is random in nature and is treated statistically a
Gaussian distribution .. Thermal noise thresholds noise floors. can be calculated based on a thermal noise floor of an ideally perfect receiver at absolute
zero where the noise level is y228.6 dBW in 1 Hz of bandwidth. For the
same receiver at room temperature 290 K., this value is y204 dBWrHz.
Crosstalk is defined by the IEEE Ref. 1. as the undesired energy appearing in one signal path as a result of coupling from another signal path. Path
in this context implies wires, waveguides, or other localized or constrained
transmission systems.
For example, on telephone circuits we may hear extraneous voice transmissions. Data circuits can be even more susceptible to extraneous signal
energy. Excessive crosstalk, in this case, manifests itself usually as error
bursts or, if crosstalk coupling loss is very low, in an unacceptable error rate.
Intermodulation noise is extraneous signal energy derived from some portion of the wanted energy when the signal energy is passed through a
nonlinear device. It is characterized by the appearance at the output of a
device of frequencies equal to the sums and differences of integral multiples
of two or more component frequencies present in the output signal. In
Figure 6.10 the input of the device shown has frequencies A and B. For the
voice channel, often 1000 and 1200 Hz are used as examples of frequencies
A and B.. The terms intermodulation noise and intermodulation products are
synonymous, sometimes referred to as IM noise.
Intermodulation products IM noise. are generated in nonlinear devices
such as amplifiers, transformers, and coils which use ferrous materials, even
waveguides. Although many amplifiers operate in a linear region, they display
what are called weak nonlinear characteristics. And as such we have to
contend with their intermodulation products.
Of principal concern are second- and third-order products. For the case of
a two-frequency input to a nonlinear device as shown in Figure 6.10, there
are the second-order products. The first of these is simply the second
harmonic 2 F1 andror 2 F2 ., which is the same as that obtained if each input
Figure 6.10. Examples of Intermodulation (IM) products. F1 , frequency 1; F 2 , frequency 2.
102
were applied separately. The other type of second-order product consists of

the sum and difference frequencies of each pair of applied frequencies. It
can be shown Ref. 7. that if each of the input frequencies has the same
amplitude, the power of the sum and difference products will be 6 dB higher
than the power of the second harmonic products.
There are three different types of third-order products. There are the
third harmonics 3F1 and 3F2 . There are the 2 F1 " F2 products, which
consist of the sum and difference of one frequency and the second harmonic
of another. Compared with the third harmonic amplitude, these products are
larger by a factor of 3F2rF1. Thus, if each of the input frequencies has the
same amplitude, the power of the 2 F1 " F2 products will be 9.6 dB higher
than that of the third harmonic power. The third type of third-order product
is the three-frequency situation considering their sums and differences such
as F1 " F2 " F3 .
Impulse noise is defined as any burst of noise which exceeds the rms noise
level by a given magnitude. The magnitude is nominally 12 dB for a 3-kHz
bandwidth Ref. 7.. Impulse noise is the result of electrical sparks or
flash-overs of a switch being turned on and off, car ignition noise, lightning,
operation of mechanical telephone switches e.g., step-by-step, crossbar, and
rotary., and electrical motor brushes. Impulse noise spikes occur randomly
and have a broad spectral content.
In most cases, impulse noise can be neglected for telephone speech
transmission. However, for data transmission it can be a major source of
errors. Impulse noise objectives are based primarily on the error susceptibility of data signals. The susceptibility depends on the type of modulation used
on a data circuit, the bit rate, and the characteristics of the transmission
medium employed. The important characteristics of the transmission medium
are a. the amount of phase and attenuation distortion and b. the rms
signal-to-peak impulse noise amplitude ratio.
Impulse noise parameters are based on the fact that digital error rates in
the absence of other impairments are approximately proportional to the
number of impulses which exceed the rms data signal by a certain number of
decibels. Reference 7 defines impulse noise as a voltage increase of 12 dB or
more above rms noise for 10 milliseconds or less. Impulse noise is measured
by a counter which responds to waveform excursions above a settable
threshold. One objective for the PSTN is no more than 15 counts hits. in 15
minutes for 80% of all connections at a threshold of 5 dB below the
received-signal level. The received-signal level is established by transmitting a
1004-Hz tone on the voice channel Ref. 7, page 157..
6.3.3 Data Circuit Performance
6.3.3.1 Introduction to Error Performance. Certainly the most common
performance measure of a data circuit is its error performance. This concept
was introduced in Section 5.9.1. There is bit error performance expressed as
bit error rate, and there is block error performance. However, with the latter,
103
one must express how long a block is either in bits or in octets. Block error
rate is coming more into favor with ATM and its fixed-length cells. A data
block is a set of contiguous bits or octets which make up a definable quantity
of information. The term frame can be used interchangeably with a block.
Bit error rate ratio. BER. is defined as the ratio of the total bits received
at a data-link destination to the bits received in error; or the contrary, the
bits received in error to the total number of bits received. It is expressed as 1
in 10 n or 1 = 10yn. For example, if 100 bits are received and one of those
bits is in error, the bit error rate BER. is 1 = 10y2 , or 1 error in 10 2 .
Let it be understood that there is no data circuit completely free of errors.
However, we may have to wait awhile for an error to occur, but eventually an
error will occur. There may be periods of time when a circuit is free of errors.
The common unit of performance in this case is the error-free second EFS..
Error rate should be handled statistically, and thus we should talk about
probability of error. This is because one of the most common causes of bit
errors is random noise peaks, which are treated statistically. So when we
express BER, we really are describing the probability of error.
Another measure of data circuit performance is throughput. To many of
us, throughput defines the number of correct bits, blocks, or frames received
which are error-free. To others, it means the number of useful bits received
that are not in error. Unless we concretely define useful bits, it is a
comparatively meaningless measure of performance. For example, should the
start and stop bits be included as useful bits? As we delve further into our
discussion of data communications, we will encounter many other bit groupings that are necessary to telecommunication engineers but have no use for
the end user, who may be a programmer. Header information is an example
in point. Headers direct traffic to a particular destination and carry out other
control functions, but carry no useful information for the data end user.
There are two generic categories of errors. These are random errors and
burst errors. If the cause of error is insufficient signal-to-noise ratio, such
errors will be random if the noise in question is random in nature such as
thermal noise.
An error burst is a string of contiguous bits with a very high probability of
being in error. There are two general causes of these error bursts: impulse
noise spikes and signal fading. If some portion of a data circuit traverses a
line-of-sight microwave link, fading may be expected. Fading may also occur
during heavy rainfall events on certain satellite circuits as well as on certain
line-of-sight microwave links. There may be other causes of error bursts such
as crosstalk.
6.3.3.2 Error Performance from an ITU-T Perspective. CCITT in the
1970s recommended a bit error rate no worse than 1 = 10y5. In the 1980s
with the advent of ISDN, CCITT tightened this value to 1 = 10y6 end-to-end
for the ISDN 64-kbps channel. In this case the 64-kbps digital channel is a
voice channel used as a bearer of data. CCITT Rec. G.821 Ref. 8. states that
fewer than 10% of 1-minute intervals should have a bit error rate worse than
104
TABLE 6.2 CCITT Error Performance Objectives for International ISDN Connections
Objective (Note 3)
Performance Classification
(a) (Degraded minutes)
(Notes 1, 2)
(b) (Severely errored seconds)
(Note 1)
(c) (Errored seconds)
(Note 1)
Fewer than 10% of 1-min intervals to have

a bit error ratio worse than 1 = 10 y 6 (Note 4)
Fewer than 0.2% of 1-s intervals to have
a bit error ratio worse than 1 = 10 y 3
Fewer than 8% of 1-s intervals to have any
errors (equivalent to 92% error-free seconds)
Note 1. The terms degraded minutes, severely errored seconds, and errored seconds are used as a
convenient and concise performance objective identifier. Their usage is not intended to imply the
acceptability, or otherwise, of this level of performance.
Note 2. The 1-min intervals mentioned in Table 1rG.821 and in the notes (i.e., the periods for
M ) 4 in Annex B ) are derived by removing unavailable time and severely errored seconds from the
total time and then consecutively grouping the remaining seconds into blocks of 60. The basic 1-s
intervals are derived from a fixed period.
Note 3. The time interval TL , over which the percentages are to be assessed, has not been specified
because the period may depend upon the application. A period of the order of any one moment is
suggested as a reference.
Note 4. For practical reasons, at 64 kbit rs, a minute containing four errors (equivalent to an error
ratio of 1.04 = 10 y6 ) is not considered degraded. However, this does not imply relaxation of the
error ratio objective of 1 = 10 y 6 .
Source: CCITT Rec. G.821, Fascicle III.5, Table 1rG.821, page 29, CCITT Blue Books, IXth Plenary
Assembly, Melbourne, November 1988.
1 = 10y6 . This 10% of 1-minute intervals is classified as degraded minutes.

Table 6.2 provides an overview of error performance recommended by
CCITT in Rec. G.821.
ITU-T Rec. G.826 makes use of block-based measurements to make
in-service measurements ISMs. easier to carry out. We review some terminology of ITU-T Rec. G.826 Ref. 9.. Consider Figure 6.11, which shows
some of the relationships to be discussed. The following definitions apply:
M s number of blocks per second
R s number of bits per second
t s 1rM
Figure 6.11. Block representation of signals. (From Figure 1, page 57, Ref. 10. Reprinted
with permission.)
105
We introduce the following definitions:

Errored Block EB .. A block in which one or more bits are in error.
Errored Second ES .. A 1-s period with one or more errored blocks.
Seerely Errored Second SES .. A 1-s period that contains G 30% EBs or
at least one severely disturbed period SDP.. For out-of-service measurements, an SDP occurs when, over a minimum period of time
equivalent to four contiguous blocks, either a. all the contiguous blocks
are affected by a high binary error density of 10y2 or b. a loss of signal
information is observed. For in-service monitoring purposes, an SDP is
estimated by the occurrence of a network defect. CCITT defines
network defect as the annexes to Rec. G.826 for different network
formats such as SDH, PDH, or cell-based see Chapter 16, ATM..
Background Block Error BBE .. An EB not occurring as part of an SES.
ES Ratio ESR .. The ratio of ESs to total seconds available time during a
fixed measurement interval.
SES Ratio SESR .. The ratio of SESs to total seconds in available time
during a fixed measurement interval.
BBE Ratio BBER.. The ratio of EBs to total blocks during a fixed
measurement interval, excluding all blocks during SESs and unavailable
time.
Error Performance Objectives (EPOs). The EPOs are summarized in Table
6.3. These objectives are measured over aailable time in a fixed measurement interval e.g., 1 month recommended.. All three objectives i.e., ESR,
SESR, and BBER. must hold concurrently to satisfy G.826, and they apply
end-to-end for a 27,500-km hypothetical reference path HRP., which is
shown in Figure 6.12.
Availability. The concept of available time is the same as was defined in
ITU-T Rec. G.821, except that the BER ) 1 = 10y3 criterion is replaced by
an SES criterion. Thus, unavailable time begins at the start of a block of 10
TABLE 6.3 Error Performance Objectives (EPOs) for G.826
Rate (Mbrs)
Bits r block
ESR
SESR
BBER
1.55
) 515
) 1555
) 55160
) 1603500
) 3500
20008000
20008000
400020,000
600020,000
15,00030,000
FFS b
0.04
0.05
0.075
0.16
a
FFS
0.002
0.002
0.002
0.002
0.002
FFS
3 = 10 y 4
2 = 10 y 4
2 = 10 y 4
2 = 10 y 4
10 y 4
FFS
No objective given due to the lack of available information.

FFS, for further study.
Source: Reference 10, Table 2, page 58.
106
Figure 6.12. Hypothetical reference path (HRP). (From Figure 2, page 57, Ref. 10. Reprinted
with permission.)
TABLE 6.4 Apportionment Rules

Percent of EPO
National Portion a
Block Allowance
17.5% to both
terminating
countries
Distance Allowance
1% per 500 km
International Portion a
Transit Allowance
2% per intermediate
country
1% per terminating
country
Distance Allowance
1% per 500 km
Satellite hops each receive 35%, but the distance of the hop is removed from the distance
allowance.
b
Four intermediate countries are assumed.
Source: Reference 10, Table 3, page 59.
consecutive SESs. The unavailable time stops and of course available time
begins. at the start of 10 consecutive ESs or better., each of which is not
severely errored.
Apportionment. The apportionment rules contained in G.826 are different
from those used in G.821. These rules are given in Table 6.4, and one should
refer to Figure 6.12 for the HRP.
The IEEE reference article Ref. 10. notes that it is not immediately
obvious that the EPO percentage figures of Table 6.4 yield 100%. However,
under the assumption of four intermediate countries and no satellite hops,
the following breakdown can be obtained:
Terminating countries:
2 = 17.5% q 2 = 1% s 37%
Intermediate counties:
4 = 2% s 8%
INTERFACE AT THE PHYSICAL LEVEL
107
Distance allowance:
27,500 km s 55 = 500 km 55 = 1%. s 55%
Total
100%
If one satellite hop is employed, it uses 35%, corresponding to a nominal
hop distance of 17,500 km.
The identification of an SES event during ISM is not so straightforward.
This is because the definition of an SES involves SDPs. SDP events can only
be measured during an out-of-service condition. Thus, equivalent ISM
events need to be defined if SDPs are to be detected while in-service. It is
recognized that there is not an exact 1 : 1 correspondence between SDP
events measured in-service and measured out-of-service. The objective of
annexes 2, 3, and 4 to G.826 is to provide ISM events that are reasonably
close to the G.826 error events.
It should be noted that G.826 generally applies to underlying transport
systems such as E1rT1 PDHChapter 7., SDHrSONET Chapter 15., and
cell-based transport Chapter 16.. Of course the impact of G.826 will be felt
on data systems being carried by that digital transport network.
The material in Section 6.3.3.2 was derived from Refs. 810.
6.3.3.3 Error Performance from a Bellcore Perspective. Reference 11

states that the bit error rate at digital cross-connects up through DS3 shall be
less than 2 = 10y10 , excluding burst errored seconds.* This value, of course,
is that of the underlying digital network, which is described in Chapter 7.
This value will be the best-case value where data bits are being mapped in
PCM time slots. Specific data offerings on the underlying digital network are
described in Section 7.6.
6.4 INTERFACE AT THE PHYSICAL LAYER
6.4.1 Introduction
A data link consists of equipment with resident software and related transmission facilities that allow two physically separated terminals data communication capability. Care must be taken when using this definition in the
multilayered setting of open systems interconnection OSI. or similar environments. What we describe below is the actual implementation of OSI
layer 1.
Let us consider a simple point-to-point data connectivity. It will consist of
a. data equipment on one side of the link and b. similar equipment on the
other side of the link. The two sides are connected by some transmission
facility. This facility may be made up of wire-pair, coaxial cable, fiber-optic
*Burst errored second is any errored second that contains at least 100 errors.
108
Figure 6.13. Functional diagram of data terminal equipment (DTE) and data communication
equipment (DCE) in a point-to-point configuration. The electrical and functional interface
between DTE and DCE is discussed in this section.
cable, or a radio system or combinations thereof. The interconnection may or

may not involve one or more switching nodes.
Our concern here is not the transmission media itself. That portion of the
link is adequately covered in Chapters 7, 8, 14, and 15. In this section we
discuss two key functional entities found in every data link. These are the
data terminal equipment DTE. and the data communication equipment DCE..
A simplified diagram of the DTE and DCE in a point-to-point configuration
Data terminal equipment includes the familiar PCsrworkstations, computersrprocessors, mass storage devices, printers, and servers. We sometimes
refer to these items of equipment as inputoutput IrO. devices.
The DCE is a functional entity such as a modem that makes the electric
signal from the DTE compatible with the transmission medium. It also
provides a number of control and interface functions. When we think of
DCE, most of us relate it to a modem, a device that generates a digitally
modulated audio tone compatible with the analog network. However, those
devices that condition a data signal for the digital network are also in the
category of DCE. Among these are the North American ATT. Digital Data
System DDS. DSUrCSU and the CCITT data conversionrconditioning
equipment typically meeting ITU-T Rec. V.110 requirements Ref. 12..
These latter items of equipment map a data bit stream into a 64-kbps digital
voice channel.
6.4.2 The DTEDCE Interface
This all-important interface is broken down into a functional interface and an
electrical interface. It is defined by many standards. The most familiar of
these standards with worldwide acceptance is the Electronic Industries
AlliancerTelecommunication Industry Association U.S.. EIArTIA-232
Ref. 2.. In this case the functional and electrical interfaces are found under
the same cover. They are detailed separately in many other standards. For
example, EIArTIA-530 defines a 25-pin functional interface between DTE
109
TABLE 6.5 EIAr

r TIA-232 Pin Assignments
Pin Number
CCITT
Number
1
2
3
4
103
104
105r 133
106
BA
BB
CAr CJ
(Note 1)
CB
6
7
8
9
10
107
102
109
CC
AB
CF
11
12
126
122r112
(Note 4)
SCFr CI
(Note 2)
13
14
15
121
118
114
SCB
SBA
DB
16
17
119
115
SBB
DD
18
19
20
141
120
108r1,r2
LL
SCA
CD
21
22
23
140r110
125
111r112
24
113
RLrCG
CE
CHrCI
(Note 2)
DA
25
142
TM
26
Circuit
(Note 3)
Description
Shield
Transmitted data
Received data
Request to sendr Ready for receiving
Clear to send
DCE ready
Signal common
Received line signal detector
(Reserved for testing)
(Reserved for testing)
Unassigned
Secondary received line signal
Detectorr Data signal rate selector
(DCE source)
Secondary clear to send
Secondary transmitted data
Transmitter signal element timing
(DCE source)
Secondary received data
Receiver signal element timing
(DCE source)
Local loopback
Secondary request to send
DTE ready
Remote loopbackrSignal quality detector
Ring indicator
Data signal rate selector
(DTErDCE source)
Transmit signal element Timing
(DTE source)
Test mode
No connection
Note 1: When hardware flow control is required Circuit CA may take on the functionality of Circuit
CJ.
Note 2: For designs using interchange circuit SCF, interchange circuits CH and CI are assigned to
pin 23. If SCF is not used, CI is assigned to pin 12.
Note 3: Pin 26 is contained on the Alt A connector only. No connection is to be made to this pin.
Note 4: Pin 11 is unassigned. It will not be assigned in future versions of EIArTIA-232. However, in
international standard ISO 2110, this pin is assigned to CCITT Circuit 126, Select Transmit Frequency.
Source: Reference 2. Copyright Electronic Industries Association rTelecommunications
Association. Reprinted with permission.
Industries
110
TABLE 6.6 EIAr

r TIA-232 Circuits by Category
Circuit
Mnemonics
CCITT
Number
AB
102
BA
BB
103
104
CA
CB
CC
CD
CE
CF
CG
CH
105
106
107
108r1,r2
125
109
110
111
CI
112
CJ
RL
LL
TM
133
140
141
142
DA
113
DB
114
DD
115
SBA
SBB
118
119
SCA
SCB
SCF
120
121
122
Circuit
Direction
Circuit
Type
Signal Common
Common
Transmitted data
Received data
To DCE
From DCE
Data
Data
Request to send
Clear to send
DCE ready
DTE ready
Ring indicator
Signal quality detector
(DTE)
(DCE)
Ready for receiving
Remote loopback
Local loopback
Test mode
To DCE
From DCE
From DCE
To DCE
From DCE
From DCE
From DCE
To DCE
Control
Control
Control
Control
Control
Control
Control
Control
From DCE
Control
To DCE
To DCE
To DCE
From DCE
Control
Control
Control
Control
To DCE
Timing
From DCE
Timing
From DCE
Timing
Secondary transmitted data

Secondary received data
To DCE
From DCE
Data
Data
Secondary request to send

Secondary clear to send
Secondary received line
signal detector
To DCE
From DCE
From DCE
Control
Control
Control
Circuit Name
Transmitter signal element

timing (DTE)
Transmitter signal element
timing (DCE)
Receiver signal element
timing (DCE)
Source: Reference 2. Copyright Electronic Industries Association r Telecommunications

Association. Reprinted with permission.
Industries
and DCE, and EIArTIA-422 and EIArTIA-423 uniquely define the electrical interface.
Table 6.5 gives the functional interface of EIArTIA-232. The table shows
the assignment and function of each pin in a 25-pin plug for a cable
connecting the DTE to the DCE. The related CCITT standard is CCITT
Rec. V.24 Ref. 13.. Table 6.6 gives the EIArTIA-232 interchange circuits by
category and provides equivalent CCITT Rec. V.24 circuit nomenclature.
r TIA-232 Electrical Interface. For data interchange circuits,
6.4.2.1 EIAr
the signal is considered in the marking condition i.e., binary 1. when the
voltage on the interchange circuit, measured at the interface point, is more
than y3 volts with respect to signal ground circuit AB.. The signal is
considered in the spacing condition i.e., binary 0. when the voltage is more
111
TABLE 6.7 Standard Signal Sense and Terminology

Interchange Voltage
Notation
Negative
Positive
Binary state
Signal condition
Function
1
Marking
OFF
0
Spacing
ON
positive than q3 volts with respect to signal ground circuit AB.. The region
between q3 volts and y3 volts is defined as the transition region. The signal
state is not uniquely defined when the voltage is in this transition region.
During the transmission of data, the marking condition is used to denote
the binary state one, and the spacing condition is used to denote the binary
state zero.
For the timing and electrical control interchange circuits, the function is
considered on when the voltage on the interchange circuit is more positive
than q3 volts with respect to circuit AB, and it is considered off when the
voltage is more negative than y3 volts with respect to circuit AB. The
function is not uniquely defined in the transition region between q3 volts
and y3 volts. Table 6.7 reflects standard signal sense and terminology.
It should be noted that there are three types of interchange signals
described in the standard: data, control, and timing.
EIArTIA-232 places the following limitations on the interchange signals
transmitted across the interface point, exclusive of external interference:
1. All interchange signals entering into the transition region shall proceed
through the transition region to the opposite signal state and shall not
reenter the transition region until the next significant change of signal
condition.
2. There shall be no reversal of direction change while the signal is in the
transition region.
3. For control interchange circuits, the time required for the signal to pass
through the transition region during a change in state shall not exceed
1 ms.
4. For data and timing interchange circuits, the time required for the
signal to pass through the transition region shall be in accordance with
the table below:
5. The maximum instantaneous rate of voltage change should not exceed
30 volts per microsecond.
Duration of Unit Interval
Maximum Allowable Transition Time
UI G 25 ms
25 ms ) UI G 125 s
125 s G UI
1 ms
4% of a unit interval
5 s
112
Definition of Unit Interval (UI) (Also called Signal Element). The part of a
signal that occupies the shortest interval of a signaling code. It is considered
to be of unit duration in building up signal combinations 1.. Unit duration
in this context is the duration of 1 bit, and thus the UI is a bit.
Note: Good engineering practice requires that the rise and fall times of
data and timing signals should be approximately equal within a range of 2 : 1
or 3 : 1..
6.4.2.2 Functional Description of Selected Interchange Circuits. The
circuit identifiers are shown in Table 6.6.
Circuit BATransmitted Data. Direction: To DCE. Signals on this circuit are
generated by the DTE and are transferred to the local DCE for transmission of data to remote DCEs. or for maintenance or control of the local
DCE.
The DTE shall hold circuit BA transmitted data. in the marking condition at all times when no data are being transmitted. In all systems, the DTE
shall not transmit data unless on condition is present on all of the following
circuits, where implemented:
1.
2.
3.
4.
Circuit
Circuit
Circuit
Circuit
CA request to send.
CB clear to send.
CC DCE ready.
CD DTE ready.
Circuit BBReceived Data. Direction: From DCE. Signals on this circuit are
generated by the DCE in response to data signals received from remote
DCEs., or by the DCE in response to maintenance or control data signals.
Circuit BB received data. shall be held in the binary one marking. condition
at all times when circuit CF received line signal detector. is in the off
condition.
On a half-duplex channel, circuit BB shall be held in the binary one
marking. condition when circuit CA request to send. is in the on condition
and for a brief interval following the on to off transition of circuit CA to
allow for the completion of transmission and decay of line reflections.
Circuit CARequest to Send. Direction: To DCE. This circuit is used to
condition the local DCE for data transmission and, on a half-duplex channel,
to control the direction of data transmission of the local DCE.
For simplex or duplex operation, the on condition maintains the DCE in
the transmit mode. The off condition maintains the DCE in a nontransmit
mode.
113
For half-duplex operation, the on condition maintains the DCE in the

transmit mode and inhibits the receive mode. The off condition maintains
the DCE in a receive mode.
A transition from off to on instructs the DCE to enter the transmit mode.
The DCE responds by taking such action as may be necessary and indicates
completion of such actions by turning on circuit CB clear to send., thereby
TABLE 6.8 EIAr

r TIA-530 Connector Contact Assignments
Contact
Number
CCITT
Number
Interchange
Points
Circuit
Category
Direction
Shield
BA
BB
CAr CJ
(Note 1)
CB
103
104
105r133
A-A
A-A
A-A
I
I
I
To DCE
From DCE
To DCE
106
A-A
From DCE
107
A-A
II
From DCE
7
8
9
10
CC
(Note 3)
AB
CF
DD
CF
102A
109
115
109
C-C
A-A
B-B
B-B
I
I
I
From DCE
From DEC
From DCE
11
12
13
14
15
DA
DB
CB
BA
DB
113
114
106
103
114
B-B
B-B
B-B
B-B
A-A
I
I
I
I
I
To DCE
From DCE
From DCE
To DCE
From DCE
16
17
18
19
BB
DD
LL
CArCJ
(Note 1)
CD
(Note 3)
104
115
141
105r133
B-B
A-A
A-A
B-B
I
I
II
I
From DCE
From DCE
To DCE
To DCE
108r1, r2
A-A
II
To DCE
1
2
3
4
5
6
20
Circuit
21
22
23
24
25
RL
CE
AC
DA
TM
140
125
102B
113
142
A-A
A-A
C-C
A-A
A-A
II
II
To DCE
From DCE
I
II
To DCE
From DCE
26
(Note 2)
Note 1: When hardware flow control is required, circuit

CJ.
Note 2: Contact 26 is contained on the Alt A connector
contact.
Note 3: In ANSIrEIA-530-87, circuits CC and CD were
category I and II circuits is not possible without circuitry
CA may take on the functionality of circuit

only. No connection is to be made to this
category I circuits. Interoperation between
similar to that in Figure B.1 in Ref. 14.
Source: Reference 16, Figure 3.9, page 16. Courtesy of Electronic Industries AlliancerTelecommunications Industry Association. Reprinted with permission.
114
indicating to the DTE that data may be transferred across the interface point
on interchange circuit BA transmitted data..
A transition from on to off instructs the DCE to complete the transmission of all data which were previously transferred across the interface point
on interchange circuit BA and then assume a nontransmit mode or a receive
mode as appropriate. The DCE responds to this instruction by turning off
circuit CB clear to send. when it is prepared to again respond to a
subsequent on condition of circuit CA request to send..
Note: A nontransmit mode does not imply that all line signals have been
removed from the communications channel.
When circuit CA request to send. is turned off, it shall not be turned on
again until circuit CB clear to send. has been turned off by the DCE.
r TIA Interface Standards: EIAr
r TIA-422, and
6.4.2.3 EIAr
r TIA-530, EIAr
r TIA-423. EIArTIA-530 Ref. 14. defines a 25-pin, and alternatively a
EIAr
26-pin, interface between DTE and DCE. It is designed for data rates up to
2.1 Mbps with serial bit streams. The electrical interfaces are described in
EIArTIA-422 Ref. 15. and EIArTIA-423 Ref. 16.. Connector contact
assignments are shown in Table 6.8, and Table 6.9 gives the interchange
circuits.
TABLE 6.9 EIAr
r TIA-530 Interchange Circuits
Circuit
Mnemonics
CCITT
Number
AB
AC
102
102B
Signal common
Signal common
BA
BB
103
104
Transmitted data
Received data
To DCE
From DCE
CA
CB
CF
CJ
CE
CC
CD
105
106
109
133
125
107
108r1, r2
Requested to send
Clear to send
Ready for receiving
Ring indicator
DCE ready
DTE ready
To DCE
From DCE
From DCE
To DCE
From DCE
From DCE
To DCE
DA
113
To DCE
DB
114
DD
115
LL
RL
TM
141
140
142
Transmit signal element timing

(DTE source)
Transmit signal element timing
(DCE source)
Receiver signal element timing
(DCE source)
Local loopback
Remote loopback
Test mode
Circuit Name
Circuit
Direction
Circuit
Type
COMMON
From DCE
DATA
CONTROL
TIMING
From DCE
To DCE
To DCE
From DCE
Source: Reference 14, Figure 4.1, page 17. Courtesy of Electronic Industries AlliancerTelecommunication Industry Association. Reprinted with permission.
115
EIArTIA-422 Ref. 15. and EIArTIA-423 Ref. 16. give the electrical
characteristics for interchange circuits and are companion standards for
EIArTIA-530 Ref. 14.. EIArTIA-422 provides for balanced voltage digital
interface circuits, and EIArTIA-423 provides for unbalanced voltage digital
interface circuits. Remember that an unbalanced connection means that one
signal lead is grounded.
The transition region at the receiver for EIArTIA-422 is between q200
mV and y200 mV. Thus any voltage more positive than q200 mV represents a spacing or 0 condition on condition. and any voltage more negative
Figure 6.14. Balanced digital interface circuit, EIArTIA-422. A, B s generator interface A, B

s load interface; C s generator circuit ground; C s load circuit ground; R s optional cable
termination resistance; Vg s ground potential difference. (From Ref. 17. Copyright Electronic
Industries Alliance.)
Figure 6.15. Unbalanced digital interface circuit, EIAr TIA-423. A, C s generator interface;
A, B s load interface; C s load circuit ground; C s generator circuit ground; Vg s ground
potential difference. (From Ref. 16. Copyright Telecommunication Industry Association.)
116
than y200 mV represents a marking or 1 condition. The nominal signal

voltages at the receiver are y6 and q6 volts DC.
EIArTIA-423 is a similar standard but is used for unbalanced circuits.
The receiver transition range is the same as EIArTIA-422, and the binary 1
and binary 0 significant states are the same. The nominal data rate maximum
is 100 kbps.
Figure 6.14 shows the EIArTIA-422 balanced digital interface circuit, and
Figure 6.15 illustrates the EIArTIA-423 unbalanced digital interface circuit.
6.5 THE QUESTION OF BANDWIDTH

6.5.1 Bandwidth Versus Bit Rate
When I started in the field of data communications, I always wondered how
many bits per second of data could fit into a unit of bandwidth. The question
still seems to bother a lot of people as my seminar experience bears out. We
first will define bandwidth, at least from one perspective. Then there will be
a very brief review by two well-known theoreticians. We then will delve into
some practical matters and attempt to answer the question.
6.5.2 Bandwidth Defined

Bandwidth provides a vital role in data communications. It can be defined as
the range between the lowest and highest frequencies used for a particular
purpose. For our purposes we should consider bandwidth as the range of
frequencies within which the performance characteristics of a device are
above certain specified limits. For filters, attenuators, amplifiers, transmission lines, and other electromagnetic active and passive equipment, these
limits are generally taken where a signal will fall 3 dB below the average
power. level in the passband, or below the level at a reference frequency.
The voice channel is one notable exception. In North America it is specifically defined at the 10-dB points about the reference frequency of 1000 Hz,
approximately in the band 2003300 Hz. The CCITT voice channel occupies
the band 3003400 Hz.
The 3-dB bandwidth concept illustrated in Figure 6.16.
6.5.3 First and Second Bandwidth Approximations

If we remain in the binary domain i.e., we deal only with two states called 1s
and 0s., we often will approximate bandwidth as 1 bit per hertz of bandwidth.
This first approximation roughly allows for nominal impairments such as
amplitudefrequency response, phase distortion, and noise. It is a rough
approximation.
THE QUESTION OF BANDWIDTH
117
Figure 6.16. The concept of the 3-dB power bandwidth.
H. Nyquist of Bell Telephone Laboratories Ref. 17. was a pioneer in

information theory. He essentially stated that theoretical error-free transmission of two bits of information can be related to the 1 Hz of bandwidth. This
is often called the Nyquist rate or Nyquist bandwidth. In other words, if we
have 1000 Hz of bandwidth, theoretically it can transport 2000 bps errorfree.
6.5.4 Shannon
Claude Shannon, also of Bell Telephone Laboratories, came up with a theory
contradicting* Nyquists 2-bit rule Ref. 18.. Shannons traditional formula
relates channel capacity C binary digits or bitsrsecond . to bandwidth W
with the following relationship:
C s W log 2 1 q SrN .
where SrN is the signal-to-noise ratio, which, in this case, must be expressed
as a numeric. If SrN were 30 dB, its numeric equivalent would be 1000. Let
the bandwidth be 3000 Hz, then
C s 3000 = 10 approximately .
s 30,000 bps
Often we will hear of the Shannon limit in reference to the above relationship. As we mentioned previously, one of the principal impairments in the
transmission channel is phase distortion. It is mainly this impairment that
*Contradicting may be a little abrupt. Where Nyquist basically considered intersymbol interference, Shannons only parameter was signal-to-noise ratio.
118
restricts us from reaching the Shannon limit. In a bandwidth-limited channel,

as the bit rate increases beyond a certain point, intersymbol interference
begins to impact bit error rate.
6.5.5 Summary
As we describe in the next chapter, we can pack more bits per hertz of
bandwidth by turning to multilevel systems or M-ary techniques, rather than
binary techniques.
Thus, we can turn to the original idea of 1 bit per hertz for binary
transmission, admitting it is a very rough first approximation. We warn that
the conventional habit of equating bandwidth to bit rate can be fraught with
danger.
REFERENCES
1. The IEEE Standard Dictionary of Electrical and Electronic Terms, 6th ed., IEEE
Std. 100-1996, IEEE, New York, 1996.
2. Interface Between Data Terminal Equipment and Data Circuit Terminating Equipment Employing Serial Binary Data Interchange, EIArTIA-232F, Telecommunication Industry Association, Washington DC, 1997.
3. 9600 Bits per Second Modem Standardized for Use of Point-to-Point 4-Wire Telephone-Type Circuits, CCITT Rec. V.29, Vol. VIII.1, IXth Plenary Assembly,
Melbourne, 1988.
4. Attenuation Distortion, CCITT Rec. G.132, IXth Plenary Assembly, Melbourne,
1988.
5. Roger L. Freeman, Telecommunication Transmission Handbook, 4th ed., John
Wiley & Sons, New York, 1998.
6. Data Communications Using Voiceband Priate Line Channels, Bell System Technical Reference Publication 41004, AT & T, New York, October 1973.
7. Transmission Systems for Communications, 4th ed., Bell Telephone Laboratories,
Holmdel, NJ, 1970.
8. Error Performance of an International Connection Forming Part of an ISDN, CCITT
Rec. G.821, Vol. III, Fascicle III.3, IXth Plenary Assembly, Melbourne, 1988.
9. Error Performance Parameters and Objecties for International Constant Bit Rate
Digital Paths at or aboe the Primary Rate, ITU-T Rec. G. 826, ITU Geneva, 1993.
10. Monsoor Shafi and Peter J. Smith, The Impact of G.826. IEEE Communications
Magazine, September 1993.
11. Transport Systems Generic Requirements: Common Requirements, Bellcorer
Telcordia GR-499-CORE Issue 1, Bellcore Piscataway, NJ, December 1995.
12. Support of Data Terminal Equipment (DTEs) with V-Series Type Interfaces by an
Integrated Serices Digital Network (ISDN), CCITT Rec. V.110, Vol. VIII.1, IXth
REFERENCES
119
13. List of Definitions for Interchange Circuits Between Data Terminal Equipment (DTE)
and Data Circuit-Terminating Equipment (DCE), CCITT Rec. V.24, IXth Plenary
Assembly, Melbourne, 1988.
14. High Speed 25-Position Interface for Data Terminal Equipment and Data CircuitTerminating Equipment, Including Alternatie 26-Position Connector, EIArTIA-530,
Telecommunication Industry Association, Washington DC, May 1992.
15. Electrical Characteristics of Balanced Voltage Digital Interface Circuits, EIArTIA422, Telecommunication Industry Association, Washington DC, December 1978.
16. Electrical Characteristics of Unbalanced Voltage Digital Interface Circuits,
EIArTIA-423, Telecommunication Industry Association, Washington DC,
December 1978.
17. H. Nyquist, Certain Topics in Telegraph Transmission Theory, AIEE Transactions, Vol. 47, April 1928.
18. C. E. Shannon and W. Weaver, The Mathematical Theory of Communications,
University of Illinois Press, Urbana, 1963.
7
THE TELECOMMUNICATIONS
NETWORK AS A VEHICLE
FOR DATA TRANSPORT
7.1 THE PUBLIC SWITCHED TELECOMMUNICATION NETWORK
7.1.1 Introduction
A major means of longer-distance data transport is the public switched
telecommunication network PSTN.. It is essentially ubiquitous and is often
the most cost-effective alternative for such transport. To intelligently decide
when and how to use the PSTN to serve data communication needs, we must
know how the network operates, as well as its advantages and limitations.
The intent of this section is to provide the reader with a review of the
operation of the PSTN from the point of view of a potential data user.
Emphasis is placed on the digital network, both transmission and switching.
The PSTN consists of a user access configuration, local switching, tandem
switching in selected situations with interconnecting trunks, and a longdistance network with its transit exchanges switches .. The PSTN access
configuration primarily consists of wire-pair subscriber loops that enter a
local serving telephone exchange via a main distribution frame. The main
distribution frame facilitates cross-connects. The loops then may be timedivision-multiplexed into groups of 24 or 30 channels in pulse code modulation PCM. configurations if the local exchange is digital. For most digital
exchanges, each switch port accepts such a channel group.
If the local serving exchange is an analog switch, each switch port accepts
one single voice channel. For the analog situation, at some point there will be
conversion to a PCM format to interface with higher levels of the local
network local exchange carrier in the United States . and the toll long
distance . network IXC or interexchange carrier network..
Given a particular desired connectivity, several switches in tandem may be
required to intervene to set up a call. In network design, the trunks that
121
122
THE TELECOMMUNICATIONS NETWORK AS A VEHICLE FOR DATA TRANSPORT
connect switches are called links. CCITT Rec. E.171 Ref. 1. states that there
shall not be more than 12 links in tandem for an international connection.
The value of 12 is broken down into four links for the international portion
of the connection, four for the national side on one end, and four for the
national side on the other end of a call. Each time a link is added to a
connection, the quality of service QoS. tends to deteriorate. For digital
connections, the deterioration is a degradation in bit error rate BER.. For
this reason, the number of links in tandem which make up a telecommunication connectivity must be limited.
It should be noted that for voice operation over the digital PSTN, the
requirements for BER are far less stringent than those for data operation.
Voice operation requires a BER of - 1 = 10y3 , whereas for data operation
the value is better than 1 = 10y6 .
Our concerns here are as follows:
1. What services are available on the PSTN for the data end user?
2. How, and to what extent, will the PSTN degrade our data signal? This
includes other degradations such as loss of traffic, out-of-sequence
delivery, and so on.
3. What limitations are there in data rate and speed of service?
7.1.2 Access to the PSTN: The Subscriber Network
The subscriber plant presents a major constraint to effective data rate. The
subscriber plant consists of wire-pair terminations on the user premises
extending to the local serving exchange. We can expect the characteristic
impedance Z0 . of the typical wire pair as connected to be 600 or 900 ohms.
Ordinarily, wire pair is an excellent transmission medium. It can support
more than 100 Mbps over limited distances. However, our in-place subscriber plant probably cannot support such data rates. We must consider the
following:
1. Whether a particular pair has inductive loading and bridged taps..
Both limit useful bandwidth, add to group delay, and notably constrain
data rate.
2. The age of the portion of the plant in question and the condition of the
loop under consideration. Some plants have been in operation well over
50 years. Cables get wet; they dry and get wet again. Insulation
deteriorates. It has been said that only about 50% of the subscriber
plants in the United States can support the ISDN basic rate 160 kbps
in the United States ..
3. Wire pairs can also be noisy. One cause is the accidental grounding of
one side of the pair. Another cause is the deterioration of balance to
ground.
INTRODUCTION TO DIGITAL NETWORKS
123
If the wire pair extends to an analog switch, we can expect an increase in

impulse noise as well as degradation of both amplitude and phase characteristics of a particular connectivity. Poor grounding in the total wire plant can
be another noise source.
The voice channel deriving from the wire pair at some point will enter
multiplex equipment. As local serving switches are modernized to digital
exchanges, that point will be on PCM channel banks between the switch
subscriber line and main distribution frame or be incorporated at the input of
the switch itself. Once the voice channel enters the channel bank, it is
restricted in frequency by a low-pass filter to the band 3003400 Hz.
Delay is another consideration for the PSTN. If a geostationary. satellite
link is involved in a connectivity, the one-way propagation time to the
satellite is about 125 s and another 125 s for the link to the distant earth
station. This totals 250 s for one-way propagation delay alone. Delay is
troublesome for voice and some image transmission systems. It notably adds
to the latency on data circuits, especially on those that require many
hand-shakes* and turn-arounds. Special problems may arise if a distant data
terminal is emulating a local area network LAN. user. Depending on the
MAC protocol involved, delay may be an overriding consideration.
7.2 INTRODUCTION TO DIGITAL NETWORKS
7.2.1 Rationale
The PSTN on a worldwide basis will shortly be all-digital. How will this
digital network impact the data user? Equally important is the private
network that will most certainly be digital. Many medium- and large-size
enterprises are turning to the development and implementation of their own
private networks. So our question is equally pertinent. It is therefore incumbent upon the data user to be fully knowledgeable of digital networks. That is
what we set out to do in this subsection.
The digital network is based on PCM. We discuss the development of a
PCM signal from the analog voice channel, its coding, the frame, and
higher-level multiplexing, among other factors. We then cover digital network
topologies and impairments. European and North American systems are
reviewed. This is followed by interfacing a data bit stream to the digital
64-kbps channel.
7.2.2 Development of a PCM Signal
When PCM was originally invented Reeves, ITT STL 1937., the only
requirement was to transmit voice speech.. Thus it was built around the
nominal 4-kHz voice channel. With the invention of the transistor, PCM
*Typically on X.25 and IP connectivities unless compensated for. Frame relay is probably the
most forgiving.
124
became a feasible product Shockley, Bell Laboratories 1948. and voice

transmission was still the only requirement. Todays digital network is optimized for speech transmission, and data transmission is a secondary issue.
There are three steps in the development of a PCM signal:
1. Sampling
2. Quantization
3. Coding
Remember that we are going to digitize an analog voice channel. We are
going from a continuously varying amplitude signal analog. to a signal made
up of discrete elements that are digital.
Sampling. The first step is sampling the voice channel. The sampling is
based on Nyquists sampling theorem. It states: If a signal f t . is sampled at
regular intervals and at a rate higher than twice the highest significant signal
frequency, then the samples contain all the information of the original signal.
The function f t . may be reconstructed from these samples by the use of a
low pass filter.
If we were to sample the nominal 4-kHz voice channel, the sampling rate
would be 8000 samples per second because the highest significant frequency
is 4000 Hz i.e., 4000 Hz = 2.. For a 15-kHz program channel, the sampling
rate would be 30,000 times a second i.e., 15,000 = 2.. An NTSC video
channel, if sampled at the Nyquist rate, would be sampled 8.4 = 10 6 times a
second. The NTSC TV video channel has a nominal bandwidth of 4.2 MHz.
The output of the sampler is a pulse amplitude modulation PAM. wave.
Such a wave, which is the result of sampling a simple sinusoid, is shown in
Figure 7.1.
With some notable exceptions, practical PCM systems involve time division multiplexing TDM. carried out in the sampling process. Sampling in
these cases does not involve just one voice channel, but many. In practice,
one system to be discussed samples 24 voice channels in sequence T1., and
another samples 30 voice channels E1.. The result is a more complex PAM
wave than is shown in Figure 7.1.
Figure 7.1. A PAM wave as a result of sampling a simple sinusoid.
125
Figure 7.2. Simplified diagram of the formation of a PAM wave and a PCM bit stream. Only
three channels are sampled sequentially in this example. ( a ) Transmitter, ( b ) receiver.
(Courtesy of GTE Lenkurt Inc., San Carlos, California.)
A simplified diagram of the processing involved is shown in Figure 7.2.

Here just three channels are sampled sequentially.
A sample is simply a measurement of the voltage level of a channel at a
moment in time. This is done by gating. For the case of a 24-channel system,
samples are taken sequentially 8000 times per second. The time it takes to
sample 24 channels just once is 125 s. If we are to sample each channel
8000 times per second, the time to sequence one sampling of 24 channels is
1r8000 s, or 125 s. If we are to sample 24 channels, the gate could be open
as long as ; 5.2 s per channel 125r24 s; 5.2.. Sequencing through each
126
channel just once is called a frame. And, of course, a frame is 125 microseconds in duration.
Quantization and Coding. The next step in the formation of a PCM signal is
quantization. In modern PCM systems, coding is carried out simultaneously
with the quantization process. Realize that quantization is simply setting
discrete voltage level possibilities. Each voltage level is assigned a binary
code, a sequence of 1s and 0s.
The number of bits required to represent a character is called a code
length, or, more properly, a coding leel. For instance, as we have learned in
Chapter 5, a binary code with four discrete binary elements a four-level
code. can code 2 4 separate and distinct meanings, or 16 characters 2 4 s 16.;
a five-level code would provide 2 5, or 32, distinct characters or meanings.
In modern PCM, we use an eight-level code deriving 256 different code
possibilities. Thus we must quantize the sample voltage levels into just 256
possibilities.
The PAM wave has an infinite number of code level voltage level.
possibilities. For all intents and purposes it is still an analog signal. If the
positive excursion of a PAM wave is between 0 and q1 volt, the reader
should ask: How many discrete values are there between 0 and 1? All values
must be considered, including 0.4578046 volts and an infinite number of
other values.
The intensity range of voice signals over an analog telephone channel is of
the order of 50 dB. The y1 to 0 to q1 volt of the PAM wave at the
excursion coder input should represent that 50-dB range. Furthermore, we
defeat the purpose of digitizing if we assign an infinite number of values to
satisfy every level encountered in the 50-dB range or a range from y1 to
q1 volt.. The key, of course, is to assign discrete levels from y1 volt to q1
volt the 50-dB range..
Actually, the assignment of these discrete values is what we call quantization. To cite an example, consider Figure 7.3. Between y1 and q1 volt, 16
quantum steps exist and are coded as shown in Table 7.1.
Figure 7.3 shows that step 12 is used twice, but in neither instance is it the
true value of the impinging sinusoid. It is a rounded-off value. These
rounded-off values are shown in the figure with a dashed line that follows the
general outline of the sinusoid. The horizontal dashed lines show the points
where the quantum step changes to the next higher or lower level if the
sinusoid curve is above or below that value. Take step 14, for example. The
curve, dropping from its maximum, is given two values of 14 consecutively.
For the first, the curve is above 14, and for the second, below. That error, in
the case of 14 for instance, from the quantized value to the true value is
called quantizing distortion or quantization noise. This distortion or noise is
a major form of imperfection found in PCM systems.
In Figure 7.3, maintaining the y1 to 0 to q1 volt relationship, let us
double the number of quantum steps from 16 to 32. What improvement
127
128
TABLE 7.1 Sixteen Coding Steps and Values

Step
Code
Step
Code
0
1
2
3
4
5
6
7
0000
0001
0010
0011
0100
0101
0110
0111
8
9
10
11
12
13
14
15
1000
1001
1010
1011
1100
1101
1110
1111
would we achieve in quantization distortion? First determine the step increment in millivolts in each case. The total excursion is 2000 millivolts i.e.,
y1000 to 0 to q1000 mV.. This excursion is divided into 16 steps, or 125 mV
per step 2000r16 s 125.. In the second case we have 32 steps, thus 2000r32
or 62.5 mV per step. Or we have 62.5 mV above or below the nearest
quantizing step. For the 32-step case, the worst quantizing distortion case
would be at the half-step level or 31.25 mV, whereas for the first case it was
62.5 volts at mid-step level.
Thus the improvement in decibels for doubling the number of quantizing
steps is
20 log 62.5r31.25 . s 20 log 2 s 6 dB
Remember it is 20 log because we are in the voltage domain..
This expression is valid for linear quantization only. Thus, increasing the
number of quantizing steps for a fixed range of input values reduces quantization distortion accordingly. Experiments with linear quantization have
shown that if 2048 uniform steps are provided, sufficient voice signal quality
is achieved.
For the 2048-quantum-step samples. case, an 11-element code is required
2 11 s 2048.. With a sampling rate of 8000 samples per second per voice
channel, binary information rate per voice channel will be 88,000 bps.
Consider that equivalent bandwidth is a function of the information rate
data rate.; the desirability of reducing this figure i.e., 11 bits per sample. is
obvious when bandwidth is at a such a premium.
There are two different types of PCM systems in the world: the European
system with its E1 family of waveforms, and the North American system with
its T1 DS1. family of waveforms.
The North American PCM system uses 256 quantum steps for PCM
voltage samplesexcept every sixth frame, where only 128 quantum steps
are utilized. Each sample consists of an 8-bit PCM word. So a quantum
step is represented by an 8-bit word except at every sixth frame where only a
7-bit word is used. The 8th bit in that frame is reserved for signaling.
129
European systems use 8-bit words for all samples, allowing a full 256
quantum steps.
How do we reduce 2048 quantum steps to only 256? It was mentioned
above that 2048 quantum steps would be required for sufficient voice fidelity,
assuming linear quantization. The reduction is accomplished by companding.
Companding is a contraction for compression and expansion. In conventional telephone practice, a compressor compresses the intensity of speech
signals at the input circuit of a communication channel by imparting more
gain to weak signals than to strong signals. At the far-end output of the
communication circuit, the expander performs the reverse function. It restores the intensity of the signal to its original dynamic range.
The standard approach, both European and North American, is to impart
larger quantizing steps to high-level signals and smaller steps for lower-level
signals. The system takes advantage of the fact that for speech transmissions
there is a much greater likelihood of encountering signals with small amplitudes than those of large amplitudes.
Figure 7.4. Quantizing and coding used with the European E1 PCM system.
130
If we allow a range of 0 to q1 volt to cover all input signal amplitude

ranges in the positive direction see Figure 7.4., then from 0 to 1r64th-volt
the signal amplitude will have 32 step gradations available for assignment and
the range from 12 volt to 1.0 volt will have only 16 step gradations.
It can be appreciated that this nonuniform companding is carried out
during the coding process. The compression and expansion functions
are logarithmic and follow one of two companding laws: the A-law and the
-law. The -law is used for the North American PCM system, and
the A-law is used for the European system. As we can see in Figures 7.4 and
7.6, the companding circuitry does not provide an exact replica of a logarithmic curve. The circuitry produces approximate equivalents using a segmented
curve, with each segment being linear.
To help understand how a PCM system works, consider that the near-end
transmitter simply measures the voltage level of the signal sample at a known
moment in time. It transmits this value as an 8-bit message to the distant
receiver which reconstructs the signal with this information. Of course the
signal is reconstructed over a series of many samplesin fact, 8000 samples
in one second. The 8-bit message is the PCM word. A PCM word occupies
a time slot. The concept of the time slot is important to understand for data
communications. In one method of data transmission, the 8-bit sample is
replaced by an 8-bit data byte or octet.
Lets return to Figure 7.4 to learn how the distant end receiver interprets
the 8-bit message it receives. By inspection of the figure, one will note that all
messages 8-bit sequences . starting with a binary 1 represent positive voltages; all negative voltages start with a binary 0 and are below the ordinate
of the curve.
Figure 7.4 has 13 linear segments closely approximating the A-law logarithmic curve, where A s 87.6. However, it should be noted that the segment
passing through the origin has four steps: two above the ordinate and two
below. Thus, really, a 16-segment representation of the curve is used.
As discussed above, the first bit of an 8-bit sequence message tells us if the
sample is a positive or negative voltage. The next three bits of the sequence
indicate the segment where the sample lies. The last four bits of the sequence
tell us where in the segment the voltage point is located.
Suppose the receiver reads an 8-bit sequence as 1101XXXX, where
XXXX are the last 4 bits. Reading from left to right, the first bit a 1. tells
us the sample is a positive voltage; it lies above the ordinate. The next three
code elements bits. indicate that it is the fourth step, or:
0
1
2
3
4
5
6
1000 and 1001

1010
1011
2200
1101
1110
1111
131
Figure 7.5. Coding of segment 4 (positive) of the E1 system. Dashed line shows position of
sample 8-bit sequence.
Figure 7.5 shows a blowup of the uniform quantizing and subsequent

straightforward coding of step 4; it illustrates the final segment coding. Note
that the quantum steps are of uniform size inside that segment, providing 16
2 4 s 16. coded quantum steps.
The North American DS1 T1. system uses a 15-segment approximation of
the logarithmic -law. Again, there are 16 segments. The segments cutting
the origin are collinear and are counted as 1, but are actually two: one above
and one below the ordinate. This leads to the 16 value i.e., there are 16
segments.. See Figure 7.6.
The quantization of the DS1 system is shown in Figure 7.6 for the positive
portion of the curve. Segment 5 representing quantum steps 64 to 80 is shown
blown up in the figure. Table 7.2 shows the DS1 T1. coding. As can be seen
again, the first code element, whether a 1 or a 0, indicates whether the
quantum step is above or below the horizontal axis the ordinate.. As with
E1, the next three code elements identify the segment, and the last four code
elements bits. identify the actual quantum level voltage point. inside the
segment.
7.2.3 The Concept of Frame
As illustrated in Figure 7.2, PCM multiplexing is carried out in the sampling
process, sampling several sources sequentially. These sources may be nominal
4-kHz voice channels or other analog information sources such as a data tone
or video. The final result of the sampling and subsequent quantization and
coding is a series of pulses, a serial bit stream that requires some indication
or identification of the beginning of a scanning sequence. This identification
tells the far-end receiver when each full sampling sequence starts and ends; it
132
Figure 7.6. Positive portion of the segmented approximation of the -law ( s 255) quantizing curve used in the North American DS1 PCM system.
times and aligns the receiver. Such identification is called framing. A full
sequence or cycle of samples is called a frame in PCM terminology.
Note that there are only 255 quantizing steps because steps 0 and 1 use
the same bit sequence, thus avoiding a sequence of all zeros no transitions ..
CCITT Rec. G.701 Ref. 2. defines a frame as a set of consecutive digit
time slots in which the position of each digit time slot can be identified by
reference to a frame alignment signal. The frame alignment signal does not
necessarily occur, in whole or in part, in each frame.
The North American PCM system carries out frame alignment in a
different way from the E1 system. As its European counterpart, the North
American system uses 8-bit time slots for each sample and samples 24
channels sequentially. Sampling each of the 24 channels just once sequentially is a frame. To each frame one bit is added, called the framing bit. One
frame therefore has 193 bits, or
8 = 24 q 1 s 193
133
TABLE 7.2 Eight-Level Coding of the North American DS1 (T1) System
Code Level
255
239
223
207
191
175
159
143
127
126
111
95
79
63
47
31
15
2
1
0
a
(Peak positive level)
(Center levels)
(Nominal zero)
(Peak negative level)
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
1
1
0
0
1
1
1
1
1
1
0
0
1
1
0
0
0
0
0
1
0
1
0
1
0
1
1
1
1
0
1
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
1
1
1a
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
1
0
0
One digit is added to ensure that the timing content of the transmitted pattern is maintained.
The framing rate i.e., sampling rate. is 8000 times per second. This is the
so-called Nyquist rate. To compute the bit rate for the North American
DS1 T1. system, multiply 8000 by 193:
193 = 8000 s 1,544,000 bps, or 1.544 Mbps
The E1 system is a 32-channel system where 30 channels carry speech on
voice channels and the remaining two channels carry supervisory signaling
and framingrsynchronization information. Each channel is an 8-bit time slot,
and we speak of time slots 031 as follows:
Time Slot Number
Type of Information or Function
0
115
16
1731
Synchronizing framing.
Speech
Signaling
Speech
In time slot 0 a synchronizing code or word is transmitted every second

frame, occupying digits 28 as follows:
0011011
In those frames without the synchronizing word, the second bit of time slot 0
is frozen at a 1 so that in these frames the synchronizing word cannot be
134
Figure 7.7. Neutral versus bipolar bit streams. Top: Alternating 1s and 0s transmitted in a
neutral mode. Bottom: Alternating 1s and 0s transmitted with alternate mark inversion (AMI),
also called bipolar.
imitated. The remaining bits of time slot 0 can be used for the transmission
of supervisory information signals.
Framing and timing should be distinguished. Framing ensures that the
PCM receiver is aligned regarding the beginning and end. of a sequence or
frame; timing refers to the synchronization of the receiver clockspecifically, that it is in step with its far-end companion. transmit clock. Timing at
the receiver is corrected via the incoming markspace and spacemark.
transitions. It is important, then, that long periods without transitions do not
occur. This was equally important in our discussion of data transmission in
Chapters 5 and 6.
7.2.4 The Line Code
PCM signals transmitted on metallic-pair cable are in the bipolar mode
biternary . as shown in Figure 7.7. The marks or 1s have only a 50% duty
cycle.* There are several advantages to this mode of line transmission:
1. No direct-current return is required; hence transformer coupling can be
used on the line.
2. The power spectrum of the transmitted signal is centered at a frequency equivalent to half of the bit rate.
It will be noted in bipolar transmission that the 0s are coded as absence of
pulses, and the 1s are alternately coded as positive or negative pulses, with
the alternation taking place at every occurrence of a 1. This mode of
transmission is more often called alternate mark inersion AMI..
*With NRZ bit streams, which we have assumed up to now, each bit has 100% duty cycle. Here
we mean that the active part of a bit occupies the entire time slot assigned. In the case of AMI
bipolar. transmission, the 1s occupy only half of the slot assigned.
135
One drawback to straightforward AMI transmission is that when a long

string of 0s is transmitted i.e., there are no transitions ., a timing problem
may come about because repeaters and decoders have no way of extracting
timing without transitions. The problem can be alleviated by forbidding long
strings of 0s. Codes have been developed that are bipolar, but with N zeros
substitution; they are called BNZS codes. For instance, a B6ZS code substitutes a particular signal for a string of six 0s.
Another such code is the HDB3 high-density binary-3., where the 3
indicates that it substitutes for binary formations with more than three
consecutive 0s. With HDB3 the second and third zeros of the string are left
unchanged. The fourth 0 is transmitted to the line with the same polarity of
the previous mark sent, which is a violation of the AMI concept. The first 0
may or may not be modified to a 1 to ensure that the successive violations are
of opposite polarity. HDB3 is used with the European E-1 system.
7.2.5 Regenerative Repeaters

Pulses passing down a digital transmission line or over a radio system suffer
attenuation and are badly distorted by the frequency characteristic of the line
and, in the case of radio, by the transmission medium and filters in the radio
equipment. A regeneratie repeater amplifies and reconstructs such badly
distorted digital signals and develops a nearly perfect replica of the original
signal at its output. Regenerative repeaters are the essential key to digital
transmission in that we can say that the noise stops at the repeater. The
principal advantage of digital systems over their analog counterparts is that
noise does not accumulate over a digital system.
Figure 7.8. Simplified functional block diagram of a regenerative repeater for use with PCM
metallic cable systems.
136
Figure 7.8 is a simplified block diagram of a regenerative repeater and

shows typical waveforms corresponding to each functional stage of signal processing. As shown in the figure, the first stage of signal processing is amplification and equalization. Equalization is often a two-step process.
The first step is a fixed equalizer that compensates for the attenuationfrequency characteristic of the nominal section, which is the standard length
of transmission line between repeaters often 6000 ft or 1850 m.. The second
equalizer is variable and compensates for departures between nominal repeater section length and the actual length and loss variations due to
temperature. The adjustable equalizer uses automatic line build-out ALBO.
networks that are automatically adjusted according to characteristics of the
received signal.
The signal output of the regenerative repeater must be accurately timed to
maintain specified pulse width and space between pulses. The timing is
derived from the incoming bit stream. The incoming signal is rectified and
clipped, producing square waves that are applied to the timing extractor,
which is a circuit tuned to the timing frequency. The output of the circuit
controls a clock-pulse generator that produces an output of narrow pulses
that are alternatively positive and negative at the zero crossings of the
square-wave input.
The narrow positive clock pulses gate the incoming pulses of the regenerator, and the negative pulses are used to run off the regenerator. Thus the
combination is used to control the width of the regenerated pulses.
Regenerative repeaters are the major source of timing jitter in a digital
transmission system. Jitter is one of the principal impairments in a digital
network, giving rise to pulse distortion and intersymbol interference.
Most regenerative repeaters transmit a bipolar AMI. waveform as shown
in Figure 7.7. Such signals can have one of three possible states in any instant
of timepositive, zero, or negativeand are often designated q, 0, and y.
The threshold circuits are gated to admit the signal at the middle of the pulse
interval. For example, if the signal is positive and exceeds a positive threshold, it is recognized as a positive pulse. If it is negative and exceeds a
negative threshold, it is regenerated as a negative pulse. If it has a value
between the positive and negative thresholds, it is recognized as a 0 i.e., no
pulse..
7.2.6 Higher-Order PCM Multiplex Systems

The PCM multiplex hierarchy is built up based on a 24-channel DS1. or
30-channel E1. group. By channel we mean the digital voice channel. These
basic groups are called level 1 in the hierarchy. Subsequent multiplex levels
are then developed i.e., levels 2, 3, and 4; and in two cases, level 5.. These
are not to be confused with the synchronous optical network SONET. and
the synchronous digital hierarchy SDH. levels covered in Chapter 15. It
should be noted that the DS1 T1. and E1 families of digital multiplex are
137
often called the plesiochronous digital hierarchy PDH., whereas SONET and
its European counterpart, SDH, are called the synchronous digital hierarchy.
Table 7.3 summarizes and compares these multiplex levels for the North
American, Japanese, and European systems. The development of the North
American hierarchy is shown in Figure 7.9, which gives respective DS line
rates and multiplex nomenclature. Regarding nomenclature, we see from the
figure that M12 accepts level 1 inputs, delivering level 2 outputs to the line. It
TABLE 7.3 PCM Multiplex Hierarchy Comparison a
Level
System Type
North America
Number of voice channels
Line bit rate (Mbps)
1
24
1.544
2
96
6.312
3
672
44.736
4
4032
274.176
Japan
24
1.544
96
6.312
480
32.064
1440
97.728
5760
400.352
Europe
30
2.048
120
8.448
480
34.368
1920
139.264
7680
564.992
See Ref. 3.
Figure 7.9. North American PCM hierarchy and multiplexing plan.
138
actually accepts 4 DS1 inputs deriving a DS2 output 6.312 Mbps.. M13
accepts level 1 inputs, delivering a level 3 output to the line. In this case, 28
DS1 inputs form one DS3 output 44,736 Mbps.. The M34 takes six DS3
inputs level 3. to form one DS4 line rate 274.176 Mbps.. DS1C is a special
case where two 1.544 Mbps DS1 rates are multiplexed to form a 48-channel
group with a line rate of 3.152 Mbps.
By simple multiplication we can see that the higher-order line rate is a
multiple of the lower inputs rate plus some number of bits. The DS1C is an
example. Here the line rate is 3152 kbps, which is 2 = 1544 kbps q 64 kbps.
The additional 64 kbps are used for multiplex synchronization and framing.
Multiplex and demultiplex timing is very important, as one might imagine.
The two DS1 signal inputs are each 1.544 Mbps plus or minus some speed
tolerance specified as "130 ppm.. The two input signals must be made alike
in repetition rate and in a rate suitable for multiplexing. This is done by bit
stuffing. In this process, bits pulses. are added to each signal in sufficient
quantity to make the signal operate at a precise rate controlled by a common
clock circuit in the multiplex. Pulses are inserted stuffed . into multiplex slots
but carry no information. Thus it is necessary to code the signal in such a
manner that these noninformation bits can be recognized and removed at the
receiving demultiplex terminal.
Consider the more general case, using CCITT terminology. If we wish to
multiplex several lower-level PCM bit streams deriving from separate tributaries into a single PCM bit stream at a higher level, a process of justification
is required called bit stuffing above.. In CCITT Rec. G.701 Ref. 2. we find
the following definition of justification: a process of changing the rate of a
digital signal in a controlled manner so that it can accord with a rate different from its own inherent rate usually without loss of information.
Positive justification as above. adds or stuffs digits; negative justification
deletes or slips digits. Stuffing is a North American term; justification is
European. If we think of computer displays and print copy, a page may be
right-hand justified, left-hand justified, or right- and left-justified as this page
TABLE 7.4 North American DS Series of Line Rates, Tolerances,

and Line Codes (Format)
Signal
Repetition Rate
(Mbps)
Tolerance
(ppm) a
DS0
DS1
DS1C
DS2
DS3
DS4
0.064
1.544
3.152
6.312
44.736
274.176
Parts per million.

Expressed in terms of slip rate.
"130
"30
"30
"20
"10
Format
Duty Cycle
(%)
Bipolar
Bipolar
Bipolar
B6ZS
B3ZS
Polar
100
50
50
50
50
100
A BRIEF OVERVIEW OF DIGITAL SWITCHING
139
TABLE 7.5 Summary of CEPT30 H 2-Related Line Rates and Codes

Level
Line Data Rate

(Mbps)
Tolerance
(ppm)
Code
Mark Peak Voltage

(V)
1
2
3
4
2.048
8.448
34.368
139.264
"50
"30
"20
"15
HDB3
HDB3
HDB3
CMI b
2.37 or 3 a
2.37
1.0
1 " 0.1c
a
b
c
2.37 V on coaxial pair; 3 V on symmetric wire pair.

Coded mark inversion.
Peak-to-peak voltage.
is. Spacing is added or deleted for such justification. With this in mind,
justification is a better term.
7.2.7 Line Rates and Codes
Table 7.4 summarizes North American DS series of PCM multiplex line
rates, tolerances of these rates, and the types of line code used. Table 7.5
gives similar information for the E1 family of multiplex line rates.
7.3 A BRIEF OVERVIEW OF DIGITAL SWITCHING

7.3.1 Advantages and Issues of PCM Switching
There are both economic and technical advantages to digital switching, and it
is in this context that we refer to PCM switching. The economic advantages
of time-division PCM switching include the following:
1. There are notably fewer equivalent cross-points for a given number of
lines and trunks than in an analog space-division switch.
2. A PCM switch is of considerably smaller size, and it requires less prime
power than its analog counterpart.
3. It has much more common circuitry i.e., common modules..
4. It is easier to achieve full availability* within economic constraints.
The technical advantages include the following:
1. It is regenerative i.e., the switch does not distort the signal; in fact, the
output signal is cleaner than the input..
2. It is noise-resistant.
*Full aailability means that each and every switch inlet can connect to each and every switch
outlet.
140
3. It is computer based and thus incorporates all the advantages of SPC

stored program control..
4. The binary message format in principle . is compatible with digital
computers and related data communications. It is also compatible with
signaling.
5. A digital exchange is lossless. There is no insertion loss as a result of a
switch being inserted in the network.
6. IT exploits the continuing cost erosion of digital logic and memory: LSI,
VLSI, and VHSIC insertion.
Two technical issues may be listed as disadvantages:
1. A digital switch degrades error performance of the system. A well-designed switch may only impact network error performance minimally,
but it still does it.
2. Switch and network synchronization and the reduction of jitter and
wander can be gating issues in system design.
7.3.2 Approaches to PCM Switching
A digital switchs architecture is made up of two elements called T and S for
time-division switching T. and space-division switching S.. We can describe
a digital switch architecture with a sequence of Ts and Ss. For example, the
AT & T No. 4ESS is a TSSSST switch, GTE No. 3 EAX is an SSTSS switch,
and the Northern Telecom DMS 100r200 is TSTS-folded. Some switches
designed to serve local networks are simply TST switches e.g., AT & Ts
5ESS.. Let us now examine these basic elements, the time T. and space S.
switch networks.
7.3.3 Time Switch
In a most simplified way, Figure 7.10 is a time switch, or what many call a
time slot interchanger. From Section 7.2 we know that a time slot represents 8
bits, the 8-bit sample of voice level for speech operation. The duration of the
PCM frame is 125 s. For the DS1 format the basic frame contains 24 time
Figure 7.10. A time-division switching showing time slot interchange. Connectivity is from
incoming time slot C to user G (outgoing time slot G ).
141
slots, and the E1 format contains 32 time slots. The time duration of a time
slot for DS1 T1. is 125r24 s 5.2803 s, and for E1 it is 125r32 s 3.906 s.
Time slot interchanging involves moving the information contained in each
time slot from the incoming bit stream location to an outgoing bit stream, but
in a different location. The outgoing location of the time slot is in accordance
with the destination of the time slot, where the information contained in the
time slot may be voice, data, video, or facsimile. Of course, at the output port
of the switch a whole new frame is generated for transmission. To accomplish
this, at least one time slot must be stored in memory write. and then called
out of memory in a changed sequence position read.. The operations must
be controlled in some manner; and some of these control actions must also
be kept in memory, such as the idle and busy condition of time slots.
Now we can identify three basic blocks of a time switch:
1. Memory for speech by time slots.
2. Memory for control
3. Time slot counter or processor
These three blocks are shown in Figures 7.11 and 7.12. The incoming time
slots can be written into the speech memory either sequentially i.e., as they
appear in the incoming bit stream. Figure 7.11. or randomly Figure 7.12.
where the order of appearance in memory is the same as the order of
Figure 7.11. Time slot interchange time switch; sequential write, random read. (From Ref. 4.)
142
Figure 7.12. Time switch, time slot interchanger, random write, sequential read. (From
Ref. 4.)
appearance in the outgoing bit stream. Now with the second version the
outgoing time slots are read sequentially because they are in the proper order
of the outgoing bit stream i.e., random write, sequential read.. This means
that the incoming time slots are written into memory in the desired output
order. The writing of incoming time slots into speech memory can be
controlled by a simple time slot counter and can be sequential i.e., in the
order in which they appear in the incoming bit stream Figure 7.11.. The
readout of the speech memory is controlled by the control memory. In this
case the readout is random where the time slots are read out in the desired
output order. The memory has as many cells as there are time slots. For the
DS1 example, there would be 24 cells. This time switch works well for a
single inletoutlet switch.
Consider a multiple port switch, which may be a tandem switch, for
example, handling multiple trunks. Here time slots from an inlet port are
destined for multiple outlet ports. Enter the space switch S .. Figure 7.13 is a
simple illustration of this concept. For example, time slot B1 is moved to the
Z trunk into time slot Z1 , and time slot Cn is moved to trunk W into time
slot Wn . However, we see that in each case of a space switch, there is no
change in time slot position.
7.3.4 Space Switch
A typical time-division space switch is shown in Figure 7.14. It consists of a
cross-point matrix made up of logic gates that allow the switching of time
143
Figure 7.13. Space switch allows time slot interchange among multiple trunks. (From Ref. 4.)
Figure 7.14. Time-division space switch cross-point array. (From Ref. 4.)
144
slots in the PCM time slot bit streams in a pattern required by the network
connectivity for that moment in time. The matrix consists of a number of
input horizontals and a number of output verticals with a logic gate at each
cross-point. The array, as shown in the figure, has m horizontals and n
verticals, and we call it an m = n array. If m s n, the switch is nonblocking.
If m ) n, the switch concentrates; and if n ) m, the switch expands.*
Sometimes other notation is used in the literature, such as n and k.
In Figure 7.14, the array consists of a number of m. input horizontals and
n. output verticals. For a given time slot, the appropriate logic gate is
enabled and the time slot passes from its input horizontal to the desired
output vertical. The other horizontals, each serving a different serial stream
of time slots, can have the same time slot e.g., a time slot from time slots
124, 130, or 1n; for instance, time slot 7 on each stream. switched into
other verticals by enabling their gates. In the next time slot position i.e., time
slot 8., a completely different path configuration could occur, again allowing
time slots from horizontals to be switched to selected verticals. The selection,
of course, is a function of how the traffic is to be routed at that moment for
calls in progress or calls being set up. The space array cross-point matrix.
does not switch time slots in the time domain as does a time switch time slot
interchanger .. This is because the occurrences of time slots are identical on
the horizontal and on the vertical. The control store memory. in Figure 7.14
enables the cross-point gates in accordance with its stored information.
In Figure 7.14, it is desired to transmit a signal from input 1 horizontal. to
output 2 vertical., and the gate at that interface would be activated by
placing an enable signal on S12 during the desired time slot. Then the
information bits of that time slot would pass through the logic gate onto the
vertical. In the same time slot, an enable signal Sm1 on the mth horizontal
would permit that particular time slot to pass to vertical. From this we can
see that the maximum capacity of the array during one time slot interval
measured in simultaneous call connections is the smaller value of m or n.
For example, if the array is 20 = 20 and a time slot interchanger is placed on
each input horizontal line. and the interchanger handles 32 time slots, the
array can serve 20 = 32 s 640 different time slots. The reader should note
how the time slot interchanger TSI. multiplies the call-handling capability of
the array when compared to its analog counterpart.
7.3.5 TimeSpaceTime Switch
Digital switches are composed of time- and space-switching stages. The letter
S designates a space-switching stage, and the letter T designates a timeswitching stage. For instance, a switch composed of a time-switching stage, a
space-switching stage, and a time-switching stage is called a TST switch.
*Concentration and expansion are the terms used, particularly for local switches, where there
are notably fewer trunks than lines.
145
Figure 7.15. A timespacetime (TST) switch. TSI, time slot interchanger.
A switch consisting of a space-switching stage, a time-switching stage, and a

space-switching stage is called an STS switch. As we discussed previously,
there are other examples such as the AT & T No. 4ESS switch, which is a
TSSSST switch.
Figure 7.15 illustrates the timespacetime TST. concept. The first stage
of the switch is a time slot interchanger TSI. or time stage which interchanges information between external incoming channels and the subsequent
space stage. Better said, there is a TSI on every input port, and it reassembles time slots on each bit stream into a desired order prior to placing them
on the horizontals of the space array the space stage.. The space stage
provides connectivity between time stages at the input and output ports. We
saw earlier that space-stage time slots need not have any relation to either
external incoming or outgoing time slots regarding number, numbering, or
position. For instance, an incoming time slot 4 can be connected to outgoing
time slot 19 via space network time slot 8.
If the space stage of a TST switch is nonblocking, blocking in such a switch
can still occur if there is no internal space-stage time slot during which the
link from the inlet time stage and the link to the outlet time stage are both
idle. The blocking probability can be minimized if the number of space-stage
time slots is large. A TST switch is strictly nonblocking if
1 s 2c y 1
where 1 is the number of space-stage time slots and c is the number of
external TDM time slots 8..
146
Figure 7.16. Typical STS switch architecture.
7.3.6 SpaceTimeSpace Switch

A spacetimespace STS. switch reverses the architecture of a TST switch.
The STS switch consists of a space cross-point matrix at the input followed by
an array of time slot interchangers whose outputs feed another cross-point
matrix. Such a switch is illustrated in Figure 7.16. Consider this operational
example with an STS switch. Suppose that incoming time slot 5 must be
connected to output time slot 12. This requires a time slot interchanger that
is available to interchange time slots 5 and 12. This can be accomplished by n
time slot interchangers in the time stage.
7.4 DIGITAL NETWORK STRUCTURE

In the United States and Canada, the long-distance PSTN network is now
all-digital. The connecting network that provides subscriber access LECs*
and independent telephone companies in the United States . will become
99.9% digital by 2002.
The network topology is fairly much the same as its prior analog counterpart. The trend is now away from a multilayer hierarchical structure to one
that may be considered either two- or three-layer hierarchy in the United
States and moving toward dynamic nonhierarchical routing DNHR.. All
reference to hierarchical structures was removed from CCITT in the 1988
Plenary Assembly. There is little to lead one to believe that the reason for
*LEC stands for local exchange carrier the local telephone company..
DIGITAL NETWORK IMPAIRMENTS AND PERFORMANCE REQUIREMENTS
147
this is because of the change to digital. It is probably because of the

improvements in transmission which have taken place in the past 30 years,
including vast implementations of both satellite communications and fiberoptic links. The structure of the U.S. network probably has more to do with
divestiture of the Bell SystemrAT & T and the advent of competing long-distance carriers. Blocking probability objectives remain essentially the
samethat is, a probability of 1 in 100 of encountering blockage during the
busy hour BH.. The busy hour is the hour of highest traffic density during a
working day.
7.5 DIGITAL NETWORK IMPAIRMENTS AND PERFORMANCE
REQUIREMENTS
7.5.1 Error Performance
Error performance requirements of the digital network are driven by the
needs of data users. The trend is toward bit-by-bit mapping of data into
digital bit streams. When this is the case, there is also a bit-by-bit impact on
data error rate. In other words, the data channel error rate will be the same
as the digital transport channel. We must also keep in mind that errors
accumulate across a digital built-up connection. Thus error rates must be
specified on an end-to-end basis to satisfy the data end user.
There are many definitions of links in telecommunications, but if we
define a link as that portion of the network between digital regenerators, we
will specify the BER for such a link as 1 = 10y9 or now even 1 = 10y10 . We
must do this to meet the error performance criteria of Section 6.3.3.
Suppose we had 10 such links in tandem which make up a built-up
connection and we specified the error performance of each link as 1 = 10y9.
What would be the error performance at the end of the 10-link connectivity
random errors assumed.? It would be 1 = 10y8 when measured over some
averaging period, say 1 week or even 24 hours. It assumes that each link is of
uniform structure such as a fairly long fiber-optic line and that errors are
random.
7.5.2 Slips
7.5.2.1 Definition. A slip is a synchronization impairment of a digital
network. Slips occur because no two free-running digital network clocks are
exactly synchronized. Consider the simple case where two switches are
connected by digital links. An incoming link to a switch terminates in an
elastic store memory. to remove transmission timing jitter. The elastic store
at the incoming interface is written into at the recovered clock rate R2. but
read out by the local clock rate R1.. If the average clock rate at the
recovered line clock R2. is different from R1. i.e., network synchronization has been lost., the elastic store will eventually overflow or underflow.
148
When R2. is greater than R1., an overflow will occur, causing a loss of
data. When R1. is greater than R2., an underflow occurs causing extraneous data being inserted into the bit stream. These disruptions of data are
called slips.
Slips must be controlled to avoid loss of frame synchronization. The most
common approach is to purposely add an artificial frame or delete a traffic
frame that does not affect framing synchronization and alignment.
Slips do not impair voice transmission per se. A slip to a listener sounds
like a click. The impairments to signaling and framing due to slips depend on
the PCM system, whether it is DS1 or E1. Consider the latter E1.. It uses
channel-associated signaling in channel 16 and may suffer multiframe misalignment due to a slip. Alignment may take up to 5 ms to be reacquired, and
calls in the process of being set up may be lost. Common channel signaling
such as CCITT Signaling System No. 7 is equipped with error detection and
retransmission features so that a slip will cause increased retransmit activity,
but the signaling function will be otherwise unaffected.
One can imagine the effect of slips on data transmission that is riding on
one or more digital 64-kbps channels. At least one octet 8 bits. of data is lost
per channel. If frame alignment is lost, all data are lost, including the slip, for
the period required to reacquire alignment. As an example, consider just one
64-kbps channel. If 5 ms is required including one time slot slip. and it takes
125 s to transmit a frame, the number of lost bits is
8 5 = 10y3 . r125 = 10y6 s 320 bits
If the slip and realignment period occur right in the midst of where one data
frame ends and the next begins, then two frames are lost and must be
recovered. Besides, many data link protocols depend on bit counting to
determine field and frame boundaries. Imagine what would happen to the
counting if we drop 8 bits or add 8 extraneous bits!
7.5.2.2 North American Slip Objectives. The North American end-to-end
slip objective is 0 slips. However, when a higher-level network synchronization clock suffers a fault, a slip rate of 255 slips per day is possible Ref. 5..
This could happen if the master network synchronizing clock is lost and all
subsidiary switch clocks are left free-running.
7.5.2.3 CCITT Slip Objectives. The CCITT slip objective is 1 slip per 70
days per plesiochronous link. Note that plesiochronous operation implies
free-running, high-stability clocks. By high stability, we mean a stability on
the order of "1 = 10y11 per month or better Ref. 6..
The slip objective per link is 1 slip in 5.8 days. This is based on the CCITT
hypothetical digital connection HRX. which consists of 13 nodes interconnected by 12 links. Thus we have 70r12 or 5.8 days. CCITT Rec. G.822 Ref.
7. states that acceptable slip rate performance on a 64-kbps international
DATA TRANSMISSION ON THE DIGITAL NETWORK
149
connection is less than 5 slips in 24 hours for more than 98.9% of the time.
Forty percent of the slip rate is allocated to the local network on each end of
the connection, 6% is allocated for each national transit toll. exchange, and
8% is allocated for the international transit portion.
7.6 DATA TRANSMISSION ON THE DIGITAL NETWORK

7.6.1 The Problem
To transmit binary data on the digital network, two problems arise. One is
the waveform. It is quite simple to fix. Computer data are usually in a serial
format with an NRZ waveform. Pulse code modulation uses AMI alternative
mark inversion, also called bipolar..
The second problem is data rate. Computer data sources have data rates
related to 75 = 2 n, such as 75, 150, 300, 600, 1200, 2400, 4800, 9600, and
14,400 bps, and these rates have no relationship whatsoever to common
digital network transmission rates of 64 kbps or the North American 56
kbps..
7.6.2 Some Solutions
One of the most commonly used and least elegant approaches is to use
analog modems such as those found in the CCITT V series recommendations
discussed in Chapter 8. In essence here, we are impressing, say, 1200 bps on
a 64-kbps DS0 or E0 channel. It seems a waste of valuable bit rate capacity.
Present PCM channel banks can come equipped with a 14.4- or 28.8-kbps
data port. In other words, 14.4 or 28.8 kbps are mapped into a 64r56-kbps
digital channel.* The result again is a waste of bits i.e., 64,000 y 14,400 s
49,600 wasted bits.. Wasted bits in our context are bits not used to transport
user data.
The propensity of data users only have analog line terminations on
premises. They have no other alternative than to use standard analog data
modems such as we describe in Chapter 8 e.g., modems based on ITU-T
Recs. V.34 and V.90..
There are users that have ISDN which provides 64-kbps digital channels at
the desktop. Some PABXs such as Nortel SL-1 permit digital operation
directly to the desktop.
Unfortunately, our standardized data rates by ITU-T and TIA. are based
on the relationship 75 = 2 n, resulting in incompatibility with the digital
64,000-bps data rate. Rate adaptation is required. One approach is to base
the adaptation on CCITT Rec.V. 110 Ref. 8.. A two-step process is used.
*In certain circumstances the North American network can only provide a 56-kbps clear
channel.
150
First the user data rate is converted to an appropriate intermediate rate

expressed by the relation 2 k = 8 kbps, where k s 0, 1, or 2. The next step
performs the second conversion from the intermediate rates. to 64 kbps.
7.6.2.1 The North American Digital Data System (DDS). The North
American DDS provides duplex point-to-point and multipoint private-line
digital data transmission at a number of synchronous data rates. It is based
on the standard 1.544-Mbps DS1 PCM line rate, where individual bit streams
have data rates that are submultiples of that line rate i.e., based on 64 kbps..
However, pulse slots bit slots. are reserved for identification in the demultiplexing of individual user bit streams as well as for certain status and control
signals and to ensure that sufficient line pulses bits. are transmitted for
clock recovery and pulse regeneration. The maximum data rate to the
subscriber for this system is 56 kbps, some 87.5% of the theoretical 64-kbps
maximum.
As we remember, the 1.544-Mbps line signal as applied to the DDS service
consists of 24 sequential 8-bit words i.e., channel time slots. plus one
additional framing bit. The entire sequence is repeated 8,000 times per
second. Note that again we have 8000192 q 1. s 1.544 Mbps, where the
value 192 is 8 = 24 see Section 7.2.2.. Thus the line rate of a DDS facility is
compatible with the DS1 T1. PCM line rate and offers the advantage of
allowing a mix of voice PCM. and data when the full dedication of a DS1
facility to data would be inefficient.
In the descriptive material Refs. 9, 10, and 11. of DDS, the 8-bit word is
called a byte. One bit of each 8-bit byte is reserved for network control and
for stuffing to meet the nominal line rate requirements. This control bit is
called the C-bit. With the C-bit removed, we see where the standard channel
bit rate is derived, namely 56 kbps or 8000 = 7. Four subrates or submultiples data rates are available to the user besides the basic 56 kbps. These are
2.4, 4.8, 9.6, and 19.2 kbps. The actual conversion is carried out in a
CSUrDSU. DDS today is being outpaced by more modern conversion
techniques such as the use of the CCITT Rec. V.110 techniques. Others
returning to the direct use of a 64r56 kbps channel or groups of fractional
DS1 channels all the way to leasing a direct DS1 connection. Usually, on
such a connection only 1.536 Mbps are usable.
7.6.2.2 Bit Integrity in a Time Slot. If we are to map data bits usually an
octet at a time. into PCM time slots and those time slot with data bits are
intermixed in a DS1 with time slots carrying voice samples, a word of caution
is advised. Those time slots carrying voice samples may require A-law to
-law conversion or vice versa, bit integrity will be lost. In other words, in all
probability bit values will be changed, completely destroying the data being
carried. Likewise, if digital padding adding loss. is employed, data integrity
INTERCONNECTS AND BYPASS
151
will also be lost. If the data time slots can be segregated from those carrying
voice samples, then integrity can be maintained in either case.
7.7 INTERCONNECTS AND BYPASS

Bypass has the connotation of bypassing the local telephone company. This
can prove very attractive, particularly to organizations that interexchange
large quantities of data on a daily basis. In North America the driving factor
is economic. Can some form of bypass save money? Another factor is
whether there is some service not available through the local telephone
company or IXC interexchange carrier. which would be available through
bypass.
There are two subsets of bypass: i. bypass by means of a private network
and ii. bypass by means of a specialty carrier, or a specialized common
carrier. Also, alternative access is becoming more and more prevalent in
North America. Here we mean the existence of two or more local telephone
companies, or organizations that offer competing services for a particular
local area.
The most straightforward way to effect bypass is via satellite, followed by
establishing private local network. Most common of these is a VSAT network. VSAT stands for very-small-aperture terminal, a type of satellite
communications network.. VSAT networks are discussed in Chapter 11.
The building of a private local network requires good planning by telecommunication professionals. Switching is assumed. Building an analog network
would be a step backwards; thus the network should be digital. It would be
wise to have full compatibility with the serving PSTN for backup or extensions see interconnect below.. This move returns us to a network optimized for voice operation. Would we have the courage to use an ATM-based
regime and design the transport and switching fabric optimized for that
operation? See Chapter 16 for an extensive discussion of ATM..
We can argue the means of transmission transport: LOS microwave or
fiber optics. Line-of-sight microwave in many instances may be more economic, at least in the short- or mid-term. Fiber has almost unlimited digital
capacity; LOS microwave does not. Local town fathers often object to LOS
microwave, with the mistaken notion that such microwave systems impair the
health of residents. Just this fact alone has brought many corporations to the
decision to opt for buried or aerial fiber links.
Interconnect is the ability of a private network to interconnect with the
public switched telecommunications network PSTN.. It is highly recommended that the design of any private network be made compatible with the
PSTN. If nothing else, we may want to use the PSTN for backup during
failure, or use it for overflow. There is one dichotomy: the LAN. LANs
operate, most generally, at much higher speeds than can be accommodated
152
on the PSTN, namely, 10 Mbps or better. However, the PSTN in North

America is even getting prepared to handle data transmission rates in this
range or greater with switched multimegabit data service SMDS.. Frame
relay is yet another attempt to meet this need.
7.8 BYPASS IN ECONOMICALLY EVOLVING NATIONS

In many countries, bypass, if permitted, is simply a way of getting telecommunication services. There are many locations in the world where there is no
telecommunication service whatsoever, or there is a 4-year wait for a simple
telephone. In many instances, the QoS of a PSTN is so poor that a private
network where permitted . can provide acceptable service.
Bypass seems very desirable. Two points should be made, however. If the
initial, or even upgraded, design of a private network does not meet expectations, the network owner remains responsible. The owner cannot turn to the
PSTN for damages.
The second point is maintenance. This will be the owners responsibility
unless the maintenance is outsourced. Outsourcing is becoming more and
more attractive.
When sizing a private network, overbuild should be considered. How
much overbuild? An economic study should answer that question. This
overbuild is not only for growth of traffic over the planning period, but also
the possibility of leasing network capacity. One of the best customers for
spare capacity may well be the PSTN. This leasing of spare capacity has
become a major source of income for electric power companies in Canada
and the United States.
REFERENCES
1. International Telephone Routing Plan, CCITT Rec. E.171, Fascicle II.2, IXth
2. Vocabulary of Digital Transmission and Multiplexing, and Pulse Code Modulation
PCM . Terms, CCITT Rec. G.701, ITU Geneva, March 1993.
3. Digital Hierarchy Bit Rates, CCITT Rec. G.702, Fascicle III.4, IXth Plenary
4. Roger L. Freeman, Telecommunication System Engineering, 3rd ed., John Wiley &
Sons, New York, 1996.
5. Bellcore Notes on the Networks, Issue 3, SR-2275, Bellcore, Morristown, NJ,
December 1997.
6. Timing Characteristics of Primary Reference Clocks, ITU-T Rec. G.811, ITU Geneva,
September 1997.
REFERENCES
153
7. Controlled Slip Rate Objecties on an International Digital Connection, CCITT Rec.

G.822, Fascicle III.3, IXth Plenary Assembly, Melbourne, 1988.
8. Support by an ISDN of Data Terminal Equipments with V-Series Types Interfaces,
ITU-T Rec. V. 110, ITU Geneva, October 1996.
9. Digital Data SystemChannel Interface Specification, PUB 62310, Bellcore,
Morristown, NJ, September 1983.
10. Digital Data Special Access Serice, Bellcore Technical Reference TR-NPL-000341,
Issue 1, Morristown, NJ, March 1989.
11. Digital Data System, Data Serice Unit Interface Specification, PUB 41450, AT & T,
New York, 1981.
8
THE TRANSMISSION OF
DATA OVER THE ANALOG
VOICE CHANNEL
8.1 BACKGROUND
Access to the PSTN is commonly by means of the standard analog telephone
channel. That access is usually a wire pair connecting the telephone to the
local serving exchange central office.. In the business environment, the telephone wire pair is usually terminated in a PABX where there is a conversion
to a digital regime internal to the PABX. When a call is routed to a distant
end-user, the PABX will place that call on a DS1 time slot, which is then
passed to the local serving exchange in a DS1 configuration Chapter 7..
Some PABXs extend digital service directly to an end-users telephone.
In most cases, whether residential or commercial, end-users are constrained by the typical analog voice channel. Baseband data transmission,
which we can typically expect as inputroutput of the DTE see Section 6.3.,
is incompatible with the analog voice channel. The device that makes the
baseband data signal compatible with the analog voice channel, is the DCE,
which we have called a modem.
In this chapter we first discuss two-wire and four-wire transmission,
because many data sets are connected to the network on a four-wire basis.
This is followed by a review of the basic impairments we can expect as our
data bit stream traverses the PSTN. The principal theme of the chapter is the
description of how modern data modems work.
Before tackling that question, we have to step back a bit to discuss the
evolution of modem technology. The principal constraint of the analog voice
channel is its limited bandwidth. Remember that it occupies the band from
300 to 3400 Hz, a 3100-Hz bandwidth. If we allow 1 bit per hertz of
bandwidth for data occupancy, the maximum data rate it can support is only
3100 bps, far less than the 34- and 56-kbps service that current modems
provide. This, really, is the basic thrust of the chapter.
155
156
THE TRANSMISSION OF DATA OVER THE ANALOG VOICE CHANNEL
8.2 TWO-WIRE VERSUS FOUR-WIRE OPERATION

8.2.1 What is Two-Wire and Four-Wire Operation?
Our telephone at home or in the office involves two-wire operation. It
connects to the serving central office exchange. as a subscriber loop or as an
equivalent subscriber loop to the serving PABX, or key system in many office
situations. A wire pair means two wires. We talk on it, and we listen on
itboth talk and listen on the same pair of wires.
Four-wire operation means we talk on one pair and listen on the other
pair. There are few situations where this actually happens. The U.S. Department of Defense AutoVon network carries four-wire operation to the subscriber in theory., as one example.
Four-wire operation is a concept internal to a telephone network; it is
used very often with higher-speed data modems, such as the CCITT V.29
modem discussed in Section 8.5.3. Private telecommunication networks will
commonly resort to four-wire operation for all but local servicethat is, on
the user side of a PABX.
In the PSTN, typical systems using four-wire operation are PCM TDM.,
FDM, radio, coaxial cable, some wire-pair, and fiber-optic basic transport
systems.
To further our data skills, we must have a firm understanding of two-wire
and four-wire operation and of the impairments involved.
8.2.2 Two-Wire Transmission
By its very nature a telephone conversation requires transmission in both
directions. When both directions are carried on the same wire pair, we call it
two-wire transmission. A more proper definition of two-wire transmission is
that when oppositely directed portions of a single telephone conversation
occur over the same electrical transmission channel or path, we call this
two-wire operation.
8.2.3 Four-Wire Transmission
Carrier, fiber optics, and radio systems require that oppositely directed
portions of a single conversation occur over separate transmission channels
or paths or using mutually exclusive time periods.. Thus we have two wires
for the transmit path and two wires for the receive path, or a total of four
wires for a full-duplex two-way. telephone conversation.
Nearly all long-distance toll. telecommunication connections traverse
four-wire links. In many, if not most, cases, from the near end-user the
connection to the long-distance network is partially, or in some cases totally,
two-wire. Likewise, the distant end-user is also connected to the long-
TWO-WIRE VERSUS FOUR-WIRE OPERATION
157
Figure 8.1. Typical long-distance telecommunications connection. Conceptually, this figure

has been drawn to show operation of a hybrid or term set. N is the matching network.
distance network via a similar two-wire link. Such a long-distance connection

is shown conceptually in Figure 8.1. Schematically, the four-wire connection
shown in the figure is as though it were a wire-line, single channel with
amplifiers. More than likely it would be one channel of a multichannel
carrier* on cable andror multiplex on radio. However, the amplifiers and
their implied direction of transmission serve to convey the ideas of this
section.
As shown in Figure 8.1, conversion from two-wire to four-wire operation is
carried out by a terminating set, more commonly referred to in the industry
as a term set. This set contains a four-winding balanced transformer or a
resistive network, the latter being less common. The balanced transformer or
resistive network is called a hybrid. The hybrid is the key element in a term
set. A term set may also include some signaling devices, such as a break-out
of E & M signaling. This is a carrier-derived supervisory signaling system.
Our concern here is the operation of the hybrid.
Operation of a Hybrid. A hybrid, for telecommunications work, is a transformer. For a simplified description, a hybrid may be viewed as a power
splitter with four sets of wire-pair connections i.e., a four-port device..
Figure 8.2 shows a functional block diagram of a hybrid. Two of the wire-pair
connections belong to the four-wire path, one pair connecting to the transmit
port and one pair connecting to the receive port. The third pair is the
*In all probability, one channel input on a PCM channel bank.
158
Figure 8.2. Operation of a hybrid transformer.
connection to the two-wire link L., which we can imagine connecting,

probably through a local switch, to a subscriber. The last wire pair connects
the hybrid to a resistance capacitance balancing network N. which electrically balances the hybrid with the two-wire connection to the subscribers
subset over the frequency range of the balancing network. An artificial line
may also be used for this purpose.
The hybrid function permits signals to pass from any pair through the
transformer to both adjacent pairs but blocks signals to opposite pairs as
shown in Figure 8.2. Signal energy entering from the four-wire side divides
equally, half dissipating in the balancing network and half going to the
desired two-wire connection. Ideally, no signal energy in this path crosses
over to the four-wire transmit side. This is an important point, which is
discussed in the following text.
Signal energy entering from the two-wire side, connecting eventually. to a
subset, divides equally, half dissipating in the impedance of the four-wire side
receive path and half going to the four-wire transmit path. Here the ideal
situation is that no energy is to be dissipated in the balancing network i.e.,
there is a perfect balance .. The balancing network is supported to display the
characteristic impedance of the two-wire line subscriber connection. to the
hybrid.
The reader notes that in the description of the hybrid, in every case,
ideally half the signal energy entering the hybrid is used to advantage and
half is dissipated, wasted. Also keep in mind that any passive device inserted
in a circuit such as a hybrid has an insertion loss. As a rule of thumb we say
that the insertion loss of a hybrid is 0.5 dB. Hence there are two losses here,
of which the reader must not lose sight Ref. 1.:
ybrid insertion loss:
0.5 dB
Hybrid dissipation loss: 3.0 dB half power.
3.5 dB total.
AMPLITUDE DISTORTION AND PHASE DISTORTION
159
8.3 ECHO AND SINGING: TELECOMMUNICATION NETWORK

IMPAIRMENTS
The operation of the hybrid with its two-wire connection on one end and
four-wire connections on the other leads us to the discussion of two phenomena that, if not properly designed for, may lead to major impairments to the
data bit stream being transported. These impairments are echo and singing.
Echo. As the name implies, echo in telecommunication systems is the
return of a talkers voice for speech operation. When the operation is
data transmission, it is the return of data energy in the form of a
modulated tone on the four-wire receive side entering the data modem.
If of sufficient level, the modulated tone will corrupt the desired receive
signal. We can expect the echo to have the same effect if that modem is
operating in a two-wire mode. In this mode, the local echo cancellers
will not be able to control the interference.
Singing. Singing is the result of sustained oscillations due to positive
feedback in amplifying circuits i.e., the amplifiers in Figure 8.1..
Singing can be thought of as echo which is completely out of control.
Singing will have the same effect as echo on the return path to the
modem but can be much more disruptive.
The primary cause of echo and singing generally can be attributed to the
mismatch in a hybrid between the balancing network see Figure 8.2. and its
two-wire connection associated with the subscriber loop. It is at this point
that the major mismatch usually occurs. The problem is particularly acute
when there is two-wire switching between the two-wire side of the hybrid and
subscriber loops.
When subscriber loops terminate in a channel bank, which is more or less
the rule on modern networks, echo and singing are much less of a problem
for the data user. If both sides of a data circuit are operating in a four-wire
mode, echo and singing are negligible.
On long-distance toll. circuits, echo and singing are controlled by adding
loss to the circuit. Up to 6 dB may be added with modern all-digital networks
up to a certain round-trip delay, where the echo is controlled by echo
cancellers.
8.4 AMPLITUDE DISTORTION AND PHASE DISTORTION

8.4.1 Introduction
As mentioned above, the analog voice channel, as defined by CCITT,
occupies the band from 300 to 3400 Hz, providing a bandwidth of 3100 Hz.
160
Figure 8.3. Limits of overall loss of a circuit relative to that at 1020 Hz. It should be noted that
the figure expresses an amplitude distortion specification. (From Figure 1rM.1020, page 56,
CCITT Rec. M.1020, Ref. 4.)
Not all of the hertz are useful for data transmission. As we approach either
band edge, amplitude distortion and phase distortion become intolerable.
In this section we review amplitude and phase distortion as an introduction and background for our subsequent discussion of conditioning and
equalization. Section 6.3.2.3 gave a more in-depth review of these impairments.
8.4.2 Amplitude Distortion
Amplitude distortion, attenuation distortion, and amplitude response are
different names for the same phenomenon. Let us define this impairment as
the relative attenuation at any frequency with respect to that at reference
frequency. Common reference frequencies for the voice channel are 1000 Hz
in North America and 800 Hz in Europe.
CCITT Rec. G.132 Ref. 2. recommends no more than 9-dB attenuation
distortion relative to 800 Hz between 400 and 3000 Hz on an international
connection. Such distortion would be untenable for a data circuit. Figure 8.3
shows the typical limits required of attenuation distortion on a data circuit.
8.4.3 Phase Distortion
Many sources in the literature tell us that phase distortion is the principal
bottleneck to data rate. Phase distortion is the result of delay. Delay varies
from band center to band edge. In other words, our data waveform travels
161
the fastest at band center, and as we approach band edge, it starts to slow
down. What we are saying is that the velocity of propagation is not constant
across the band of interest.
Of course delay is minimal, and delay distortion i.e., phase distortion. is
the flattest around band center, which is in the vicinity of 1700 and 1800 Hz.
By flattest we mean where there is a comparatively minimum variation in
delay. Delay distortion is often expressed by a related parameter called
enelope delay distortion EDD.. For example, a certain specification may read
that the EDD between 1000 and 2600 Hz will not exceed 1000 s. Note that
1800 Hz is right in the center of this band.
One rule of thumb often used is that envelope delay distortion in the band
of interest should not exceed the period of one bit. With a non-return-to-zero
NRZ. waveform, the period of a bit is simply calculated by
Bit period seconds . s 1r bit rate .
Some Examples: The period of a bit i.e., the time duration of one bit. for
2400 bps bit stream is 1r2400 s or 416.66 s; for 56,000 bps, it is 1r56,000.
s 17.857 s; for 1 Mbps, it is 1r1 = 10 6 . s or 1 s; for 1.544 Mbps, it is
1r1.544 = 10 6 . or 0.6476 s. In a similar manner, given the period of a bit
in seconds, we can calculate the bit rate in bps, assuming an NRZ waveform.
For this discussion, we will assume that envelope delay distortion values are
the same as group delay distortion values.
If the envelope delay distortion is 1 ms from 1000 to 2600 Hz, the
maximum bit rate assumes binary transmission. is 1 kbps. We will find that
there are two ways to get around this dilemma of a limited bit rate. The first
is to use equalization discussed below., and the second is to reduce
the modulation rate. One way to accomplish the latter is to reduce the bit
rate. Another approach would be to use an M-ary wave form which reduces
the modulation rate without sacrificing the bit rate. M-ary waveforms are
discussed in Section 8.5. For further reading see CCITT Rec. G.133 Ref. 3..
8.4.4 Conditioning and Equalization
Line conditioning in the United States is a tariffed offering by the serving
common carrier to the data end user. In other words, the telephone company
or administration guarantees that a certain leased line or lines will meet
some minimum standards regarding amplitude distortion and envelope delay
distortion. Because these lines are special, the telephone company charges
the customer accordingly. Table 8.1 shows bandwidth parameter limits for
typical conditioned line offerings for the United States.
CCITT Rec. M.1020 provides requirements for special conditioning of
leased lines for data transmission. Figure 8.3 shows requirements for amplitude distortion i.e., overall loss relative to 1020 Hz., and Figure 8.4 shows
limits of group delay relative to the minimum measured group delay in the
162
TABLE 8.1 Bandwidth Parameter Limits a, b

Attenuation Distortion
(Frequency Response)
Relative to 1004 Hz
Channel
Conditioning
Frequency
Range (Hz)
Basic
5002500
3003000
10002400 d
3002700 d
3003000
5002800 d
3003000 d
C1
C2
Variation
(dB) c
y2
y3
y1
y2
y3
y1
y2
to q8
to q12
to q3
to q6
to q12
to q3
to q6
C3
(access line)
5002800 d
3003000 d
y0.5 to q1.5
y0.8 to q3
C3
(Trunk)
5002800 d
3003000 d
y0.5 to q1
y0.8 to q2
C4
5003000 d
3003200 d
y2 to q3
y2 to q6
C5
5002800 d
3003000 d
y0.5 to q1.5
y1 to q3
Envelope Delay
Distortion
Frequency
Range (Hz)
Variation
(microseconds)
8002600
1750
10002400 d
8002600
1000
1750
10002600 d
6002600 d
5002800 d
10002600 d
6002600 d
5002800 d
10002600 d
6002600 d
5002800 d
10002600 d
8002800 d
6003000 d
5003000 d
10002600 d
6002600 d
5002800 d
500
1500
3000
110
300
650
80
260
500
300
500
1500
3000
100
300
600
C-conditioning applies only to the attenuation and envelope delay characteristics.

Measurement frequencies will be 4 Hz above those shown. For example, the basic channel will
have y2- to q8-dB loss, with respect to the 1004-Hz loss, between 504 and 2504 Hz.
c
(q ) means loss with respect to 1004 Hz. (y ) means gain with respect to 1004 Hz.
d
These specifications are tariffed items.
b
Source: Reference 5.
Figure 8.4. Limits of group delay relative to the minimum measured group delay in the 500to 2800-Hz band. (From Figure 2r1020, page 56, CCITT Rec. M.1020, Ref. 4.)
163
band from 500 to 2800 Hz. For our discussion, group delay distortion. values
are very close to envelope delay values, and thus we may use them interchangeably. For example, in Figure 8.4, between 1000 and 2600 Hz, the
envelope delay distortion is 500 s.
Conditioning, therefore, is what the telephone company administration .
can do for the data end user. Equalization is what the data end-user can do
to improve either raw if nonconditioned lines are used. or residual amplitude distortion and envelope delay distortion in the case of conditioned lines.
There are several methods of performing equalization. One of the most
common is to use several networks in tandem. Such networks tend to flatten
response. In the case of amplitude distortion, they add attenuation increasingly toward band center where end-to-end loss is the least, and less and less
is added as band edges are approached. Essentially the amplitude equalizer
provides a mirror response opposite to that of the channel response, making
the net amplitude response comparatively flat.
The delay equalizer operates in a fairly similar fashion. Voice channel
delay increases toward channel edges parabolically from channel center. The
delay equalizer adds delay in the center, with less and less delay added as
band edges are approached, much like an inverted parabola. Thus the delay
response is flattened at some small cost in absolute delay, which, in most data
systems, has no effect on performance. However, care must be taken with the
effect of a delay equalizer on an amplitude equalizer, and, conversely, the
effect of an amplitude equalizer on a delay equalizer. Their design and
adjustment must be such that the flattening of the channel for one parameter
does not entirely distort the channel for the other parameter. Figure 8.5
shows typical envelope delay response in a voice channel along with the
opposite response of an equalizer to flatten the envelope delay characteristics
of the voice channel.
Another type of equalizer is the transversal type of filter, which is shown
in Figure 8.6. This equalizer is useful where it is necessary to select among,
or to adjust, several attenuation amplitude . and phase characteristics. As
shown in Figure 8.6, the basis of the filter is a tapped delay line to which the
input is presented. The output is taken from a summing network which adds
or sums the outputs of the taps. Such a filter is adjusted to the desired
response equalization of both phase and amplitude . by adjusting the tap
contributions.
If the characteristics of a line are known, which is probably the case for a
leased facility, another method of equalization is predistortion of the output
signal of the data set e.g., modem.. Some devices designed for this requirement consist of a shift register and a summing network. If the equalization
needs to be varied, then a feedback circuit from the receiver to the transmitter would be required to control the shift register. Such a type of active
predistortion is valid for binary transmission only.
A major drawback to the equalizers discussed with the exception of the
last with a feedback circuit. is that they are useful only on dedicated or
164
Figure 8.5. A delay (phase) equalizer tends to flatten delay characteristic of a voice channel.
(From Ref. 6. Copyright BTL. Reprinted with permission.)
Figure 8.6. A transversal filter.
leased circuits where the circuit characteristics are known and remain fixed.
Obviously, if a switched circuit is used, such as on a dial-up connection, a
variable automatic equalizer is required, or manual conditioning would be
necessary on every circuit in the switched systems that would possibly be
transmitting data.
Circuits are conventionally equalized on the receiving end. This is called
post-equalization. Equalizers must be balanced and present the proper
impedance to the line. Administrations telephone companies. may choose to
condition trunks and attempt to eliminate the need to equalize station lines.
The economy of considerably fewer equalizers is obvious. In addition, each
circuit that would possibly carry comparatively high-speed data in the system
would have to be equalized, and the equalization must be good enough that
any possible circuit combination would meet the overall requirements. If
DATA MODEMS
165
equalization requirements become greater i.e., parameters are more stringent., then consideration may have to be given to the restriction of the
maximum number of circuits trunks. in tandem. CCITT Rec. E.171. recommends no more than 12 circuits in tandem on an international connection.
Equalization to meet amplitudefrequency response attenuation distortion. requirements is less exacting on the overall system than is EDD. With
modern digital networks, considering the network in isolation, the principal
contributors to amplitude and phase distortion are the input and output
3.4-kHz filters. The measured performance for amplitude and phase distortion should prove out considerably better than the requirements of CCITT
Rec. M.1020. Now add a subscriber loop to each end of the connection to
permit a DCErDTE access, and we may have some fairly different results.
Automatic equalization for both amplitude and delay is widely used. All
CCITT-recommended modems operating at 2400 bps and higher have automatic equalizers built in. These equalizers are self-adaptive and require a
short adaptation or training period after a circuit is switched, on the order of
1 s or less. This can be carried out during the transmission of synchronization
sequences. Not only is the modem clock being averaged for the new circuit
on the transmission of a synchronization idle signal, but the self-adaptive
equalizer adjusts for optimum equalization at the same time.
8.5 DATA MODEMS

8.5.1 Where We Are and Where We Are Going
In Section 6.3.2.2 modems were introduced. In Section 8.1 we showed that a
data rate up to 3100 bps could optimistically be achieved using some form of
binary transmission. This was a hypothetical case where we had the 3100-Hz
bandwidth of the standard voice channel. In this assumption, we did not
cover the fact that those frequencies near band edge had excessive amplitude
and phase distortion. For all intents and purposes, the really useful part of
the band is 6002800 Hz, which could derive perhaps some 2200 bps.
2200 bps is quite far from todays speeds of 28 and 56 kbps. How do we get
from here to there? Where we fall down is that we were considering just one
bit per hertz of occupancy. How can we improve the bit packing to 2, 3, 4, or
5 or more bits per hertz?
8.5.2 Getting More Bits per Hertz

In Section 6.3.2 modulation was introduced where the three generic modulation types available to the modem designer were discussed. These are
amplitude modulation AM., frequency modulation FM., and phase modulation PM..
166
Because of the influence of automatic landline and radio telegraphy, the

nomenclature has been changed to the following:
Amplitude shift keying ASK. for AM

Frequency shift keying FSK. for FM
Phase shift keying PSK. for PM
In each case an audio tone is used to carry the information. It is called a tone
because if we connect a loudspeaker to the tone generator, the tone can be
heard having a musical quality. This happens because the tone, when detected, is in the audio-frequency range of the human ear. It is also in the
range of the standard voice channel, between 300 and 3400 Hz as one would
expect.
In the case of ASK, for binary transmission, a binary 1 is tone on and a
binary 0 is tone off. A binary sequence, then, is transmitted as a series of
tone-on and tone-off sequences.
The two binary states for FSK represent a higher-frequency tone for the
mark or binary 1 and a lower-frequency tone for the space or binary 0. At the
FSK receiver, two filters are used, one centered on each tone frequency. If
there is tone energy in the higher-frequency filter, then a mark or binary 1
has been received; if there is tone energy in the lower-frequency filter slot,
then a space or binary 0 has been received.
Phase-shift keying modems use only one tone, usually centered around
1700 or 1800 Hz, where phase distortion is flattest. The phase of the tone can
be retarded. Suppose a binary 0 is assigned to the condition of no phase
retardation and a binary 1 is assigned to 180 phase retardation. This now
defines BPSK or binary phase shift keying. For example, if we receive a
signal with 180 of phase retardation, we receive a binary 1; if we receive a
signal an 1800-Hz tone. with no phase retardation, we receive a binary 0.
Phase modulation is usually illustrated diagrammatically by a circle. Figure
8.7 shows our BPSK example.
Carry this concept one step further. Retard the phase in 90 increments,
or 0, 90, 180, and 270. Figure 8.8 shows where we divide the circle into
Figure 8.7. A spatial representation of a data tone with BPSK

modulation. Commonly a circle is used when we describe phase
modulation. The frequency of the data tone is usually 1800 Hz.
DATA MODEMS
167
Figure 8.8. A spatial representation of a data tone with

QPSK modulation. Such a spatial representation is often
called signal constellation.
quadrants, representing the following:

Phase Change (Degrees)
Equivalent Binary Number
0
90
180
270
00
01
10
11
We have increased the bit packing capacity to 2 bits per hertz of bandwidth.
This is so because for every transition, 2 bits are transmitted. For instance, if
there were a transition to 270, the bit sequence 11 would be transmitted.
This can be carried still one step further, where the circle has eight
positions at 45 increments. This is illustrated in Figure 8.9. Now each
transition represents 3 bits. The bit packing has been increased to 3 bits per
hertz of bandwidth. If there were 3000 Hz of bandwidth available, its
theoretical capacity is 9000 bps. We call this waveform 8-ary PSK.
The bit packing values discussed here e.g., 1 bitrHz, 2 bitsrHz, 3
bitsrHz. are theoretical values. Practical values are less optimistic. For
example, QPSK may achieve no more than 1.2 bits per hertz in the real
world.
Figure 8.9. A spatial representation of 8-ary PSK.

Note that the binary sequences have been selected
arbitrarily. For an interface with the distant end, a
certain specification or ITU recommendation is dictated, such as ITU-T Rec. V.34. Thus there is agreement by all parties concerned regarding the values of
the signal points in the constellation.
168
Figure 8.10. An example of a 16-QAM signal constellation (i.e., 8 phases and 2 levels.)
A very important concept here is that bandwidth is dictated by transitions

per second, which we call bauds. Only in the binary regime e.g., BPSK. are
bauds and bits per second the same.
Up to this point we have described how to achieve 2 and 3 bits per hertz of
bit packing. Suppose now we not only vary the phase as we did above, but we
also vary to amplitude or level of the signal. Of course we will vary it in fixed
increments. Figure 8.10 shows a simple example of this idea. We have eight
phases as we did with 8-ary PSK. and two amplitudes or levels. This
illustrates a method of achieving 4 bits per hertz of bit packing. The CCITT
V.29 modem uses this technique. It is a 9600-bps modem operating at 2400
baud. For every transition in this case, 4 bits are transmitted. The waveform
is called 16-QAM, where QAM stands for quadrature amplitude modulation.
8.5.3 Specific High-Speed Modems
A well-known data modem used for 9600-bps operation is described in
CCITT Rec. V.29 Ref. 7.. It is standardized for use in point-to-point
four-wire telephone-type circuits. The main characteristics of this modem are
as follows:
1. Fall back data rates of 7200 and 4800 bps
2. Full-duplex and half-duplex operation
3. Combined amplitude and phase modulation with synchronous mode
operation
4. Automatic equalizer
5. Optional inclusion of a multiplexer for combining data rates of 7200,
4800, and 2400 bps
6. Line signal at 1700 Hz " 1 Hz
DATA MODEMS
169
7. Signal space coding. At 9600 bps the scrambled data stream to be

transmitted is divided into groups of four consecutive data bits called
quad bits.. The first bit, Q1 , of each quadbit is used to determine the
signal element amplitude to be transmitted. The second quadbit element Q2 , third Q3 , and fourth Q4 are encoded as a phase change
relative to the phase of the immediate preceding element.
It should be noted that this modem achieves a theoretical bit packing of 4
bits per hertz. Three of these bits derive from using 8-ary PSK, and the
fourth derives from using the equivalent of two amplitude levels. Thus it is a
2400-baud modem capable of transmitting 9600 bps.
Table 8.2 shows the CCITT Rec. V.29 phase encoding, Table 8.3 gives the
amplitudephase relationships, and Figure 8.11 is the signal space diagram
for 9600 bps operation. The figure has 16 points in space representing the 16
quadbit possibilities, 2 4 s 16. Thus the modem achieves 4 bits per hertz of
theoretical bit packing.
The CCITT Rec. V.32 Ref. 8. modem is designed for full-duplex operation at 9600 bps on two-wire circuits for use on the PSTN and on leased
telephone-type circuits. Some of the highlights of the V.32 modem are as
follows:
1. Full-duplex operation over the PSTN and on two-wire point-to-point
leased circuits
2. Channel separation by echo cancellation techniques i.e., separation of
transmit and receive channels on same wire pair.
TABLE 8.2 Phase Encoding for the CCITT Rec. V.29 Modem
Q2
Q3
Q4
Phase Change (deg)
0
0
0
0
1
1
1
1
0
0
1
1
1
1
0
0
1
0
0
1
1
0
0
1
0
45
90
135
180
225
270
315
TABLE 8.3 AmplitudePhase Relationships with the CCITT Rec. V.29 Modem
Q1
Relative Signal
Element Amplitude
0, 90, 180, 270
0
1
3
5
45, 135, 225, 315
'2
3'2
Absolute Phase (deg)
170
Figure 8.11. Signal space diagram for the CCITT V.29 modem when operating at 9600 bps.
3. Quadrature amplitude modulation QAM. for each channel with synchronous line transmission at 2400 bauds.
4. Any combination of the following data signaling rates may be implemented: 9600-, 4800-, and 2400-bps synchronous.
5. At 9600-bps operation, two alternative modulation schemes are provided. One uses 16 carrier states, and the other uses trellis coding with
32 carrier states. However, modems providing the 9600-bps data signaling rate shall be capable of interworking using the 16-state alternative.
6. There is an exchange of rate sequences during start-up to establish the
data rate, coding, and any other special facilities requiring coordination.
7. There is an optional provision for an asynchronous startstop. mode
of operation in accordance with CCITT Rec. V.14.
8. The carrier tone. frequency is 1800 Hz " 1 Hz. The receiver can
operate with frequency offsets up to "7 Hz.
CCITT Rec. V.32 offers two alternatives for signal element coding of the
9600-bps data rate: i. nonredundant and ii. trellis coding.
With nonredundant coding, the scrambled data bit stream to be transmitted is divided into groups of four consecutive bits. The first 2 bits in time,
Q1 n and Q2 n , where the subscript n designates the sequence number of the
group, are differentially encoded into Y 1 n and Y 2 n , in accordance with
Table 8.4. Bits Y 1 n , Y 2 n , Q3 n , and Q4n are then mapped into coordinates of
the signal state to be transmitted according to the signal space diagram
shown in Figure 8.12 and as listed in Table 8.5.
When using the second alternative with trellis coding, the scrambled data
stream to be transmitted is divided into two groups of four consecutive data
DATA MODEMS
171
TABLE 8.4 Differential Quadrant Coding for 4800-bps Operation and for Nonredundant
Coding at 9600 bps
Inputs
Previous Outputs
Q1n
Q2 n
Y1n y 1
Y 2ny1
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Phase Quadrant
Change
(deg)
q90
q180
q270
Outputs
Y1n
Y 2n
Signal State
for 4800 bps
0
1
0
1
0
0
1
1
1
1
0
0
1
0
1
0
1
1
0
0
0
1
0
1
1
0
1
0
0
0
1
1
B
C
A
D
A
B
D
C
C
D
B
A
D
A
C
B
Source: Reference 8, Table 1rV.32, page 236.
Figure 8.12. Sixteen-point signal constellation with nonredundant coding for 9600-bps operation with subset A, B, C, and D states used for 4800-bps operation and for training. (From
Figure 1r V.32, page 236, CCITT Rec. V.32, Ref. 8.)
172
TABLE 8.5 Two Alternative Signal State Mappings for 9600-bps Operation
Coded Inputs a
Trellis Coding
( Y 0)
Y1
Y2
Q3
Q4
Re
Im
Re
Im
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
10
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
y1
y3
y1
y3
1
1
3
3
y1
y1
y3
y3
1
3
1
3
y1
y1
y3
y3
y1
y3
y1
y3
1
3
1
3
1
1
3
3
y4
0
0
4
4
0
0
y4
y2
y2
2
2
2
2
y2
y2
y3
1
y3
1
3
y1
1
y3
1
1
y1
3
y1
y1
3
y1
3
y1
y3
1
y3
1
y2
y2
2
2
2
2
y2
y2
4
0
0
y4
y4
0
0
4
Nonredundant
Coding
3
y1
1
y3
1
1
y1
3
y1
y1
See Tables 8.4 and 8.6.
Source: Reference 8, Table 3rV.32, page 238.
bits. As shown in Figure 8.13, the first 2 bits in time Q1 n and Q2 n in each
group, where the subscript n designates the sequence number of the group,
are first differentially encoded into Y 1 n and Y 2 n in accordance with Table
8.6. The two differentially encoded bits Y 1 n and Y 2 n are used to input a
systematic convolutional coder which generates redundant bit Y 0 n . The
redundant bit and the four information carrying bits Y 1 n , Y 2 n , Q3 n , and Q4n
are then mapped into coordinates of the signal element to be transmitted
according to the signal space diagram shown in Figure 8.13 and as listed in
Table 8.5.
173
DATA MODEMS
Figure 8.13. A 32-point signal constellation with trellis coding for 9600-bps operation and
states A, B, C, and D used for 4800-bps operation and for training. The binary numbers
denote Y 0 n , Y1n , Y 2 n , Q3 n , and Q4n . (From Figure 3r V.32, page 239, CCITT Rec. V.32, Ref.
8.)
TABLE 8.6 Differential Encoding for Use with Trellis-Coded Alternative at 9600 bps
Inputs
Previous Outputs
Outputs
Q1n
Q2 n
Y1n y 1
Y 2ny1
Y1n
Y 2n
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
1
0
1
1
0
1
0
0
1
1
0
1
0
0
1
Source: Reference 8, Table 2r V.32, page 237.
174
For 4800-bps operation the data bit stream to be transmitted is divided

into groups of two consecutive data bits Q1 n and Q2 n , where Q1 n is first in
time and the subscript n designates the sequence number of the group,
differentially encoded into Y 1 n and Y 2 n according to Table 8.4. Figure 8.13
shows subsets A, B, C, and D of the signal states used for 4800-bps operation.
CCITT Rec. V.33 Ref. 9. defines a 14,400-bps modem standardized for
use on point-to-point four-wire leased telephone-type circuits. The modem is
intended to be used primarily on special-quality leased circuits that are
typically defined in CCITT Rec. M.1020 see Section 8.4.4.. The modems
principal characteristics are:
A fallback data rate of 12,000 bps

A capability of operating in a duplex mode with continuous carrier
Combined amplitude and phase modulation i.e., QAM. with synchronous mode of operation
Inclusion of eight-state trellis-coded modulation
Optional inclusion of a multiplexer for combining data rates of 12,000,
9600, 7200, 4800, and 2400 bps
A carrier frequency of 1800 Hz " 1 Hz
At the 14,400-bps data rate, the scrambled data stream to be transmitted

is divided into groups of six consecutive data bits. As illustrated in Figure
8.14, the first 2 bits in time Q1 n and Q2 n in each group are first differentially
coded into Y 1 and Y 2 in accordance with Table 8.7. The two differentially
encoded bits Y 1 n and Y 2 n are used as input to a systematic convolutional
encoder which generates redundant bit Y 0 n . This redundant bit and the six
information-carrying bits Y 1 n , Y 2 n , Q3 n , Q4n , Q5n , and Q6 n are then
mapped into coordinates of the signal element to be transmitted in accordance with the signal space diagram shown in Figure 8.14. Figure 8.15 shows
the signal constellation and mapping for trellis-coded modulation at 12,000
bps. The self-synchronizing generating polynomial is 1 q Xy18 q Xy23.
8.5.4 Scrambling and Its Rationale
Many of the more sophisticated higher-speed modems include a scrambler.
This is a device that pseudorandomly shuffles bits in the serial bit stream
before modulation. There is a companion descrambler at the other end of the
circuit that shuffles these bits back to their proper order after demodulation
in the receive modem. The generating polynomial is given for each modem.
Thus the receive modem knows a priori the scrambling. Before it can
properly descramble the receive bit stream, it must be time-slaved to the
transmit modem to effect proper start of descrambling and to maintain
synchronization throughout the scrambling period.
DATA MODEMS
175
Figure 8.14. Signal constellation and mapping for trellis-coded modulation at 14,400 bps.
Binary numbers refer to Q6 n , Q5 n , Q4n , Q3 n , Y 2 n , Y 2 n , Y1n , and Y0 n . A, B, C, and D refer to
synchronizing signal elements. (From Figure 2r V.33, page 255, CCITT Rec. V. 33, Ref. 9.)
The rationale for scrambling is to optimize transmission performance by

maintaining a minimum number of signal element transitions. Remember
that we derive timing from transitions. Scrambling also minimizes the probability of repetitive data patterns. We also will find scrambling used on
higher-level PCM digital systems to maintain the 1s density. The 1s
density in this case just ensures that there will be a high probability of a
minimum number of transitions. In PCM transmission the 1 is the active
state; the 0 is the passive, or no-voltage state.
The scrambler here is not to be confused with a link encryption system or
scrambler. that is used to maintain data security i.e., to prevent unauthorized listening in.. They operate in a similar manner. However, in the case
of the link encryption system, such parameters as the generating polynomial
are confidential.
176
TABLE 8.7 Differential Encoding for Use with Trellis Coding

Inputs
Previous Outputs
Outputs
Q1n
Q2 n
Y1n y 1
Y 2ny 2
Y1n
Y 2n
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
0
1
1
0
0
1
1
1
1
0
0
1
1
0
0
0
1
0
1
1
0
1
0
0
1
1
0
1
0
0
1
Source: Reference 9, Table 1Ar V.33, page 253.
Figure 8.15. Signal space diagram and mapping for trellis-coded modulation at 12,000 bps.
Binary numbers refer to Q5 n , Q4n , Q3 n , Y 2 n , Y1n , and Y 0 n . A, B, C, and D refer to
synchronizing signal elements. (From Figure 3r V.33, page 256, CCITT Rec. V.33, Ref. 9.)
DATA MODEMS
177
8.5.5 Introduction to Trellis-Coded Modulation (TCM)

In Section 4.5.2, forward error correction FEC. was introduced. FEC allows
the system designer to reduce signal-to-noise ratio a certain amount depending on the BER requirement. That amount of reduction in dB. is called
coding gain. For example, for a certain modulation type, an SrN of 9.6 dB is
required for a BER of 1 = 10y5. Using FEC with given parameters, the same
BER can be achieved with an SrN of 5 dB. The results is a coding gain of
9.6 y 5, or 4.6 dB. We pay for this by an increase in symbol rate ergo
bandwidth . due to the redundancy added.
When working with the analog voice channel, we have severe bandwidth
limitations, nominally 3 kHz. When implementing FEC and maintaining a
high data rate, our natural tendency is to use bit packing schemes, turning
from binary modulation to M-ary modulation, typically QAM schemes. A
good example is the V.29 modem described above, which has a two-dimensional signal constellation: one dimension in phase and the other in amplitude as illustrated in Figures 8.10 and 8.11. These figures have 16 points. As
the number of points increases, the receive modem becomes more prone to
errors introduced by noise hits and intersymbol interference. The problem
can be mitigated by the design of the M-ary signal set.
In the introductory discussion of Section 8.5.2, coding and modulation
were carried out separately. With trellis-coded modulation TCM., there is a
unified concept. Here coding gain is achieved without bandwidth expansion
by expanding the signal set.
In conventional, multilevel amplitude andror phase. modulation systems,
during each modulation interval, the modulator maps m binary symbols bits.
into one of M s 2 m possible transmit signals, and the demodulator recovers
m bits by making an independent M-ary nearest-neighbor decision on each
signal received. Previous codes were designed for maximum Hamming free
distance. Hamming distance refers to the number of symbols in which two
codes symbols or blocks differ regardless of how these symbols differ
Ref. 10..
TCM broke with this tradition and maximized free Euclidean distance
distance between maximum likely neighbors in the signal constellation .
rather than Hamming distance. The redundancy necessary for TCM would
have to come from expanding the signal set to avoid bandwidth expansion.
Ungerboeck Ref. 10. states that, in principle, TCM can achieve coding gains
of about 78 dB over conventional uncoded multilevel modulation schemes.
Most of the achievable coding gain would be obtained by expanding the
signal sets only by a factor of 2. This means using signal set sizes 2 mq1 for
transmission of m bits per modulation interval.
All TCM systems Ref. 10. achieve significant distance gains with as few as
4, 8, and 16 states. Roughly speaking, it is possible to gain 3 dB with 4 states,
4 dB with 8 states, nearly 5 dB with 16 states, and up to 6 dB with 128 or
more states. Doubling the number of states does not always yield a code with
178
larger free distance. Generally, limited distance growth and increasing numbers of nearest neighbors and neighbors with next-largest distances are the
two mechanisms that prevent realizing coding gains from exceeding the
ultimate limit set by channel capacity. This limit can be characterized by the
signal-to-noise ratio at which the channel capacity of a modulation system
with 2 m q 1.-ary signal set equals m bpsrHz Ref. 10; also see Ref. 11..
Practical TCM systems can achieve coding gains of 36 dB at spectral
efficiencies equal to or larger than 2 bitsrHz.
8.5.6 The V.34 Modem: 28,800 bps and 33,600 bps

8.5.6.1 General. The V.34 modem Ref. 12. is designed to operate in the
PSTN and on point-to-point leased telephone-type circuits. The following are
some of the salient characteristics of the V.34 modem:
1. Duplex and half-duplex modes
2. Channel separation by echo cancellation techniques
3. Quadrature amplitude modulation QAM. for each channel with synchronous line transmission at selectable symbol rates including the
mandatory rates of 2400, 3000, and 3200 symbols per second and
optional rates of 2743, 2800, and 3429 symbols per second
4. Synchronous primary channel data rates of:
33,600 bps
28,800 bps
24,000 bps
19,200 bps
14,400 bps
9600 bps
4800 bps
31,200 bps
26,400 bps
21,600 bps
16,800 bps
12,000 bps
7200 bps
2400 bps
5. Trellis coding for all data rates

6. An optional auxiliary channel with a synchronous data rate of 200 bps,
a portion of which may be provided to the user as an asynchronous
secondary channel
7. Adaptive techniques that enable the modem to achieve close to the
maximum data rate the channel can support in each direction
8. Exchange of rate sequence during start-up to establish the data rate
9. Automoding to V-Series modems supported by ITU-T Rec. V.32 bis
automode procedures and Group 3 facsimile machines
DATA MODEMS
179
8.5.6.2 Selected Definitions

Constellation Shaping. A method for improving noise immunity by introducing a nonuniform two-dimensional probability distribution for transmitted
signal points. The degree of constellation shaping is a function of the
amount of constellation expansion.
Data Mode Modulation Parameters. Parameters determined during start-up
and used during data mode transmission.
Frame Switching. A method for sending a fractional number of bits per
mapping frame, on average, by alternating between sending an integer
b y 1 bits per mapping frame and b bits per mapping frame according to
a periodic switching pattern.
Line Probing. A method for determining channel characteristics by sending
periodic signals, which are analyzed by the modem and used to determine
data mode modulation parameters.
Nonlinear Encoding. A method of improving distortion immunity near the
perimeter of a signal constellation by introducing a nonuniform two-dimensional 2D. signal point spacing.
Precoding. A nonlinear equalization method for reducing equalizer noise
enhancement caused by amplitude distortion. Equalization is performed
at the transmitter using precoding coefficients provided by the remote
modem.
Preemphasis. A linear equalization method where the transmit signal spectrum is shaped to compensate for amplitude distortion. The preemphasis
filter is selected using a filter index provided by the remote modem.
Shell Mapping. A method of mapping data bits to signal points in a multidimensional signal constellation, which involves partitioning a two-dimensional signal constellation into rings containing an equal number of signal
points.
Trellis Encoding. A method of improving noise immunity using a convolutional coder to select a sequence of subsets in a partitioned signal
constellation. The trellis encoders employed in the V.34 modem are all
four-dimensional 4D. and they are used in a feedback structure where the
inputs to the trellis encoder are derived from the signal points.
8.5.6.3 An Overview of Selected Key Areas of Modem Operation
Line Signals. The primary channel supports synchronous data rates of 2400
33,600 bps in multiples of 2400 bps. An auxiliary channel with a synchronous
data rate of 200 bps may also be optionally supported. The primary and
auxiliary data rates are determined during phase 4 of modem start-up
according to the procedures described in paragraphs 11.4 and 12.4 of the
180
TABLE 8.8 Symbol Rates

Symbol Rate, S
2400
2743
2800
3000
3200
3429
1
8
7
5
4
10
1
7
6
4
3
7
Source: Reference 12, Table 1r V.34, ITU-T Rec. V.34.
reference document Ref. 12.. The auxiliary channel is used only when the
call and answer modems have both declared this capability. The primary
channel data rates may be asymmetric.
Symbol Rates. The symbol rate is S s arc . = 2400 " 0.01% 2D symbols
per second, where a and c are integers from the set given in Table 8.8 in
which symbol rates are shown rounded to the nearest integer.. The symbol
rates 2400, 3000, and 3200 are mandatory; 2743, 2800, and 3429 are optional.
The symbol rate is selected during phase 2 of modem start-up according to
procedures described in paragraph 11.2 or 12.2 of the reference publication.
Asymmetric symbol rates are optionally supported and are used only when
the call and answer modems have both declared this capability.
Carrier Frequencies. The carrier frequency is dre. = S Hz, where d and e
are integers. One of two carrier frequencies can be selected at each symbol
rate, as given in Table 8.9, which provides the values of d and e and the
corresponding frequencies rounded to the nearest integer. The carrier frequency is determined during phase 2 of modem start-up according to the
procedures given in paragraph 11.2 or 12.2 of the reference publication.
Asymmetric carrier frequencies are supported.
TABLE 8.9 Carrier Frequencies Versus Symbol Rate

Low Carrier
High Carrier
Symbol Rate, S
Frequency
Frequency
2400
2743
2800
3000
3200
3429
1600
1646
1680
1800
1829
1959
2
3
3
3
4
4
3
5
5
5
7
7
1800
1829
1867
2000
1920
1959
3
2
2
2
3
4
4
3
3
3
5
7
DATA MODEMS
181
Figure 8.16. Transmit spectra templates for indices 0 to 5. Note: Tolerance for transmit
spectrum is "1 dB. (From Figure 1r V.34, ITU-T Rec. V.34, Ref. 12.)
Preemphasis
TRANSMIT SPECTRUM SPECIFICATIONS. The transmit spectrum specifications
use a normalized frequency, which is defined as the ratio frS, where f is the
frequency in hertz and S is the symbol rate. The magnitude of the transmitted spectrum conforms to the templates shown in Figures 8.16 and 8.17 for
normalized frequencies in the range from dre s 0.45. to dre q 0.45.. The
transmitted spectrum is measured using a 600- pure resistive load.
Figure 8.16 requires parameter , which is given in Table 8.10; and Figure
8.17 requires parameters and , which are given in Table 8.11.
SELECTION METHOD.
The transmitted spectrum is specified by a numerical

index. The index is provided by the remote modem during phase 2 of start-up
procedures as defined in paragraph 11.2 or 12.2 of the reference publication.
Electrical Characteristics of Interchange Circuits. For the primary channel,

where an external physical interface is provided, the electrical characteristics
will conform to ITU-T Rec. V.10 or V.11. The connector and pole assignments specified in ISO 2110 Amd. 1.0 or ISOrIEC 11569, column V-Series
) 20,000 bps, is used. Alternatively, when the DTEDCE interface speed is
not designed to exceed 116 kbps, these same connectors may be used with
characteristics conforming to ITU-T Rec. V.10 only.
Where an external physical interface is provided for the secondary channel, electrical characteristics are in accordance with ITU-T Rec. V.10.
182
Figure 8.17. Transmit spectra templates for indices 6 to 10. Note: Over the range specified,
the tolerance for the transmit spectrum magnitude is "1 dB. (From Figure 2r V.34, ITU-T
Rec. V.34, Ref. 12.)
TABLE 8.10
Parameter for Indices 0 to 5

Index
0
1
2
3
4
5
0 dB
2 dB
4 dB
6 dB
8 dB
10 dB
TABLE 8.11
Parameters and for Indices 6 to 10
Index
6
7
8
9
10
0.5 dB
1.0 dB
1.5 dB
2.0 dB
2.5 dB
1.0 dB
2.0 dB
3.0 dB
4.0 dB
5.0 dB
DATA MODEMS
183
Scrambler. A self-synchronizing scrambler is included in the modem for the

primary channel. Auxiliary channel data are not scrambled. Each direction of
transmission uses a different scrambler. According to the direction of transmission, the generating polynomial is
Call mode:
Answer mode:
1 q Xy18 q Xy23
1 q Xy5 q Xy23
8.5.6.4 Framing. Figure 8.18 gives an overview of the frame structure. The
duration of a superframe is 280 ms and consists of J data frames, where
J s 7 for symbol rates 2400, 2800, 3000, and 3200, and J s 8 for symbol rates
2743 and 3429. A data frame consists of P mapping frames, where P is given
in Table 8.12. A mapping frame consists of four four-dimensional 4D.
symbol intervals. A 4D symbol interval consists of two 2D symbol intervals. A
bit inversion method is used for superframe synchronization.
Mapping frames are indicated by the time index i, where i s 0 for the
first mapping frame of signal B1 as defined below. and is incremented by 1
for each mapping frame thereafter.
Sequence B1 consists of one data frame of scrambled 1s at the end of
start-up using selected data mode modulation parameters. Bit inversions for
superframe synchronization are inserted as if the data frame were the last
Figure 8.18. Overview of framing and indexing. (From Figure 3rV.34, ITU-T Rec. V.34,
Ref. 12.)
184
TABLE 8.12
Framing Parameters
Symbol Rate, S
2400
2743
2800
3000
3200
3429
7
8
7
7
7
8
12
12
14
15
16
15
data frame in a superframe. Prior to transmission of B1, the scrambler, trellis

encoder, differential encoder, and precoding filter tap delay line are initialized to zeros.
The 4D symbol intervals are indicated by the time index m s 4 i, where j
s 0, 1, 2, 3. is a cyclic time index that indicates the position of the 4D symbol
interval in a mapping frame. The 2D symbol intervals are indicated by the
time index n s 2 m q k, where k s 0, 1. is a cyclic time index that indicates
the position of the 2D symbol interval in a 4D symbol interval.
Mapping Frame Switching. An integer number of data bits are transmitted
in every data frame. The total number of primary and auxiliary channel data
bits transmitted in a data frame is denoted by
N s R = 0.28rJ
where R is the sum of the primary channel and auxiliary channel data rates.
The total number of primary and auxiliary. data bits transmitted in a
mapping frame varies between b y 1 low frame. and b high frame. bits
according to a periodic switching pattern SWP., of period P, such that the
average number of data bits per mapping frame is NrP. the value of b is
defined as the smallest integer not less than NrP. The number of high
frames in a period is the remainder:
r s N y b y 1. P
for 1 F r F P
SWP is represented by 1216-bit binary numbers where 0 and 1 represent

low and high frames, respectively. The leftmost bit corresponds to the first
mapping frame in a data frame. The rightmost bit is always 1.
SWP may be derived using an algorithm that employs a counter as follows:
Prior to each data frame the counter is set to zero. The counter is incremented by r at the beginning of each mapping frame. If the counter is less
than P, send a low frame; otherwise, send a high frame and decrement the
counter by P.
185
DATA MODEMS
TABLE 8.13
b and Switching Pattern (SWP) as a Function of Data Rate and Symbol Rate
2400 symr s 2743 symr s 2800 symr s 3000 symr s 3200 symr s
P s 12
P s 12
P s 14
P s 15
3429 symr
P s 16
P s 15
Data
Rate, R
SWP
SWP
SWP
SWP
SWP
SWP
2400
2600
4800
5000
7200
7400
9600
9800
12,000
12,200
14,400
14,600
16,800
17,000
19,200
19,400
21,600
21,800
24,000
24,200
26,400
26,600
28,800
29,900
8
9
16
17
24
25
32
33
40
41
48
49
56
57
64
65
72
73
FFF
6DB
FFF
6DB
FFF
6DB
FFF
6DB
FFF
6DB
FFF
6DB
FFF
6DB
FFF
6DB
FFF
6DB
14
15
21
22
28
29
35
36
42
43
49
50
56
57
63
64
70
71
FFF
56B
FFF
56B
FFF
56B
FFF
56B
FFF
56B
FFF
56B
FFF
56B
FFF
56B
FFF
56B
14
15
21
22
28
28
35
35
42
42
48
49
55
56
62
63
69
70
1BB7
0489
15AB
0081
0A95
3FFF
0489
1FBF
0081
1BB7
3FFF
15AB
1FBF
0A95
1BB7
0489
15AB
0081
13
14
20
20
26
27
32
33
39
39
45
46
52
52
58
59
64
65
71
71
3DEF
1249
0421
3777
2D6B
0081
7FFF
2AAB
14A5
3FFF
3DEF
1249
0421
3777
2D6B
0081
7FFF
2AAB
14A5
3FFF
12
13
18
19
24
25
30
31
36
37
42
43
48
49
54
55
60
61
66
67
72
73
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
FFFF
5555
12
12
17
18
23
23
28
29
34
35
40
40
45
46
51
51
56
57
62
63
68
68
0421
36DB
3DEF
0889
14A5
3F7F
7FFF
1555
2D6B
0001
0421
36DB
3DEF
0889
14A5
3F7F
7FFF
1555
2D6B
0001
0421
36DB
Table 8.13 gives the values for b and SWP for all combinations of data
rate and symbol rate. In the table, SWP is represented by a hexadecimal
number. For example, at 19,200 bps and symbol rate 3000, SWP is 0421 hex.
or 0000 0100 0010 0001 binary..
Multiplexing of Primary and Auxiliary Channel Bits. The auxiliary channel
bits are time division multiplexed with the scrambled primary channel bits.
The number of auxiliary channel bits transmitted per data frame is W s 8
at symbol rates 2400, 2800, 3000, and 3200, and W s 7 at symbol rates 2743
and 3429. In each mapping frame, the bit I1 i, 0 is used to send either an
auxiliary channel bit or a primary channel bit according to the auxiliary
channel multiplexing pattern, AMP, of period P see Figure 8.19.. AMP can
be represented as a P-bit binary number where a 1 indicates that an auxiliary
channel bit is sent and a 0 indicates that a primary channel bit is sent. AMP
depends only on the symbol rate and is given in Table 8.14 as a hexadecimal
number. The leftmost bit corresponds to the first mapping frame in a data
frame.
186
Figure 8.19. Encoder block diagram. (From Figure 4rV.34, ITU-T Rec. V.34, Ref. 12.)
TABLE 8.14
Auxiliary Channel Multiplexing Parameters
Symbol Rate, S
AMP
2400
2743
2800
3000
3200
3429
8
7
8
8
8
7
12
12
14
15
16
15
6DB
56B
15AB
2AAB
5555
1555
The auxiliary channel multiplexing pattern may be derived using an

algorithm similar to the algorithm for SWP, the frame switching pattern.
Prior to each data frame, a counter is set to zero. The counter is incremented
by W at the beginning of each mapping frame. If the counter is less than P, a
primary channel bit is sent; otherwise, an auxiliary channel bit is sent, and
the counter is decremented by P.
8.5.6.5 Encoder. Figure 8.19 is a functional block diagram of the encoder.
Signal Constellation. Signal constellations consist of complex-valued signal
points that lie on a two-dimensional rectangular grid.
All signal constellations used with the V.34 modem are subsets of a
960-point superconstellation. Figure 8.20 shows one-quarter of the points in
DATA MODEMS
187
Figure 8.20. One-quarter of the points in the superconstellation. (From Figure 5r V.34, page
14, ITU-T Rec. V.34, Ref. 12.)
the superconstellation. These points are labeled with decimal integers between 0 and 239. The point with the smallest magnitude is labeled 0, the
point with the next larger magnitude is labeled 1, and so on. When two or
more points have the same magnitude, the point with the greatest imaginary
component is taken first. The full superconstellation is the union of the four
quarter-constellations obtained by rotating the constellation in Figure 8.20 by
0, 90, 180, and 270.
A signal constellation with L points consists of Lr4 points from the
quarter-constellation in Figure 8.20 with labels 0 through Lr4 y 1, along
with the 3 Lr4 points that are obtained by 90, 180, and 270 rotations of
these signal points.
8.5.7 The V.90 ModemMaximum Data Rate: 56 kbps
8.5.7.1. Principal Characteristics. The V.90 modem Ref. 12. is designed
for full duplex operation over the PSTN. It normally operates on a two-wire
basis where interference between the send and receive channels is minimized
188
using local echo cancellation techniques. The modulation downstream toward the end-user . is PCM at 8000 symbols per second. It provides synchronous data signaling rates in the downstream direction from 28,000 to
56,000 bps in increments of 8000r6 bps.
In the upstream direction toward the local serving exchange or other,
similar node. V.34 Section 8.5.6. modulation is employed. The bit stream is
synchronous with upstream data rate from 4800 bps to 28,800 bps in increments of 2400 bps, with optional support of 31,200 and 33,600 bps.
Adaptive techniques are used which enable the modem to achieve close to
the maximum data signaling rates the channel can support on each connection. The modem will also negotiate V.34 operation if a connection will not
support V.90 operation. When the modem is started up, it exchanges rate
sequences to establish the operational data signaling rate.
8.5.7.2 Definitions
Analog Modem. The analog modem is the modem of the pair that, when in
data mode, generates ITU-T Rec. V.34 signals and receives ITU-T Rec.
G.711 PCM signals that have been passed through a G.711 decoder. The
modem is typically connected to the PSTN.
Note: ITU-T G.711, Pulse Code Modulation of Voice Frequencies,
deals with the coding of PCM signals derived from audio signals, as well as
with the decoding of those signals back to their audio equivalents. See
Chapter 7 for details.
Digital Modem. The digital modem is the modem of the pair, that, when in
the data mode, generates ITU-T Rec. G.711 signals and receives ITU-T
Rec. V.34 signals that have been passed through a Rec. G711 encoder.
The modem is connected to a digitally switched network through a digital
interface we.g., a basic rate interface BRI. or a primary rate interface
PRI.x. For a discussion of BRI and PRI, see Chapter 13..
Downstream. Transmission in the direction from the digital modem toward
the analog modem.
Uchord. Ucodes are grouped into eight Uchords. Uchord 1 contains Ucodes 0
to 15; Uchord 2 contains Ucodes 16 to 31; . . . ; and Uchord 8 contains
Ucodes 112 to 127.
Ucode. The universal code used to describe both a -law and A-law PCM
codeword. All universal codes are given in decimal notation in Table 8.15.
The -law and A-law codewords are the octets to be passed to the digital
interface by the digital modem and are given in hexadecimal notation. All
modifications defined in ITU-T Rec. G.711 have already been made. The
MSB in the -law PCM and A-law PCM columns in Table 8.15 correspond to the polarity bit of the Rec. G.711 character signals. A linear
representation of each PCM code word is also given.
DATA MODEMS
TABLE 8.15
189
The Universal Set of PCM Codewords
Ucode
-Law
PCM
-Law
Linear
Ucode
-Law
PCM
-Law
Linear
A-Law
PCM
A-Law
Linear
FF
D5
33
DE
428
F4
560
FE
D4
24
34
DD
460
F7
592
FD
16
D7
40
35
DC
492
F6
624
FC
24
D6
56
36
DB
524
F1
656
FB
32
D1
72
37
DA
556
F0
688
FA
40
D0
88
38
D9
588
F3
720
F9
48
D3
104
39
D8
620
F2
752
F8
56
D2
120
40
D7
652
FD
784
F7
64
DD
136
41
D6
684
FC
816
F6
72
DC
152
42
D5
716
FF
848
10
F5
80
DF
168
43
D4
748
FE
880
11
F4
88
DE
184
44
D3
780
F9
912
12
F3
96
D9
200
45
D2
812
F8
944
13
F2
104
D8
216
46
D1
844
FB
976
14
F1
112
DB
232
47
D0
876
FA
1008
15
F0
120
DA
248
48
CF
924
E5
1056
16
EF
132
C5
264
49
CE
988
E4
1120
17
EE
148
C4
280
50
CD
1052
E7
1184
18
ED
164
C7
296
51
CC
1116
E6
1248
19
EC
180
C6
312
52
CB
1180
E1
1312
20
EB
196
C1
328
53
CA
1244
E0
1376
21
EA
212
C0
344
54
C9
1308
E3
1440
22
E9
228
C3
360
55
C8
1372
E2
1504
23
E8
244
C2
376
56
C7
1436
ED
1568
24
E7
260
CD
392
57
C6
1500
EC
1632
25
E6
276
CC
408
58
C5
1564
EF
1696
26
E5
292
CF
424
59
C4
1628
EE
1760
27
E4
308
CE
440
60
C3
1692
E9
1824
28
E3
324
C9
456
61
C2
1756
E8
1888
29
E2
340
C8
472
62
C1
1820
EB
1952
30
E1
356
CB
488
63
C0
1884
EA
2016
31
E0
372
CA
504
64
BF
1980
95
2112
32
DF
396
F5
528
65
BE
2108
94
2240
A-Law
PCM
A-Law
Linear
190
TABLE 8.15
( Continued )
Ucode
-Law
PCM
-Law
Linear
A-Law
PCM
A-Law
Linear
Ucode
-Law
PCM
-Law
Linear
A-Law
PCM
A-Law
Linear
66
67
BD
2236
97
2368
97
9E
8828
B4
8960
BC
2364
96
2496
98
9D
9340
B7
9472
68
BB
2492
91
2624
99
9C
9852
B6
9984
69
BA
2620
90
2752
100
9B
10364
B1
10496
70
B9
2748
93
2880
101
9A
10876
B0
11008
71
B8
2876
92
3008
102
99
11388
B3
11520
72
B7
3004
9D
3136
103
98
11900
B2
12032
73
B6
3132
9C
3264
104
97
12412
BD
12544
74
B5
3260
9F
3392
105
96
12924
BC
13056
75
B4
3388
9E
3520
106
95
13436
BF
13568
76
B3
3516
99
3648
107
94
13948
BE
14080
77
B2
3644
98
3776
108
93
14460
B9
14592
78
B1
3772
9B
3904
109
92
14972
B8
15104
79
B0
3900
9A
4032
110
91
15484
BB
15616
80
AF
4092
85
4224
111
90
15996
BA
16128
81
AE
4348
84
4480
112
8F
16764
A5
16896
82
AD
4604
87
4736
113
8E
17788
A4
17920
83
AC
4860
86
4992
114
8D
18812
A7
18944
84
AB
5116
81
5248
115
8C
19836
A6
19968
85
AA
5372
80
5504
116
8B
20860
A1
20992
86
A9
5628
83
5760
117
8A
21884
A0
22016
87
A8
5884
82
6016
118
89
22908
A3
23040
88
A7
6140
8D
6272
119
88
23932
A2
24064
89
A6
6396
8C
6528
120
87
24956
AD
25088
90
A5
6652
8F
6784
121
86
25980
AC
26112
91
A4
6908
8E
7040
122
85
27004
AF
27136
92
A3
7164
89
7296
123
84
28028
AE
28160
93
A2
7420
88
7552
124
83
29052
A9
29184
94
A1
7676
8B
7808
125
82
30076
A8
30208
95
A0
7932
8A
8064
126
81
31100
AB
31232
96
9F
8316
B5
8448
127
80
32124
AA
32256
Source: Table 1r V.90, ITU-T Rec. V. 90, ITU Geneva, September 1998, Ref. 13.
DATA MODEMS
191
8.5.7.3 Overview of V.90 Operation. As we know, the analog voice channel is limited to the frequency band of 3003400 Hz. This severely limits the
maximum data rate to be carried on that channel. In more advanced
countries the analog voice channel terminates the PCM channel banks at the
local serving switch central office.. Advantage is taken of this fact by the
V.90 modem, which emulates the PCM analog-to-digital conversion and
coding in the downstream direction. Depending on the quality of the subscriber loop, nearly 56 kbps can be achieved.
The V.90 modem operates asymmetrically providing the higher bit rate
downstream. In the upstream direction, the V.90 modem operates just like
the V.34 modem see Section 8.5.6.. In this case, the theoretical maximum bit
rate achievable is 33,600 bps.
The V.90 modem operates in a digital mode in the upstream direction. It
is very sensitive to quantization noise of the connecting circuits to the distant
end-user. If it detects additional digital-to-analogranalog-to-digital signal
processing i.e., a hybrid network. in those intervening circuits, it will back off
and operate in the V.34 mode in the upstream direction.
As we will remember from Chapter 7, PCM transmits samples of voltage
levels with a discrete coding process. There are a theoretical 256 coding steps
to express a voltage 255 steps in the real world.. Companding is utilized,
either A-law or -law, where, as a voltage level measurement approaches
zero, more and more companding steps are available. The PCM gradations
i.e., distance between discrete steps. are so small that they can be obliterated by noise. Thus, with the V.90 modem, only the higher levels steps
127255., are utilized. There are 128 such levels. The V.90 ships data in
chunks of 8 bits which some may wish to call a byte .. The transport rate is
8000 bytes or octets per second thats the Nyquist sampling rate.. Of course
this is the same rate used by the PSTN. The 8000 bytes per second of the
V.90 must be synchronized with the 8000 time slots a second of the digital
PSTN.
The V.90 constantly probes the downstream bit stream operation. When it
detects degradation, it drops back 92 levels for 52-kbps operation. Using
fewer levels provides more robust operation, but at a lower rate.
The V.90 consists of two modems: the digital modem and the analog
modem.
8.5.7.4 V.90 Encoder. Figure 8.21 is an overview of the encoder and
represents one data frame. Data frames in the digital modem have a
six-symbol structure. Each symbol position within the data frame is called a
data frame interal and is indicated by a time index, I s 0, . . . , 5, where I s 0
is the first in time. Frame synchronization between the digital modem
transmitter and analog modem receiver is established during training procedures.
192
Figure 8.21. Digital modem encoder block diagram. (From Figure 1r V.90, ITU-T Rec. V.90,
Ref. 13.)
8.5.7.5 Mapping Parameters. Mapping parameters established during

training or rate renegotiation procedures are as follows:
Six PCM code sets, one for each data frame interval 05, where data
frame interval I has Mi members
K, the number of modulus encoder input data bits per data frame
Sr , the number of PCM code sign bits per data frame used as redundancy for spectral shaping and
S, the number of spectral shaper input data bits per data frame, where
S q Sr s 6. Table 8.16 shows the data signaling rates achieved by the
valid combinations of K and S during data mode. Table 17rV.90 in
reference publication . shows the valid combinations of K and S used
during phase 4 and rate renegotiation procedures
8.5.7.6 Input Bit Parsing. D equal to S q K . serial data bits, d 0 to d Dy1

where d 0 is first in time, are parsed into S sign input bits and K modulus
encoder bits. D 0 to d Sy1 form S0 to sSy1 and d S to d Dy1 form b 0 to bKy1 .
The K modulus encoder bits and the S sign bits are used as specified in the
paragraphs below.
8.5.7.7 Modulus Encoder. K bits enter the modulus encoder. The data
signaling rates associated with each value of K are tabulated in Table 8.16.
There are six independent mapping moduli, M0 to M5 , which are the number
of members in the PCM code sets defined for data frame interval 0 to data
frame interval 5, respectively. Mi is equal to the number of positive levels in
the constellation to be used in data frame interval I as signaled by the analog
DATA MODEMS
TABLE 8.16
Data Signaling Rates for Different K and S
K, Bits Entering
Modulus Encoder
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
193
S, Sign Bits Used for

User Data
Data Signaling
Rate, kbitr s
From
To
From
To
6
5
4
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
5
4
3
28
28
28
28
29 1r3
30 2r3
32
33 1r3
34 2r3
36
37 1r3
38 2r3
40
41 1r3
42 2r3
44
45 1r3
46 2r3
48
49 1r3
50 2r3
52
53 1r3
54 2r3
56
28
29 1r3
30 2r3
32
33 1r3
34 2r3
36
37 1r3
38 2r3
40
41 1r3
42 2r3
44
45 1r3
46 2r3
48
49 1r3
50 2r3
52
53 1r3
54 2r3
56
56
56
56
Source: Table 2r V.90, ITU-T Rec. V.90, ITU-Geneva, September 1998, Ref. 13.
modem using the CP sequences defined in paragraph 8.5.2 of the reference

document Ref. 13..
The values of Mi and K shall satisfy the inequality 2 K F 5is0 Mi .
The modulus encoder converts K bits into six numbers, K 0 to K 5 , using
the following algorithm.
Note: Other implementations are possible, but the mapping function must
be identical to that given in the algorithm described below.
1. Represent the incoming K bits as an integer, R 0 :
R 0 s b 0 q b1*2 1 q b 2*2 2 q qbKy1*2 Ky1
2. Divide R 0 by M0 . The remainder of this division gives K 0 , and the
quotient becomes R1 for use in the calculation for the next data frame
194
interval. Continue for the remaining five data frame intervals. This
gives K 0 to K 5 as
K i s R i modulo Mi ,
where 0 F K i - Mi ; R iq1 s R i y K i . rMi
3. The numbers K 0 , . . . , K 5 are the output of the modulus encoder,

where K 0 corresponds to data frame interval 0 and K 5 corresponds to
data frame interval 5.
8.5.7.8 Mapper. There are six independent mappers associated with the
six data frame intervals. Each mapper uses a tabulation of Mi PCM codes
that make up the positive constellation points of data frame interval i
denoted Ci . The PCM codes to be used in each data frame interval are
specified by the analog modem during training procedures. The PCM code
that is denoted by the largest smallest. Ucode is herein called the largest
smallest. PCM code. The members of Ci shall be labeled in descending
order so that label 0 corresponds to the largest PCM code in Ci , while label
Mi y 1 corresponds to the smallest PCM code in Ci . Each mapper takes K i
and forms Ui by choosing the constellation point in Ci labeled by K i .
8.5.7.9 Spectral Shaping. The digital modem output line signal spectrum
shall be shaped, if spectral shaping is enabled. Spectral shaping only affects
the sign bits of transmitted PCM symbols. In every data frame of six symbol
intervals, Sr sign bits are used as redundancy for spectral shaping while the
remaining S sign bits carry user information. The redundancy, Sr , is specified
by the analog modem during training procedures and can be 0, 1, 2, or 3.
When Sr s 0, spectral shaping is disabled.
Note: The initial state of the spectral shaper does not affect the performance of the analog modem and is therefore left to the implementor.
Sr s 0, S s 6. The PCM code sign bits, $ 0 to $ 5 , shall be assigned using
input sign bits s0 to s5 and a differential coding rule:
$ 0 s s0 [ $ 5 of the previous data frame . ; and
$ i s si [ $ iy1
for i s 1, . . . , 5
where [ stands for modulus-2 addition.

Sr s 1, S s 5. Sign bits S0 to S4 are parsed to one 6-bit shaping frame per
data frame in accordance with Table 8.16.
The odd bits are differentially encoded to produce the output pXj in
accordance with Table 8.17.
Finally, a second differential encoding shall be performed to produce the
initial shaping sign bit assignment, t j 0. to t j 5., using the rule
t j k . s pXj k . [ t jy1 k .
DATA MODEMS
TABLE 8.17
195
Parsing Input Sign Bits to Shaping Frames
Data Frame
Interval
Sr s 1, S s 5
p j (0) s 0
p j (0) s 0
p j (1) s s 0
p j (1) s s 0
p j (2) s s1
p j (2) s s1
p j (3) s s 2
p j q 1(0) s 0
p j q 1(1) s s1
p j (4) s s 3
p j q 1(1) s s 2
p j q 2 (0) s 0
p j (5) s s 4
p j q 1(2) s s 3
p j q 2 (1) s s 2
Sr s 2, S s 4
Sr s 3, S s 3
p j (0) s 0
p j (1) s s 0
p j q 1(0) s 0
Source: Table 3r V.90, ITU-T Rec. V.90, ITU Geneva, September 1998, Ref. 13.
TABLE 8.18
Odd Bit Differential Coding
Data Frame
Interval
S r s 1, S s 5
X
p j (0)
0
1
X
p j (1)
2
3
4
5
X
p j (0)
s0
s p j (1) [
X
p j (2)
S r s 2, S s 4
X
p j y 1(5)
s p j (2)
p j (3) s p j (3) [ p j (1)

X
p j (4)
X
p j (5)
s p j (4)
s p j (5) [
X
p j (3)
X
p j (1)
S r s 3, S s 3
X
p j (0) s 0
s0
s p j (1) [
X
p j y 1(1)
X
p j (2) s p j (2)
X
p j q 1(0) s 0
X
X
p j q 1(1) s p j q 1(1) [ p j (1)
X
p j q 1(2) s p j q 1(2)
X
p j (1)
X
s p j (1) [ p j y 1(1)
X
p j q 1(0) s 0
X
X
p j q 1(1) s p j q 1(1) [ p j (1)
X
p j q 2 (0) s 0
X
p j q 2 (1)
X
s p j q 2 (1) [ p j q 1(1)
The spectral shaper converts each bit t j k ., to a PCM code sign bit $ k as
described in Section 8.5.6.9.
Sr s 2, S s 4. Sign bits S0 to S3 are parsed to 3-bit shaping frames per data
frame as shown in Table 8.17. The odd bit in each shaping frame is
differentially encoded to produce differentially coded outputs pXj and pXjq1 in
accordance with Table 8.18.
Finally, a second differential encoding shall be performed on each shaping
frame to produce the initial shaping sign bit assignments t j 0. to t j 2. and
t jq1 0. to t jq1 2. using the differential encoding rule:
t jq1 k . s pXjq1 k . [ t j k .
The spectral shaper converts each bit t j k . to PCM code sign bit $ k and
converts each bit t jq1 k . to PCM code sign bit $ kq3 as described in Section
8.5.6.9.
196
Sr s 3, S s 3. The sign bits S0 to S2 are parsed to three 2-bit shaping

frames per data frame as shown in Table 8.17. The odd bits in each shaping
frame are differentially encoded to produce differentially encoded outputs
pXj , pXjy1 and pXjy2 as shown in Table 8.18.
Finally, a second differential encoding shall be performed on each shaping
frame to produce the initial shaping sign bit assignments t j 0. to t j 1., t jq1 0.
to t jq1 1., and t jq2 0. to t jq2 1. using the differential encoding rule:
t jq1 k . s pXjq1 k . [ t j k .
t jq2 k . s pXjq2 k . [ t jq1 k .
The spectral shaper converts each bit t j k . to PCM code sign bit $ k , each bit
t jq1 k . to PCM code sign bit $ kq2 , and each bit t jq2 k . to PCM code sign bit
$ kq4 as described in Section 8.5.6.9.
8.5.7.10 Spectral Shaper. The spectral shaper operates on a spectral
shaper frame basis. For the cases Sr s 2 and S r s 3, there are multiple
shaper frames per six-symbol data frame. Spectral shaper operation for each
shaper frame within a data frame called shaping frame j in this section. is
identical except that they affect different data frame PCM sign bits as shown
in Table 8.19.
The spectral shaper modifies the initial sign bits t j 0., t j 1., . . . . to corresponding PCM code sign bits $ 0 , $ 1 . . . . without violating the constraint
described below, so as to optimize a spectral metric. The constraint of the
spectral is described using the 2-state trellis diagram shown in Figure 8.22.
TABLE 8.19
Shaping Frame to Data Frame Signal Relationship
Data Frame
Data Frame
Interval
Sr s 1, S s 5
Sr s 2, S s 4
Sr s 3, S s 3
PCM Sign Bit
t j (0)
t j (0)
t j (0)
$0
t j (1)
t j (1)
t j (1)
$1
t j (2)
t j (2)
t j q 1(0)
$2
t j (3)
t j q 1(0)
t j q 1(1)
$3
t j (4)
t j q 1(1)
t j q 2 (0)
$4
t j (5)
t j q 1(2)
t j q 2 (1)
$5
DATA MODEMS
197
Figure 8.22. Trellis diagram used to constrain the spectral shaper.
In a given spectral shaping frame j, the spectral shaper shall modify the
initial sign sequence, t j k ., according to one of the following four sign
inversion rules:
Rule A: Do nothing.
Rule B: Invert all sign bits in spectral shaping frame j.

Rule C: Invert even-numbered t j 0., t j 2., etc.. sign bits in spectral
shaping frame j.
Rule D: Invert odd-numbered t j 1., t j 3., etc.. sign bits in spectral
shaping frame j.
The trellis diagram describes the sequence of sign inversion rules that are
allowable. For example, when the spectral shaper is in state Q j s 0 at the
beginning of frame j, only rules A and B are allowable in frame j. The
current state Q j together with the sign inversion rule selected for frame j
determine the next state Q jq1 according to the trellis diagram.
The look-ahead depth parameter, 1d, is an integer between 0 and 3
selected by the analog modem during training procedures. 1d of 0 and 1 are
mandatory in the digital modem. 1d of 2 and 3 are optional.
To select the sign inversion rule for the jth spectral shaping frame, the
spectral shaper shall use the PCM symbol magnitudes produced by the
mapper for spectral shaping frames j, j q 1, . . . , j q 1d. It shall compute
the spectral metric that would result if each of the allowable sequences of
sign inversion rules for frames j through j q 1d, starting from the current
state Q j in frame j, were to be used. The shaper shall select the sign
inversion rule for frame j that minimizes the spectral metric, ww n x, defined in
Section 8.5.6.9 up to and including the final symbol of spectral shaping frame
j q 1d. The selection determines the next state Q jq1 .
The shaper shall then set PCM code signs $ i for shaping frame j according
to the selected sign inversion rule for shaper frame j.
198
8.5.7.11 Spectral Shape Filter. The analog modem determines the spectral shape filter function used in the digital modem by selecting parameters
of the following transfer function:
T z. s
1 y a1 zy1 . 1 y a2 zy1 .
1 y b1 zy1 . 1 y b2 zy1 .
where a1 , a2 , b1 , and b 2 are parameters having absolute values less than or

equal to 1. The parameters, a1 , a2 , b1 , and b 2 , are specified by the analog
modem during training procedures and are represented in the 8-bit twoscomplement format with 6 bits after the binary point. The digital modem
shall perform the spectral shaping according to the spectral shaping metric,
ww n x characterized by the filter:
F z. s
1
T z.
1 y b1 zy1 . 1 y b2 zy1 .
1 y a1 zy1 . 1 y a2 zy1 .
The filter input, x w n x, shall be a signed signal proportional to the linear value
corresponding to PCM codes being transmitted. The relationship between
PCM codes and corresponding linear values is given in Table 1rV.90. ww n x
shall be computed as follows:
1. yw n x s x w n x y b1 x w n y 1x q a1 yw n y 1x
2. w n x s yw n x y b 2 yw n y 1x q a2 w n y 1x
3. ww n x s 2 w n x q ww n y 1x
8.5.7.12 Sign Assignment. Six sign bits generated by the spectral shaper
described above are attached to six unsigned mapper outputs U0 to U5 to
complete the mapping of the data frame intervals. A sign bit 0 means that the
transmitted PCM codeword represents a negative voltage and a sign bit 1
means it will represent a positive voltage.
8.5.7.13 MUX. The signed PCM codewords, PCM i , are transmitted from
the digital modem sequentially with PCM 0 being first in time.
8.5.7.14 Analog Modem. The characteristics of the analog modem described in this section apply when in the V.90 mode. After fallback to the
V.34 mode, the analog modem has the characteristics as defined in ITU-T
Rec. V.34 Section 24-7.9..
Data Signaling Rates. The modem supports data rates of 4800 to 28,800 bps
in increments of 2400 bps, with optional support for 31,200 and 33,600 bps.
DATA MODEMS
199
The 200-bps V.34 auxiliary channel is not supported. The data rate is
determined during phase 4 of the modem start-up in accordance with the
procedures described in paragraph 9.4 of the reference document.
Symbol Rates. The analog modem supports the symbol rate 3200. It may
also support 3000 and the optional symbol rate 3429 as defined in ITU-T
Rec. V.34 Section 8.5.5.. The other V.34 symbol rates2400, 2743, and
2800are not supported. The symbol rate is selected by the analog modem
during phase 2 of modem start-up in accordance with paragraph 9.2 of the
reference publication.
Carrier Frequencies. The analog modem supports the carrier frequencies
specified in ITU-T Rec. V.34, paragraph 5.3 Section 8.5.5. for the appropriate symbol rate. The carrier frequency is determined during phase 2 of
modem start-up in accordance with the procedures specified in paragraph 9.2
of the reference document.
Preemphasis. The analog modem supports the preemphasis filter characteristics specified in ITU-T Rec. V.34, paragraph 5.4. The filter selection is
provided by the digital modem during phase 2 of modem start-up in accordance with the procedures specified in paragraph 9.2 of the reference
document.
Scrambler. The analog modem includes a self-synchronizing scrambler as
specified in ITU-T Rec. V.34, Section 7, using the generating polynomial,
GPA, given in Equation 7.2rV.34 Section 8.5.5..
Framing. The analog modem uses the framing method specified for the V.34
primary channel described in Section 8 of ITU-T Rec. V.34.
Encoder. The analog modem uses the encoder specified for the V.34 primary channel in Section 9, ITU-T Rec. V.34.
Interchange Circuits. The requirements of this section apply to both modems.
List of Interchange Circuits (V.34rV.90). Reference in this section to ITU-T
Rec. V.24 interchange circuit numbers are intended to refer to the functional
equivalent of such circuits and are not intended to imply the physical
implementation of such circuits. For example, references to circuit 103
should not be understood to refer to the functional equivalent of circuit 103
see Table 8.20..
200
TABLE 8.20
Interchange Circuits
Interchange Circuit
No.
Description
Notes
102
103
104
105
106
Signal ground or common return

Transmitted data
Received data
Request to send
Ready for sending
107
108r1 or
108r 2
109
Data set ready

Connect data set to line
Data terminal ready
Data channel received line signal detector
125
133
Calling indicator
Ready for receiving
Note 1: Thresholds and response times are not applicable because a line signal
detector cannot be expected to distinguish received signals from talker echoes.
Note 2: Operation of circuit 133 shall be in accordance with 4.2.1.1 of Recom mendation V.43.
Source: Table 6rV.90, ITU-T Rec. V.90, ITU Geneva, September 1998, Ref. 13.
Asynchronous Character-Mode Interfacing. The modem may include an

asynchronous-to-synchronous converter interfacing to the DTE in an asynchronous or start-stop character . mode. The protocol for the conversion is in
accordance with ITU-T Recs. V.14, V.42, and V.80. Data compression may
also be employed.
Portions of Section 8.5.7 have been extracted from ITU-T Rec. V.90 of 9r98, Ref. 13.
REFERENCES
1. Roger L. Freeman, Telecommunication Transmission Handbook, 4th ed., John
2. Attenuation Distortion, CCITT Rec. G.132, Fascicle III.1, IXth Plenary Assembly,
Melbourne, 1988.
3. Group Delay Distortion, CCITT Rec. G.133, Fascicle III.1, IXth Plenary Assembly,
Melbourne, 1988.
4. Characteristics of Special Quality International Leased Circuits with Special Bandwidth Conditioning, CCITT Rec. M.1020, ITU Geneva, March 1993.
5. Data Communications Using Voiceband Priate Line Channels, Bell System Technical Reference, PUB 41004, AT & T, New York, 1973.
6. Transmission Systems for Communications, 5th ed., Bell Telephone Laboratories,
Holmdel, NJ, 1982.
7. 9600 Bits per Second Modem Standardized for Use on Point-to-Point 4-Wire Leased
Telephone Circuits, CCITT Rec. V.29, Fascicle, VIII.1, IXth Plenary Assembly,
Melbourne, 1988.
REFERENCES
201
8. A Family of 2-Wire, Duplex Modems Operating at Data Signaling Rates up to 9600

Bitsr s for Use on the General Switched Telephone Network and on Leased Telephone-Type Circuits, CCITT Rec. V.32, ITU Geneva, March 1993.
9. 14,400 Bits per Second Modem Standardized for Use on Point-to-Point 4-Wire Leased
Telephone-Type Circuits, CCITT Rec. V.33, Fascicle VIII.1, IXth Plenary Assembly, Melbourne, November 1988.
10. Gottfried Ungerbock,
Trellis-Coded Modulation with Redundant SetsParts I
and II, IEEE Communications Magazine, Vol. 27, No. 2, February 1987.
11. Andrew J. Viterbi et al., A Pragmatic Approach to Trellis-Coded Modulation,
IEEE Communication Magazine, Vol. 27, No. 7, July 1987.
12. A Modem Operating at Data Signaling Rates of up to 33,600 Bitsrs for Use on the
General Switched Telephone Network and on Leased Point-to-Point 2-Wire Telephone-Type Circuits, ITU-T Rec. V.34, ITU Geneva, February 1998.
13. A Digital Modem and Analog Modem Pair for Use on the Public Switched Telephone
Network PSTN . at Data Signaling Rates of Up to 56,000 Bitsrs Downstream and
Up to 33,600 Bitsrs Upstream, ITU-T Rec. V.90, ITU Geneva, September 1998.
9
DATA COMMUNICATIONS IN THE
OFFICE ENVIRONMENT, PART 1
9.1 INTRODUCTION
In the present office environment the personal computer PC. is nearly
ubiquitous. Some may be stand-alone devices, but in the majority of cases
they are interconnected by one means or another. They may also connect, via
a server, to a mainframe computer, high-speed printers, and mass storage
devices.
A workstation consists of a PC, and an interconnection capability is
implied. It is defined by the IEEE Ref. 1. as a device used to perform tasks
such as data processing and word processing. Often a workstation is just a PC
with keyboard and display. An office may be viewed as a conglomeration of
management information systems MIS.. MIS represents the entire electronic
data processing capabilities of an organization.
The interconnection of MIS assets on a local level is the subject of this
chapter. This interconnection is almost universally carried out by means of a
local area network LAN.. The IEEE Ref. 1. defines a LAN as a nonpublic
data network in which serial transmission is used without store and forward
techniques for direct data communication among data stations located on the
users premises.
A LAN is made cost-effective because a single transmission medium
interconnects all devices on the network. There are several simple assumptions on why we can do this with just one transmission medium such as a wire
pair. First, transactions are usually of very short duration. Consider a LAN
transmitting at a 10-Mbps data rate. It transmits 10,000 bits in 1 ms, or one
page of text of this book. Now if each user only transmits for a period of
1 ms, and we allow 1 ms between transactions for circuit setup, then in
1 second, 500 users could be accommodated. It is also apparent that a user
would not have to wait very long to access the medium.
203
204
DATA COMMUNICATIONS IN THE OFFICE ENVIRONMENT, PART 1
Another reason that this can be accomplished with one medium such as a
wire pair is that distances are very short, probably never more than 200 m
500 ft.. Furthermore, a wire pair over comparatively short distances can
handle considerably more than a 10-Mbps data rate. 100 Mbps and even
1 Gbps are not uncommon rates with modern systems such as Ethernet
CSMArCD..
In the following sections we discuss several of the more popular nonproprietary. LANs which will likely be encountered in an office, factory, or
campus environment.
9.2 DISTINGUISHING CHARACTERISTICS OF LANS

A LAN covers a limited geographical area. In fact the maximum extension of
a LAN is a major factor in its design and selection. A LAN may have an
extension of less than 100 ft covering a portion of a floor in a building. It may
cover not only an entire floor but several floors. Some LANs can extend over
a large university campus. A LAN has a single owner; it is a private network.
Its principal purpose is cost sharing of high-value computer assets.
LANs have much higher data rates when compared to wide-area networks
WANs.. Many operational WANs use transport systems optimized for voice;
data transmission is a compromise. In other words, WANs are constrained by
the intervening media; LANs are only constrained by the laws of physics.
A common bit rate for LANs still is 10 Mbps. 100 Mbps and 1 Gbps are
becoming equally prevalent. These systems are optimized for data, and
excellent BER performance values can be achieved without error correction,
from 1 = 10y8 to 1 = 10y12 and better. There are three different transmission media used with LANs: wire pair, coaxial cable, and fiber optics.* The
popularity of coaxial cable is diminishing. Wire pair is the transmission
medium of choice, although comparatively distance-limited. Fiber optics,
however, is rapidly taking its rightful place in the enterprise environment. It
is the least constrained for data rate among the three transmission media.
Another advantage of LANs is that they permit coupling of multivendor
equipments into an integral system, particularly when based on open system
interconnection OSI. Chapter 2.. However, it should be pointed out that
connectivity via OSI layer 1 does not guarantee compatibility of layer 2 and
above with equipment of different vendors. Such compatibility can often be
achieved by conversion software.
LANs can be extended using repeaters; LANs can also be interconnected
among themselves using devices called bridges. Smart bridges and routers can
direct traffic to other disparate LANs or to a WAN see Section 10.3..
*There are radio-based LANs, now popularly called wireless LANs Section 10.1.5..
DISTINGUISHING CHARACTERISTICS OF LANs
205
There are two generic transmission techniques utilized by LANs: baseband

and broadband. Baseband transmission in this context can be loosely defined
as the direct application of the baseband signal to the transmission medium.
Broadband transmission regarding LANs. is where the baseband signal from
the data device is translated in frequency to a particular band of frequencies
in radio-frequency spectrum and applied to a coaxial cable medium. Broadband transmission requires a modem to carry out the translation. Baseband
transmission may require some sort of signal conditioning device. With
broadband transmission we usually think of simultaneous multiple carriers
that are separated in the frequency domain. Broadband LANs derive from
cable TV technology.
Baseband LANs have completely dominated the marketplace. For this
reason, considerably more space in this chapter is dedicated to baseband
LANs than to broadband LANs. Different LANs are defined by the transmission medium, the bit rates., and the protocol.
Many of the widely used LAN protocols have been developed or standardized in North America through the offices of the Institute of Electrical and
Electronic Engineers IEEE.. The American National Standards Institute
ANSI. has subsequently accepted and incorporated these standards, and
they now bear the ANSI imprimatur. A notable exception is FDDI, which is
completely ANSI-sponsored. The IEEE develops LAN standards in the
IEEE 802 committee, which is currently organized into these subcommittees:
802.1*
High-Level Interface
802.2*
Logical Link Control LLC.
802.3*
CSMArCD Networks
802.4*
Token Bus Networks
802.5*
Token Ring Networks
802.6*
Metropolitan Area Networks
802.7*
Broadband Technical Advisory Group
802.8
Fiber-Optic Technical Advisory Group
802.9*
Integrated Data and Voice Networks
802.10*
Standard for Interoperable LAN Security
802.11*
Wireless LANs
802.12*
Demand Priority Working Group
802.13
802.14*
Cable TV Working Group
802.15
Wireless Personal Area Working Group
802.16
Broadband Wireless Access LMDS. Working Group
Asterisk *. indicates published standards .
In some applications, the baseband modulates a radio frequency and then the resulting signal
is applied to the cable medium.
206
Figure 9.1. LAN architecture related to OSI.
9.3 HOW LAN PROTOCOLS RELATE TO OSI

LAN protocols utilize only OSI layers 1 and 2, the physical and data-link
layers, respectively. The data-link layer is split into two sublayers: medium
access control MAC. and logical link control LLC.. These relationships to
OSI are shown in Figure 9.1.
The sublayers, logical link control and medium access control, carry out
four basic functions:
1. They provide one or more service access points SAPs.. An SAP is a
logical interface between two adjacent layers.
2. Before transmission, they assemble data into a frame with address and
error-detection fields.
3. On reception, they disassemble the frame and perform address recognition and error detection.
4. They manage communications over the link.
The first function and those related to it are performed by the LLC sublayer.
The last three functions listed are handled by the MAC sublayer. Section 9.4
provides a brief description of LLC operation. Section 9.5 gives overviews of
several MACs and their related protocols.
9.4 LOGICAL LINK CONTROL (LLC)
The LLC sublayer constitutes the top sublayer in the data-link layer as shown
in Figure 9.2. The LLC is common to various MAC methods that are defined
and supported by IEEE 802 committees wISO 8802 activity Ref. 2.x.
LOGICAL LINK CONTROL (LLC)
207
Figure 9.2. LLC relationship in the LAN

reference model.
The LLC is a service specification to the network layer OSI layer 3., to
the MAC sublayer, and to the LLC sublayer management function. The
service specification to the network layer provides a description of the
various services that the LLC sublayer, plus underlying layers and sublayers,
offers to the network layer. The service specification to the MAC sublayer
provides a description of the services that the LLC sublayer requires of the
MAC sublayer. These services are defined so as to be independent of both
a. the form of the medium access methodology and b. the nature of the
medium itself. The service specification to the LLC sublayer management
function provides a description of the management services that are provided
to the LLC sublayer. All of these service specifications are given in the form
of primitives that represent in an abstract way the logical exchange of
information and control between the LLC sublayer and the identified service
function i.e., network layer, MAC sublayer, or LLC sublayer management
function..
The LLC provides two types of data-link operation. The first type of
operation is a data-link connectionless mode of service across a data link with
minimum protocol complexity. This type of operation is attractive when
higher layers provide any essential recovery and sequencing services so that
these do not need replicating in the data-link layer. In addition, this type of
operation may prove useful in applications where it is not essential to
guarantee the delivery of every data-link layer data unit. The connectionless
mode is described in terms of logical data links.
The second type of operation provides a data-link connection-mode service across a data link which is comparable to such protocol procedures as
high-level data-link control HDLC; see Chapter 3.. This service includes
support of sequenced delivery of data units and a comprehensive set of
data-link layer error recovery techniques. This second type of service is
described in terms of data-link connections.
The LLC standard Ref. 2. identifies two distinct classes of LLC
operation. Class I provides data-link connectionless-mode service only. Class
II provides data-link connection-mode service plus data-link connectionlessmode service. Either class of operation may be supported.
208
Figure 9.3. Service primitives.
The basic LLC protocols are peer protocols for use in multistation,
multiaccess environments. Because of the multistation, multiaccess environment, it is possible for a station to be involved in a multiplicity of peer
protocol data exchanges with a multiplicity of different stations over a
multiplicity of different logical data links andror data-link connections that
are carried by a single physical layer PHY. over a single physical medium.
Each unique tofrom pairing at the data-link layer defines a separate logical
data link or data-link connection with separate logical parameters and
variables.
9.4.1 LLC Sublayer Services and Primitives
The services of a layer or sublayer. are the capabilities which it offers to a
user in the next higher layer or sublayer.. In order to provide its service, a
layer or sublayer. builds its functions on the services which it requires from
the next lower layer or sublayer.. Figure 9.3 illustrates the four generic
primitives in relation to service providers and service users.
Services are specified by describing the information flow between the
N-user and the N-layer or sublayer.. The information flow is modeled by
discrete, instantaneous events, which characterize the provision of service.
Each event consists of passing a service primitive from one layer or sublayer.
to the other through an N-layer or sublayer. SAP associated with the
N-user. Service primitives convey the information required in providing a
particular service. These service primitives are an abstraction in that they
209
specify only the service provided rather than the means by which the service
is provided. This definition of service is independent of any particular
interface implementation.
Services are specified by describing the service primitives and parameters
that characterize each service. A service may have one or more related
primitives that constitute the activity that is related to the particular service.
Each service primitive may have zero or more parameters that convey the
information required to convey the service.
There are four generic primitives described as follows:
Request. The request primitive is passed from the N-user to the N-layer
or sublayer. to request that a service be initiated.
Indication. The indication primitive is passed from the N-layer or sublayer. to the N-user to indicate an internal N-layer or sublayer. event
which is significant to the N-user. This event may be logically related to
a remote service request, or may be caused by an event internal to the
N-layer or sublayer..
Response. The response primitive is passed from the N-user to the N-layer
or sublayer. to complete a procedure previously invoked by an indication primitive.
Confirm. The confirm primitive is passed from the N-layer or sublayer. to
the N-user to convey the results of one or more associated previous
service requests..
Note the similarity with HDLC described in Chapter 3..
Possible relationships among primitive types are shown by means of
time-sequence diagrams in Figure 9.4. The figure also shows the logical
relationship of primitive types. Primitive types that occur earlier in time
connected by dotted lines in the figure. are the logical antecedents of
subsequent primitive types.
9.4.1.1 Unacknowledged Connectionless-Mode Service. This is defined as the data transfer service that provides the means by which network
entities LAN accesses . can exchange link service data units LSDUs. without
the establishment of a data-link connection. The data transfer can be
point-to-point, multicast, or broadcast.
9.4.1.2 Connection-Mode Services. This set of services provides the
means for establishing, using, resetting, and terminating data-link layers
connections. These connections are point-to-point connections between linklayer service access points LSAPs..
The connection establishment service provides a means by which a network entity can request, or be notified of, the establishment of data-link layer
connections.
210
Figure 9.4. Time-sequence diagram.
The connection-oriented data transfer service provides the means by

which a network entity can send or receive LSDUs, over a data-link layer
connection. This service also provides data-link layer sequencing, flow control, and error recovery.
The connection reset service provides the means by which established
connections can be returned to their initial state.
The connection termination service provides the means by which a network entity can request, or be notified of, the termination of data-link layer
connections.
The connection flow control service provides the means to control the flow
of data associated with a specified connection, across the network
layerrdata-link layer interface.
9.4.1.3 Type 1 and Type 2 Operation
Type 1 Operation. With Type 1 operation, protocol data units PDUs. are
exchanged between LLCs without the need for establishment of a
data-link connection. In the LLC sublayer these PDUs are not acknowledged, nor is there any flow control or error recovery in the Type 1
procedures.
211
Figure 9.5. LLC PDU format. DSAP Address s destination service access point address field;
SSAP Address s source service access point address field; Control s control field (16 bits for
formats that include sequence numbering, and 8 bits for formats that do not; Information s
information field; * s multiplication; M s an integer value equal to or greater than 0. (Upper
bound of M is a function of the medium access control methodology used.)
Type 2 Operation. With Type 2 operation, a data-link connection is established between two LLCs prior to any exchange of information-bearing
PDUs. The normal cycle of communication between two Type 2 LLCs
on a data-link connection consists of the transfer of PDUs containing
information from the source LLC to the destination LLC, acknowledged by PDUs in the opposite direction.
9.4.2 LLC PDU Structure
9.4.2.1 Overview. Figure 9.5 shows the LLC PDU format. As the figure
indicates, each LLC PDU contains two address fields: the destination service
access point DSAP. address field and the source service access point SSAP.
address field. Each of these fields contains only a single address. The DSAP
address field identifies one or more SAPs for which the LLC information
field is intended. The SSAP address field identifies the specific SAP from
which the LLC information field was initiated. Figures 9.6 and 9.7 show the
address field formats.
An individual address is usable as both an SSAP and a DSAP address; a
null address is usable as both an SSAP and DSAP address; and a group
address is usable only as a DSAP address.
All 1s in the DSAP address field i.e., the address type designation bit set
to 1 and the seven address bits set to 1. is predefined to be the global
DSAP address. The DSAP address designates a group consisting of all
DSAPs actively being serviced by the underlying MAC SAP address es..
All 0s in the DSAP or SSAP address field i.e., the address type designation bit set to 0 and seven address bits set to 0. is predefined to be the null
address. The null SAP address designates the LLC that is associated with the
underlying MAC SAP address and is not used to identify any SAP to the
network layer or any SAP to an associated layer management function.
Addresses 01000000 and 11000000 are designated as the individual and group
addresses, respectively, for an LLC sublayer management function at the
station. Other addresses with the next to low-order bit set to 1 are reserved
for International Standards Organization ISO. definition.
212
Figure 9.6. DSAP and SSAP address field formats.
Figure 9.7. Global DSAP address field format.
The control field consists of one or two octets that are used to designate
command and response functions and that contain sequence numbers when
required. The content of this field is described below.
The information field consists of any integral number including zero.
of octets.
The bit order is as follows. Addresses, commands and responses, and
sequence numbers are delivered torreceived from the MAC sublayer least
significant bit first i.e., the first bit of a sequence number that is
deliveredrreceived has the weight 2**0.. The information field is delivered
to the MAC sublayer in the same bit order as received from the network
layer. The information field is delivered to the network layer in the same bit
order as received from the MAC sublayer.
213
Definition of an Invalid LLC PDU. An invalid LLC PDU is defined as one

which meets at least one of the following conditions:
1. It is identified as such by the PHY or MAC sublayer.
2. It is not an integral number of octets in length.
3. It does not contain two properly formatted address fields, one control
field, and, optionally, an information field in their proper order.
4. Its length is less than 3 octets one-octet control field. or 4 octets
two-octet control field..
Invalid LLC PDUs are ignored.
9.4.2.2 Control Field Formats. The three formats defined for the control
field, shown in Figure 9.8, are used to perform numbered information
transfer, numbered supervisory transfer, unnumbered control, and unnumbered transfer functions. The numbered information transfer and supervisory
transfer functions apply only to Type 2 operation. The unnumbered control
and unnumbered information transfer functions apply to either Type 1 or
Type 2 operation but not both. depending on the specific function selected.
Information Transfer Format I. The I-format PDU is used to perform a
numbered information transfer in Type 2 operation. Except where otherwise
Figure 9.8. LLC PDU control field formats.
214
specified e.g., UI, TEST, FRMR, and XID., it is the only LLC PDU that
may contain an information field. The functions N S ., N R ., and pollrfinal
PrF. are independent wi.e., each I-format PDU has an N S . sequence
number, an N R . sequence number that acknowledges or does not acknowledge additional I-format PDUs at the receiving LLC, and a PrF bit that is
set to 1 or 0x. One notes the similarity to HDLC given in Chapter 3.
Supervisory Format S. The S-format PDU is used to perform data-link
supervisory control functions in Type 2 operation, such as acknowledging
I-format PDUs, requesting retransmission of I-format PDUs, and requesting
a temporary suspension of transmission of I-format PDUs. The functions of
N R . and PrF are independent i.e., each S-format PDU has an N R .
sequence number that does or does not acknowledge additional I-format
PDUs at the receiving LLC, and it also has a PrF bit that is set
to 1 or 0..
Unnumbered Format U. The U-format PDUs are used in either Type 1 or
Type 2 operation, depending upon the specific function utilized, to provide
additional data-link control functions and to provide unsequenced information transfer. The U-format PDUs contain no sequence numbers but include
a PrF bit that is set to 1 or 0.
9.4.2.3 Control Field Parameters
Type 1 Operation Parameters. The only parameter that exists in Type 1
operation is the PrF bit. The PrF bit set to 1 is only used in Type 1
operation with the XID and TEST commandrresponse PDU functions. The
poll P. bit set to 1 is used to solicit poll. a correspondent response PDU
with the F bit set to 1 from the addressed LLC. The final F. bit set to 1
is used to indicate that the response PDU which is sent by the LLC is the
result of a soliciting poll. command PDU P bit set to 1..
Type 2 Operation Parameters
MODULUS.
Each I PDU is sequentially numbered with a number that has a

value between 0 and MODULUS minus ONE where MODOLUS is the
modulus of the sequence numbers.. The modulus is equal to 128 for Type 2
LLC control field format. The sequence numbers cycle through the entire
range.
The maximum number of sequentially numbered I PDUs that may be
outstanding i.e., unacknowledged . in a given direction on a data-link connection at any given time never exceeds one less than the modulus of the
sequence numbers. This restriction prevents any ambiguity in the association
of sent I PDUs with sequence numbers during normal operation andror
error recovery action.
215
LLC PDU STATE VARIABLES AND SEQUENCE NUMBERS. A station LLC maintains
a send state variable V S . for the I PDUs it sends and a receive state variable
V R . for the I PDUs it receives on each data-link connection. The operation
of V S . is independent of the operation of V R ..
The send state variable denotes the sequence number of the next insequence I PDU to be sent on a specific data-link connection. The send state
variable takes on a value between 0 and MODULUS minus ONE where
MODULUS equals 128 and the numbers cycle through the entire range..
The value of the send state variable is incremented by one with each
successive I PDU transmission on the associated data-link connection, but
will not exceed N R . of the last received PDU by more than MODULUS
minus ONE. Again the similarity to HDLC should be noted.
Only I PDUs contain N S ., the send sequence number of the sent PDU.
Prior to sending an I PDU, the value of N S . is set equal to the value of the
send state variable for that data-link connection.
The receive state variable denotes the sequence number of the next
in-sequence I PDU to be received on a specific data-link connection. The
receive state variable takes on a value between 0 and MODULUS minus
ONE where MODULUS equals 128 and the numbers cycle through the
entire range.. The value of the receive state variable associated with a
specific data-link connection is incremented by one whenever an error-free,
in-sequence I PDU is received whose send sequence number N S . equals the
value of the receive state variable for the data-link connection.
All I-format PDUs and S-format PDUs contain N R ., the expected
sequence number of the next received I PDU on the specified data-link
connection. Prior to sending an I-format PDU or S-format PDU, the value of
N R . is set equal to the current value of the associated receive state variable
for that data-link connection. N R . indicates that the station sending the
N R . has received correctly all I PDUs numbered up through N R . y 1 on
the specified data-link connection.
POLLrFINAL PrF. BIT.
The poll P. bit is used to solicit poll. a response from

the addressed LLC. The final F. bit is used to indicate the response PDU
sent as the result of a soliciting poll. command.
The pollrfinal PrF. bit serves as a function in Type 2 operation in both
command PDUs and response PDUs. In command PDUs the PrF bit is
referred to as the P bit. In response PDUs it is referred to as the F bit. PrF
bit exchange provides a distinct commandrresponse linkage that is useful
during normal operation and recovery operations.
A command PDU with the P bit set to 1 is used on a data-link
connection to solicit a response PDU with the F bit set to 1 from the
addressed LLC on that data-link connection.
Only one PDU with a P bit set to 1 will be outstanding in a given
direction at a given time on the data-link connection between any specific
pair of LLCs. Before an LLC issues another PDU on the same data-link
216
connection with the P bit set to 1, the LLC will have received a response
PDU with the F bit set to 1 from the addressed LLC. If no valid response
PDU is received within a system-defined P-bit timer timeout period, the
re.sending of a command PDU with the P bit set to 1 is permitted for
error recovery purposes.
A response PDU with the F bit set to 1 is used to acknowledge receipt
of a command PDU with the P bit set to 1.
Following the receipt of a command PDU with the P bit set to 1, the
LLC sends a response PDU with the F bit set to 1 on the appropriate
data-link connection at the earliest possible opportunity. The LLC is permitted to send appropriate response PDUs with the F bit set to 0 at any
medium access opportunity on an asynchronous basis without the need for a
command PDU..
9.4.2.4 Commands and Responses. The commandrresponse CrR. bit,
located as the low-order bit in the SSAP field, is used to distinguish between
commands and responses. Table 9.1 lists appropriate commands and related
responses.
Type 1 Operation Commands and Responses. The Type 1 commands and
responses are all U-format PDUs. The U-format PDU command encodings
for Type 1 operation are listed in Figure 9.9.
The unnumbered information UI. command PDU is used to send information to one or more LLCs. Use of the UI command PDU is not dependent
on the existence of a data-link connection between the destination and the
source LLCs, and its use will not affect the V S . and V R . variables
associated with any data-link connections. There is no LLC response PDU to
the UI command PDU.
TABLE 9.1 LLC Commands and Related Responses

Information Transfer Format Commands
Information Transfer Format Responses
IInformation
IInformation
Supervisory Format Commands

RRreceive ready
RNRreceive not ready
REJreject
Supervisory Format Responses

RRreceive ready
RNRreceive not ready
REJreject
Unnumbered Format Commands

UIunnumbered information
DISCdisconnect
SABMEset asynchronous balanced
mode extended
XIDexchange identification
TESTtest
Unnumbered Format Responses

UAunnumbered acknowledgement
DMdisconnected mode
FRMRframe reject
XIDexchange identification
TESTtest
217
Figure 9.9. Type 1 operation command control field bit assignments.
Reception of the UI command PDU is not acknowledged nor is the

sequence number verified; therefore, the data contained in a UI PDU
may be lost if a logical data-link exception such as a transmission error
or a receiver-busy condition. occurs during the sending of the command
PDU.
A UI command PDU has either an individual, group, global, or null
address as the destination DSAP address and has the originators individual
address as the SSAP address.
The XID command PDU is used to a. convey the types of LLC services
supported for all LLC services. and the receive window size on a per-datalink connection basis to the destination LLC and b. cause the destination
LLC to respond with the XID response PDU at the earliest opportunity. The
XID command PDU has no effect on any mode or sequence numbers
maintained by the remote LLC. An XID command PDU has either an
individual, group, global, or null address as the destination DSAP address
and has the originators individual address as the SSAP address.
The information field of an XID basic format command PDU consists of
an 8-bit XID format identifier field plus a 16-bit parameter field that is
encoded to identify the LLC services supported plus the receive window size,
as shown in Figure 9.10. The receive window size k . is the maximum number
that the send state variable V S . can exceed the N R . of the last received
PDU.
TEST COMMAND.
The TEST command PDU is used to cause the destination

LLC to respond with the TEST response PDU at the earliest opportunity,
thus performing a basic test of the LLC to LLC transmission path. An
information field is optional with the TEST command PDU. If present,
however, the received information field is returned, if possible, by the
addressed LLC in the TEST response PDU. The TEST command PDU has
no effect on any mode or sequence numbers maintained by the remote LLC
and may be used with an individual, group, global, or null DSAP address and
with an individual, group, or global DA address.
218
Figure 9.10. XID information field basic format.
TYPE 1 OPERATION RESPONSES. The U-format PDU response encodings for

Type 1 operation are shown in Figure 9.11.
Type 2 Operation Commands and Responses. Type 2 commands and

responses consist of I-format, S-format, and U-format PDUs.
The function of the information, I, command and response is to transfer
sequentially numbered PDUs containing an octet-oriented information field
across a data-link connection. The encoding of the I PDU control field for
Type 2 operation is shown in Figure 9.12.
The I PDU control field contains two sequence numbers: N S ., send
sequence number, which indicates the sequence number associated with the I
PDU; and N R ., receive sequence number, which indicates the sequence
Figure 9.11. Type 1 operation response control field bit assignments.
219
Figure 9.12. Information transfer format control field bits, I PDU, Type 2 operation.
number as of the time the PDU is sent. of the next expected I PDU to be
received and consequently indicates that the I PDUs numbered up through
N R . y 1 have been received correctly. Supervisory, S, PDUs are used to
perform numbered supervisory functions such as acknowledgments, temporary suspension of information transfer, or error recovery.
PDUs with the S format do not contain an information field and, therefore, do not increment the send state variable at the sender or the receive
state variable at the receiver. The encoding of the S-format PDU control
field for Type 2 operation is shown in Figure 9.13.
An S-format PDU contains an N R ., receive sequence number, which
indicates, at the time of sending, the sequence number of the next expected I
PDU to be received and consequently indicates that all received I PDUs
numbered up through N R . y 1 have been received correctly.
When sent, an RR receive ready. or REJ PDU indicates the clearance of
any busy condition at the sending LLC that was indicated by the earlier
sending of an RNR receive not ready. PDU.
The RR PDU is used by an LLC to indicate that it is ready to receive an I
PDUs.. I PDUs numbered up through N R . y 1 are considered to be
acknowledged.
The reject REJ. PDU is used by an LLC to request resending of I PDUs
starting with the PDU numbered N R .. I PDUs numbered up through
N R . y 1 are considered as acknowledged. It is possible to send additional I
PDUs awaiting initial sending after the re-sent I PDUs.
With respect to each direction of sending on a data-link connection, only
one sent REJ condition is established at any given time. The sent REJ
220
Figure 9.13. Supervisory format control field bits.
condition will be cleared upon receipt of an I PDU with an N S . equal to

N R . of the REJ PDU.
The receive not ready RNR. PDU is used by an LLC to indicate a busy
condition i.e., a temporary inability to accept subsequent I PDUs.. I PDUs
numbered up through N R . y 1 are considered as acknowledged. I PDUs
numbered N R . and any subsequent I PDUs received, if any, are not
considered as acknowledged; the acceptance status of these PDUs is indicated in subsequent exchanges.
UNNUMBERED FORMAT COMMANDS AND RESPONSES. Unnumbered, U, commands and responses are used in Type 2 operation to extend the number of
data-link control functions. PDUs sent with the U format do not increment
the state variables on the data-link connection at either sending or receiving
LLC. The encoding of the U-format commandrresponse PDU control field
is given in Figures 9.14 and 9.15.
Set Asynchronous Balanced Mode Extended (SABME) Command. The

SABME command is used to establish a data-link connection to the destination LLC in the asynchronous balanced mode. No information is permitted
with the SABME command PDU. The destination LLC confirms receipt of
the SABME command PDU by sending a UA unnumbered acknowledgement. response PDU on the data-link connection at the earliest opportunity
according to whether a DL-CONNECT primitive. response or a DL-
221
Figure 9.14. Unnumbered format control field bits.
Figure 9.15. Unnumbered command and response control field bit assignments.
DISCONNECT request primitive is passed from the network layer to the

LLC sublayer. Upon acceptance of the SABME command PDU, the destination LLCs send and receive state variables which are set to zero. If the UA
response PDU is received correctly, then the initiating LLC also assumes the
asynchronous balanced mode with its corresponding send and receive state
variables set to zero.
Previously sent I PDUs that are unacknowledged when this command is
actioned remain unacknowledged. Whether or not an LLC resends the
contents of the information field of unacknowledged outstanding I PDUs is
decided at a higher OSI. layer.
Disconnect (DISC) Command. The DISC command PDU is used to terminate an asynchronous balanced mode previously set by a SABME command
PDU. It is used to inform the destination LLC that the source LLC is
suspending operation of the data-link connection and that the destination
LLC should assume the logically disconnected mode. No information field is
222
Figure 9.16. FRMR information field format.
permitted with the DISC command PDU. Prior to actioning the command,
the destination LLC confirms the acceptance of the DISC command PDU by
sending a UA response PDU on that data-link connection.
Previously sent I PDUs that are unacknowledged when this command is
actioned remain unacknowledged. Whether or not the LLC resends the
contents of the information field of unacknowledged outstanding I PDUs is
decided at a higher OSI. layer.
The frame reject response FRMR. information field is shown in
Figure 9.16.
Section 9.4 consists of abstracted and edited material from Ref. 2.
MEDIUM ACCESS CONTROL
223
9.5 MEDIUM ACCESS CONTROL

9.5.1 Introduction
With baseband LAN transmission, one medium interconnects all participating stations. In such a scheme, only one LAN station can transmit at a time.
If two stations transmit at the same time, they will interfere with each other,
each corrupting the others traffic. There are three approaches to control
medium access: random access or contention first come, first served., token
passing, and time-slot assignment schemes. In this chapter and Chapter 10
we deal with the first two. A typical time-slot assignment scheme is distributed queue dual bus, which is described in Ref. 4.
9.5.2 Carrier Sense Multiple Access with Collision Detection
(CSMAr
r CD) Basic Concepts
9.5.2.1 Introduction to CSMAr
r CD. Carrier sense multiple access with
collision detection CSMArCD. Ref. 3. is a random access method for
LANs. Conceptually, this is a baseband technique for LAN transmission. As
with other baseband LANs, two or more LAN stations share a common
transmission medium. To transmit, a station waits defers. until a quiet
period on the medium i.e., no other station is transmitting . and then sends
the message in bit serial form. If, after initiating a transmission, the message
collides with that of another station, then each transmitting station intentionally sends a few additional bytes to ensure propagation of the collision
throughout the system. The station remains silent for a random amount of
time called backoff . before attempting to transmit again.
CSMArCD relationship to OSI is shown in Figure 9.17. This section deals
with the lowest two OSI layers: the data-link layer and the physical layer.
These encompass the MAC sublayer and the physical layer. The OSI data-link
layer in this case consists of the MAC and the LLC Section 9.4..
r CD. Figure 9.18 illustrates the
9.5.2.2 MAC Frame Structure for CSMAr
MAC frame format. There are nine fields: preamble, start frame delimiter
SFD., destination and source addresses, lengthrtype, MAC Client Data,
PAD when required., frame check sequence, and extension. The extension
field is required for 1-Gbps operation in the half-duplex mode only. Of these
nine fields, all are of fixed length except the client data, PAD, and extension
fields. The nine fields may contain an integer number of octets between the
minimum and maximum values determined by the specific implementation of
the CSMArCD MAC see Section 9.6.5..
In the tables in Section 9.6.5, the minimum and maximum frame size limits
refer to that portion of the frame from the destination address field through
the frame check sequence inclusive. When dealing with Figure 9.18, the
octets of a frame are transmitted from top to bottom and the bits of each
octet are transmitted from left to right.
224
225
Figure 9.18. MAC frame format.
Elements of a MAC Frame

PREAMBLE FIELD. A preamble field contains seven octets and is used to allow
the PLS physical layer signaling. circuitry to reach its steady-state synchronization with received frame timing. The preamble pattern is alternating
binary 1s and 0s i.e., 10101010.. The nature of the pattern is such that, for
Manchester coding, it appears as a periodic waveform on the medium that
enables bit synchronization. It should be noted that the pattern ends with a
binary 0.
Start Frame Delimiter (SFD) Field. The SFD field is the sequence 10101011.
It immediately follows the preamble and indicates the start of frame.
ADDRESS FIELDS.
Each MAC frame contains two address fields: destination

address field and the source address field, in that order see Figure 9.19..
The destination address field specifies the destination addressee s. for which
the frame is intended. The source address field identifies the station from
which the frame was initiated.
a. Each address field is 48 bits long.
b. The first bit least significant bit, LSB. is used in the destination
address field as an address-type designation bit to identify the destina-
226
Figure 9.19. Address field format.
tion address either as an individual or as a group address. If this bit is

0, it indicates that the address field contains an individual address. If
the bit is 1, it indicates that the address field contains a group address
that identifies none, one or more, or all of the stations connected to
the LAN. In the source address field, the first bit is reserved and
set to 0.
c. The second bit is used to distinguish between locally or globally
administered addresses. For globally administered or U, universal.
addresses, the bit is set to 0. If an address is to be assigned locally, this
bit is set to 1. Note that for the broadcast address, this bit is also set
to 0.
d. Each octet of each address field is transmitted least significant bit first.
If there are all 1s in the destination address field, it is predefined to be
the broadcast address.
LENGTHrTYPE FIELD. This is a two-octet field that takes on one of two
meanings depending on its numeric value. For numerical evaluation, the first
octet is the most significant octet in this field.
a. If the value of this field is less than or equal to the value of

maxValidFrame, then the lengthrtype field indicates the number of
MAC client data octets contained in the subsequent data field of the
frame length interpretation ..
MaxValidFrame s maxFrameSize
y 2XaddressSize q LengthOrTypeSize q crcSize . r8
Substitute the following in the above equation: MaxFrameSize in bits.
uses data from tables in Section 9.6.5 e.g., for 10 Mbps, 1518 octets
=8..
2XaddressSize s 32 bits
LengthOrTypeSize s 16 bits
CrcSize s 32 bits
227
b. If the value of this field is greater than or equal to 1536 decimal equal
to 0600 hexadecimal ., then the lengthrtype fields indicates the nature
of the MAC client protocol Type interpretation ..* The length and
type interpretations of this field are mutually exclusive.
Regardless of the interpretation of the lengthrtype field, if the length of
the data field is less than the minimum required for proper operation of the
protocol, a PAD field a sequence of octets. is added at the end of the data
field but prior to the FCS field, as discussed below. The procedure that
determines the size of the PAD field is specified in Section 4.2.8 of Ref. 3.
The lengthrtype field is transmitted and received with the high-order octet
first.
DATA AND PAD FIELDS.
The data field MAC client data. contains a sequence

of n octets. Full data transparency is provided in the sense that any arbitrary
sequence of octet values may appear in the data field up to the maximum
number specified by the implementation of the standard that is used. Keep in
mind that there is a minimum frame size required i.e., 64 octets. for proper
operation of the protocol.
The length of the PAD field required for MAC client data that is n
octets long is max w0, minFrameSize y 8x n q 2XaddressSize q 48.x bits.
The maximum possible size of the data field is maxFrameSize y
2XaddressSize q 48.r8 octets.
FCS. FIELD. A CRC is used by the transmit and
receive algorithms to generate a CRC value for the FCS field. The FCS field
contains a 32-bit CRC value. The value is computed as a function of the
contents of the source address, destination address, length, MAC client data,
and PAD that is, all fields except the preamble, SFD, FCS, and extension..
The encoding is defined by the following generating polynomial:
FRAME CHECK SEQUENCE
G x . s x 32 q x 26 q x 23 q x 22 q x 16 q x 12 q x 11 q x 10
q x 8 q x7 q x 5 q x 4 q x 2 q x q 1
Mathematically, the CRC value corresponding to a given frame is defined by
the following procedure:
a. The first 32 bits of the frame are complemented.
b. The n bits of the frame are then considered to be the coefficients of a
polynomial M x . of degree n y 1. The first bit of the destination
address field corresponds to the x ny1. term, and the last bit of the
data field corresponds to the x 0 term..
*Type field assignments are administered by the Registration Authority, IEEE Standards Dept.,
PO Box 1331, Piscataway, NJ 08855-1331.
228
c. M x . is multiplied by x 32 and divided by G x ., producing a remainder

R x . of degree F 31.
d. The coefficients of R x . are considered to be a 32-bit sequence.
e. The bit sequence is complemented and the result is the CRC.
The 32 bits of the CRC value are placed in the frame check sequence field so
that the x 31 term is the leftmost bit of the first octet, and the x 0 term is the
rightmost bit of the last octet. The bits of the CRC are thus transmitted in
the order x 31 , x 30 , . . . , x 1, x 0 ..
EXTENSION FIELD.
The extension field follows the FCS field and is made up

of a sequence of extension bits, which are readily distinguished from data
bits. The length of the field is in the range of zero to slotTime-minFrameSize. bits, inclusive. The contents of the extension field are not included in
the FCS computation.
Carrier extension is employed in the half-duplex mode only and at operating speeds above 100 Mbps. The slotTime employed at speeds of 100 Mbps
and below is inadequate to accommodate network topologies of the desired
physical extent. Carrier extension provides a means by which the slotTime
can be increased to a sufficient value for the desired topologies, without
increasing the minFrameSize parameter, because this would have deleterious
effects. Nondata bits, referred to as extension bits, are appended to frames
that are less than slotTime bits in length so that the resulting transmission is
at least one slotTime in duration. Carrier extension can be performed only if
the underlying physical layer is capable of sending and receiving symbols that
are readily distinguished from data symbols, as is the case in most physical
layers that use a block encodingrdecoding scheme. The maximum length of
the extension is equal to the quantity slotTime y minFrameSize ..
The MAC continues to monitor for collisions while it is transmitting
extension bits, and it will treat any collision that occurs after the threshold
slotTime. as a late collision.
Order of Bit Transmission. Each octet of the MAC frame, with the exception of the FCS, is transmitted low-order bit first.
Inalid MAC Frame. An invalid MAC frame is defined as one that meets
at least one of the following conditions:
1. The frame length is inconsistent with the length field.
2. It is not an integral of octets in length.
3. The bits of the incoming frame exclusive of the FCS field itself. do not
generate a CRC value identical to the one received. The contents of
invalid MAC frames are not passed to the LLC.
It should be noted that the LLC PDU is inserted into the MAC client data
field as shown in Figure 9.18.
229
9.5.2.3 CSMAr
r CD Basic Operation
Transmission Without Contention. When an LLC sublayer requests the
transmission of a frame, the transmit data encapsulation component of the
MAC sublayer constructs the frame from the LLC-supplied data. It appends
a preamble and an SFD to the beginning of a frame. A PAD is also
appended, if required. It also appends destination and source addresses, a
length count field, and an FCS. The frame is then handed to the transmit
media access management component in the MAC sublayer for transmission.
The transmit media access management then attempts to avoid contention
with other traffic on the medium by monitoring the carrier sense signal
provided by the PLS and deferring to passing traffic. When the medium is
clear, frame transmission starts. The MAC sublayer then provides a serial
stream of bits to the PLS interface for transmission.
The PLS performs the task of actually generating the electrical signals on
the medium that represent the bits of the frame. Simultaneously, it monitors
the medium and generates the collision detect signal, which, in the contention-free case, remains off for the duration of the frame.
Reception Without Contention. At each LAN receiving station, the arrival of
a frame is first detected by the PLS, which responds by synchronizing with
the incoming preamble and by turning on the carrier sense signal. As the
encoded bits arrive from the medium, they are decoded and translated back
into binary data. The PLS passes subsequent bits up to the MAC sublayer,
where the leading bits are discarded, up to and including the end of the
preamble and SFD.
Meanwhile, the receive media access management component of the MAC
sublayer, having observed carrier sense, has been waiting for the incoming
bits to be delivered. Receive media access management collects bits from the
PLS as long as the carrier sense signal remains on. When the carrier sense
signal is removed, the frame is truncated to the octet boundary, if necessary,
and passed to receive data decapsulation for processing.
Receive data decapsulation checks the frames destination address field to
decide whether the frame should be received by this station. If so, it passes
the destination address DA., the source address SA., and the LLC data
unit LLCDU. to the LLC sublayer along with the appropriate status code
indicating reception complete or reception too long. It also checks for
invalid MAC frames by inspecting the FCS to detect any damage to the
frame en route and by checking for proper octet boundary alignment of the
frame. Frames with a valid FCS may also be checked for proper octet
boundary alignment.
Access Interference and Recovery. If multiple LAN stations attempt to
transmit at the same time, it is possible for them to interfere with each
others transmissions, in spite of their attempts to avoid this by deferring.
230
When transmissions from two stations overlap, the resulting contention is

called collision. A given station can experience collision during the initial part
of its transmission the collision window. before its transmitted signal has had
time to propagate to all stations on the medium. Once the collision window
has passed, a transmitting station is said to have acquired the medium.
Subsequent collisions are avoided because all other properly functioning.
LAN stations can be assumed to have noticed the signal by means of carrier
sense. and to be deferring to it. The time to acquire the medium is thus
based on the round-trip propagation time of the physical layer whose
elements include the PLS, the physical medium attachment PMA., and the
physical medium itself.
In the event of a collision, the transmitting stations physical layer initially
notices the interference on the medium and then turns on the collision detect
signal. This is noticed in turn by the transmit media access management
component of the MAC sublayer, and the collision handling begins. First, the
transmit media access management enforces the collision by transmitting a
bit sequence called a jam. This ensures that the duration of the collision is
sufficient to be noticed by the other transmitting stations. involved in the
collision. After the jam is sent, the transmit media access management
terminates the transmission and schedules another transmission attempt after
a randomly selected time interval. Retransmission is attempted again in the
face of repeated collisions. Because repeated collisions indicate a busy
medium, however, the transmit media access management attempts to adjust
to the medium load by backing off voluntarily delaying its own retransmissions to reduce its load on the medium.. This is accomplished by expanding
the interval from which the random retransmission time is selected on each
successive transmit attempt. Eventually, either the transmission succeeds, or
the attempt is abandoned on the assumption that the medium has failed or
has become overloaded.
At the receiving end, the bits resulting from the collision are received and
decoded by the PLS just as are the bits of a valid frame. Fragmentary frames
received during collisions are distinguished from valid transmissions by the
MAC sublayers receive media access management component.
Carrier Deference. Even when it has nothing to transmit, the CSMArCD
MAC sublayer monitors the physical medium for traffic by watching the
carrier sense signal provided by the PLS. Whenever the medium is busy, the
CSMArCD MAC sublayer defers to the passing frame by delaying any
pending transmissions of its own. After the last bit of a passing frame i.e.,
when carrier sense changes from true to false., the CSMArCD MAC
sublayer continues to defer for a specified interframe spacing period.
If, at the end of the interframe spacing, a frame is waiting to be transmitted, transmission is initiated independently of the value of carrier sense.
When transmission has completed or immediately, if there was nothing to
transmit., the CSMArCD MAC sublayer resumes its original monitoring of
carrier sense.
231
When a frame is submitted by the LLC sublayer for transmission, the

transmission is initiated as soon as possible, but in conformance with the
rules of deference stated above.
Collision Handling. Once a CSMArCD sublayer has finished deferring and
has started transmission, it is still possible for it to experience contention
for the medium. Collisions can occur until acquisition of the network has
been accomplished through the deference of all other stations CSMArCD
sublayers.
The dynamics of collision handling are largely determined by a single
parameter called the slot time. This single parameter describes three important aspects of collision handling:
1. It is an upper bound on the acquisition time of the medium.
2. It is an upper bound on the length of a frame fragment generated by a
collision.
3. It is the scheduling quantum for retransmission.
To fulfill all three functions, the slot time must be larger than the sum of the
physical layer round-trip propagation time and the media access layer maximum jam time. The slot time is determined by the parameters of the
implementation.
Collision Detection and Enforcement. Collisions are detected by monitoring
the collision detect signal provided by the physical layer. When a collision is
detected during a frame transmission, the transmission is not terminated
immediately. Instead, the transmission continues until additional bits specified by jam size have been transmitted counting from the time collision
detect went on.. This collision enforcement or jam guarantees that the
duration of the collision is sufficient to ensure its detection by all transmitting stations on the network. The content of the jam is unspecified; it may be
any fixed or variable pattern convenient to the media access implementation,
but the implementation will not be intentionally designed to the 32-bit CRC
value corresponding to the partial . frame transmitted prior to the jam.
Collision Backoff and Retransmission. When a transmission attempt has
terminated due to a collision, it is retried by the transmitting CSMArCD
sublayer until it is successful or a maximum number of attempts attempt
limit. have been made and all have terminated due to collisions. The
scheduling of the retransmissions is determined by a controlled randomization process called truncated binary exponential backoff. At the end of enforcing a collision jamming., the CSMArCD sublayer delays before attempting
to retransmit the frame. The delay is an integer multiple of the slotTime. The
number of slot times to delay before the nth retransmission attempt is chosen
as a uniformly distributed random integer r in the range
0 F r F 2k,
where k s min n, 10 .
232
If all limit attempts fail, this event is reported as an error. Algorithms used to
generate the integer r are designed to minimize the correlation between the
numbers generated by any two stations at any given time.
It is noted that the values given above define the most aggressive behavior
that a station may exhibit in attempting to retransmit after a collision. In the
course of implementing the retransmission scheduling procedure, a station
may introduce extra delays that will degrade its own throughput, but in no
case may a stations retransmission scheduling result in a lower average delay
between retransmission attempts than the procedure defined above.
9.6 CSMAr
r CDCURRENT STATUS AND ADVANCED OPERATION
9.6.1 General
The original CSMArCD system operated at a nominal 10-Mbps data rate.
Subsequently, the system has undergone modifications for 100-Mbps and
1000-Mbps capabilities. Numerous variants on transmission media have also
evolved. Table 9.2 lists 22 varieties of CSMArCD, variously called Ethernet.
9.6.2 Half-Duplex and Full-Duplex
Half-duplex operation is where two or more stations share a common
transmission medium. Full-duplex operation, on the other hand, allows
simultaneous communication between a pair of stations using point-to-point
means dedicated channel .. Full-duplex operation does not require that
transmitters defer, nor do they monitor or react to receive activity, as there is
no contention for a shared medium in this mode. Full-duplex mode can only
be used when all of the following are true:
a. The physical medium is capable of supporting simultaneous transmission and reception without interference.
b. There are exactly two stations connected with a full-duplex point-topoint link. Since there is no contention for use of a shared medium,
the multiple access i.e., CSMArCD. algorithms are unnecessary.
c. Both stations on the LAN are capable of, and have been configured to
use, full-duplex operation.
The most common configuration envisioned for full-duplex operation consists
of a central bridge also known as a switch . with a dedicated LAN connecting
each bridge port to a single device. Repeaters as defined in the reference
standard are outside the scope of full-duplex operation.
9.6.3 Compatibility InterfacesPhysical Layer
In this section five important compatibility interfaces are defined in what is
architecturally the physical layer. These are shown in Figure 9.20.
CSMArCD CURRENT STATUS AND ADVANCED OPERATION
233
rCD
TABLE 9.2 Variants of CSMAr
Type Name
Bit Rate
Transmission Medium
Comments
10BASE2
10 Mbps
RG 58 coaxia l cable
10BASE5
10 Mbps
Coaxial cable
10BASE-F
10 Mbps
Fiber-optic cable
10BASE-FB
10 Mbps
Fiber-optic cable
Has MAU port on one repeater

that can connect to another
repeater
10BASE-FB
segment
10 Mbps
Fiber-optic cable
Fiber-optic link segment provides

connectivity between two
10BASE-FB ports on repeaters
10BASE-FL
segment
10 Mbps
Fiber-optic cable
Fiber-optic link segment provides

connectivity between two
10BASE-FL MAUs
10BASE-FP
segment
10 Mbps
Fiber-optic cable
Fiber-optic mixing segment,

including one 10BASE-FP star
and all attached fiber pairs
10BASE-FP
star
10 Mbps
Fiber-optic cablerstar
Passive device couples fiber pairs

together to form 10BASE-FP
segment
10BASE-T
10 Mbps
Two wire pairs
Ordinary telephone pair cable
100BASE-FX
100 Mbps
Two optical fibers
100BASE-T
10 & 100 Mbps
Shielded wire pair;

twisted pair
Limited length shielded cable,

twisted pair
100BASE-T2
100 Mbps
Two twisted pairs
Category 3 or better balanced

cabling
Thicknet
100BASE-T4
100 Mbps
Four twisted pairs
Category 3, 4, or 5 UTP
100BASE-TX
100 Mbps
Two twisted pairs
Category 5 UTP or STP
100BASE-X
100 Mbps
General reference to 100BASE-TX

and 100BASE-FX
1000BASE-CX 1000 Mbps
Specialty shielded
copper cable
Specialty shielded copper balanced

jumper cable assembles
1000BASE-LX 1000 Mbps
Fiber-optic cable
Long wavelength, multimode or

single mode
1000BASE-SX 1000 Mbps
Fiber-optic cable
Short wavelength over multimode

fiber
1000BASE-T
1000 Mbps
Four wire pairs
Balanced category 5
1000BA SE-X
1000 Mbps
General category
Embodies three types of

transmission media: 1000BASE-CX,
1000BASE-SX, and 1000BASE-LX
10BROAD36
10 Mbps
Single broadband
cable
1BASE5
1 Mbps
Two twisted pairs
Telephone wire pair
234
235
a. Medium-Dependent Interfaces (MDI). To communicate in a compatible

manner, all stations shall adhere rigidly to the exact specification of
physical media signals defined in clause 8 and beyond. in IEEE Std.
802.3 1998 Edition, Ref. 3., and to the procedures that define correct
behavior of a station. The medium-independent aspects of the LLC
sublayer and the MAC sublayer should not be taken as detracting from
this point; communication by way of the ISOrIEC 8802-3 wANSIrIEEE
Std 802.3x local area network requires complete compatibility at the
physical medium interface that is, the physical cable interface ..
b. Attachment Unit Interface (AUI). It is anticipated that most DTEs will
be located some distance from their connection to the physical cable.
A small amount of circuitry will exist in the medium attachment unit
MAU. directly adjacent to the physical cable, while the majority of
the hardware and all of the software will be placed within the DTE.
The AUI is defined as a second compatibility interface. While conformance with this interface is not strictly necessary to ensure communication, it is highly recommended, since it allows maximum flexibility in
intermixing MAUs and DTEs. The AUI may be optional or not
specified for some implementations of this standard that are expected
to be connected directly to the medium and so do not use a separate
MAU or its interconnecting AUI cable. The PLS and PMA are then
part of a single unit, and no explicit AUI implementation is required.
c. Media Independent Interface (MII). It is anticipated that some DTEs
will be connected to a remote PHY, andror to different medium-dependent PHYs. The MII is defined as a third compatibility interface.
While conformance with implementation of this interface is not strictly
necessary to ensure communication, it is highly recommended, since it
allows maximum flexibility in intermixing PHYs and DTs. The MII is
optional.
d. Gigabit Media Independent Interface (GMII). The GMII is designed to
connect a gigabit-capable MAC or repeater unit to a gigabit PHY.
While conformance with implementation of this interface is not strictly
necessary to ensure communication, it is highly recommended, since it
allows maximum flexibility in intermixing PHYs and DTEs at gigabit
speeds. The GMII is intended for use as a chip-to-chip interface. No
mechanical connector is specified for use with the GMII. The GMII is
optional.
e. Ten-Bit Interface (TBI). The TBI is provided by the 1000BASE-X
PMA sublayer as a physical instantiation of the PMA service interface.
The TBI is highly recommended for 1000BASE-X systems, since it
provides a convenient partition between the high-frequency circuitry
associated with the PMA sublayer and the logic functions associated
with the PCS and MAC sublayers. The TBI is intended for use as a
chip-to-chip interface. No mechanical connector is specified for use
with the TBI. The TBI is optional.
236
Figure 9.21. CSMA r CD MAC functions.
9.6.4 Layer Interface Requirements

The interface requirements for CSMArCD are as follows:
a. The interface between the MAC sublayer and its client includes
facilities for transmitting and receiving frames, and it provides per-operation status information for use by higher-layer error recovery procedures.
b. The interface between the MAC sublayer and the physical layer
includes signals for framing carrier sense, receive data valid, transmit
initiation. and contention resolution collision detect ., facilities for
passing a pair of serial bit streams transmit, receive. between the two
layers, and a wait function for timing.
9.6.4.1 Functional Capabilities of the MAC Sublayer. Figure 9.21 outlines the CSMArCD media access control functions. This is followed by the
237
function listing below from Ref. 3.:

a. For Frame Transmission
1. Accepts data from the MAC client and constructs a frame
2. Presents a bit-serial data stream to the physical layer for transmission on the medium
Note: Assumes data passed from the client sublayer are octet
multiples.
b. For Frame Reception
1. Receives a bit-serial data stream from the physical layer
2. Presents to the MAC client sublayer frames that are either
broadcast frames or directly addressed to the local station
3. Discards or passes to network management all frames not addressed to the receiving station
c. In half-duplex mode, defers transmission of a bit-serial stream whenever the physical medium is busy
d. Appends proper FCS value to outgoing frames and verifies full octet
boundary alignment
e. Checks incoming frames for transmission errors by way of FCS and
verifies octet boundary alignment
f. Delays transmission of frame bit stream for specified interframe gap
period
g. In half-duplex mode, halts transmission when collision is detected
h. In half-duplex mode, schedules retransmission after a collision until a
specified retry limit is reached
i. In half-duplex mode, enforces collision to ensure propagation
throughout network by sending jam message
j. Discards received transmissions that are less than a minimum length
k. Appends preamble, start frame delimiter, DA, SA, lengthrtype field,
and FCS to all frames, and inserts PAD field for frames whose data
length is less than a minimum value
l. Removes preamble, start frame delimiter, DA, SA, lengthrtype field,
FCS, and PAD field if necessary . from received frames
m. Appends extension bits to the first or only. frame of a burst if it is
less than slotTime bits in length when in half-duplex mode at speeds
above 100 Mbps
n. Strips extension bits from received frames when in half-duplex mode
at speeds above 100 Mbps
238
9.6.5 Allowable ImplementationsParameterized Values

Table 9.3 identifies the parameter values used with the 10-Mbps implementation of CSMArCD. Table 9.4 gives the parameter values of 1BASE5.
Table 9.5 provides the parameter values for 100-Mbps implementations, and
Table 9.6 gives the parameter values for 1000-Mbps implementations.
TABLE 9.3 Parameters for 10-Mbps Implementations
Parameters
Values
slotTime
interFrameGap
attemptLimit
backoffLimit
jamSize
maxFrameSize
minFrameSize
burstLimit
512 bit times

9.6 s
16
10
32 bits
1518 octets
512 bits (64 octets)
Not applicable
Note: The spacing between two successive noncolliding packets,

from start to idle at the end of the first packet to start of preamble
of the subsequent packet, can have a minimum value of 47 bit
times, at the AUI receive line of the DTE. This interFrameGap
shrinkage is caused by variable network delays, added preamble
bits, and clock skew.
Source: Section 4.4.2 of Ref. 3.
TABLE 9.4 Parameters for 1BASE5 Implementations

Parameters
Values
slotTime
interFrameGap
attemptLimit
backoffLimit
jamSize
maxFrameSize
minFrameSize
burstLimit
512 bits times

96 s
16
10
32 bits
1518 octets
Not applicable

Parameters
Values
slotTime
interFrameGap
attemptLimit
backoffLimit
jamSize
maxFrameSize
minFrameSize
burstLimit
512 bit times

0.96 s
16
10
32 bits
1518 octets
Not applicable
239

Parameters
Values
slotTime
interFrameGap
attemptLimit
backoffLimit
jamSize
maxFrameSize
minFrameSize
burstLimit
4096 bit times

0.096 s
16
10
32 bits
1518 octets
65, 536 bits
Note: The spacing between two noncolliding packets, from the last
bit of the FCS field of the first packet to the first bit of the preamble of
the second packet, can have a minimum value of 64 BT (bit times),
as measured at the GMII receive signals at the DTE. This interFrameGap shrinkage may be caused by variable network delays,
added preamble bits, and clock tolerances.
9.6.5.1 Definition of Slot Time

The dynamics of collision handling are largely determined by a single
parameter called the slot time. This single parameter describes three important aspects of collision handling:
1. It is an upper bound on the acquisition time of the medium.
2. It is an upper bound on the length of a frame fragment generated by a
collision.
3. It is the scheduling quantum for retransmission.
To fulfill all three functions, the slot time must be larger than the sum of the
physical layer round-trip propagation time and the media access layer maximum jam time. The slot time is determined by the parameters of a specific
implementation.
r CD System Designation. There are a number of different
9.6.5.2 CSMAr
implementations of CSMArCD. These are designated by a nomenclature
made up of three parts: A number indicating the bit rate in Mbps, followed
by either the words BASE or BROAD, standing for baseband BASE. or
broadband BROAD.. The third element is either a number or letters.. If a
number is used, it describes a segment length in hundreds of meters. For
example, in 10BASE5, the 10 indicates that it is a 10-Mbps CSMArCD
system, BASE tells us that it is a baseband system versus broadband ., and
the 5 indicates that a segment length is 500 m. With some other implementations, the third element may be a letter or a letter-number. Consider the
240
following three examples:

100BASE-T4. This is a 100-Mbps baseband CSMArCD system using 4 pairs
of Category 3, 4, or 5 UTP unshielded twisted pair..
1000BASE-LX. This is 1000-Mbps baseband CSMArCD system using longwavelength laser devices over fiber-optic cable.
1BASE5. This is a 1-Mbps CSMArCD system with 500-m segments using
two pairs of twisted-pair telephone wire.
9.6.6 Physical Signaling (PLS) and Attachment Unit Interface
(AUI) Selected Discussion
Each direction of data transfer is serviced with two making a total of four.
balanced circuits: Data and Control. The Data and Control circuits are
independently self-clocked, thereby eliminating the need for separate timing
circuits. This is accomplished by encoding of all signals. The Control circuit
signaling rate is nominally but not of necessity exactly. equal to the Data
circuit signaling rate.
The data circuits are used only for data transfer. No control signals
associated with the interface are passed on these circuits. Likewise, the
Control circuits are used only for control message transfer. No data signals
associated with the interface are passed on these circuits.
Using such an arrangement see Figure 9.20. allows the DTE full media
independence for baseband coax, baseband twisted pair, broadband coax,
and baseband fiber media so that identical PLS, MAC, and MAC clients may
be used with any of these media.
9.6.6.1 Frame Structure Across the AUI. Frames transmitted on the AUI
have the following structure:
silence:preamble:sfd:data:edt:silence:
These frame elements in sequence have the following characteristics:
Element
Characteristics
silence:
preamble:
s no transitions
s alternating CD1. and CD0. 56 bit times
ending in CD0.
s CD1.CD0.CD1.CD0.CD1.CD0.
CD1.CD1.
s 8XN instances of CD0 or CD1
s IDL
sfd:
data:
etd:
241
These elements are defined as follows:

Silence. The silence : delimiter provides an observation window for an
unspecified period of time during which no transitions occur on the
AUI. The minimum length of this period is specified by the access
procedure.
Preamble. The preamble: delimiter begins a frame transmission and
provides a signal for receiver synchronization. The signal is an alternating pattern of CD1. and CD0.. This pattern is transmitted on the
Data Out circuit by the DTE to the MAU for a minimum of 56 bit
times* at the beginning of each frame. The last bit of the preamble
wthat is, the final bit of preamble before the start of frame delimiter
sfd.x is a CD0. The DTE is required to supply at least 56 bits of
preamble to satisfy system requirements. System components consume
preamble bits in order to perform their functions. The number of
preamble bits sourced ensures an adequate number of bits are provided
to each system component to correctly implement its function.
Start of Frame Delimiter sfd .. The sfd: indicates the start of a frame,
and follows the preamble. The sfd: element is
CD1. CD0. CD1. CD0. CD1. CD0. CD1. CD1.

Data. The data: in a transmission is in multiples of eight encoded data
bits CD0s and CD1s..
End of Transmission Delimiter etd .. The etd: delimiter indicates the end
of a transmission and serves to turn off the transmitter. The signal
starts the IDL.
9.6.6.2 Signal Characteristics
Signal Encoding. Two different signal encoding mechanisms may be used by
the AUI attachment unit interface .. One of the mechanisms is used to
encode data and the other to encode control.
Data Encoding. Manchester encoding is used for the transmission of data
across the AUI. Manchester encoding is a binary signaling mechanism that
combines data and clock into bit-symbols. Each bit-symbol is split into two
halves, with the second half containing the binary inverse of the first half; a
transition always occurs in the middle of each bit-symbol. During the first
half of the bit-symbol, the encoded signal is the logical complement of the bit
value being encoded. During the second half of the bit-symbol, the encoded
*A bit time is the duration of one bit. It is equal to the inverse of the bit rate.
242
Figure 9.22. Examples of Manchester coding.
signal is the uncomplemented value of the bit being encoded. Thus, a CD0 is
encoded as a bit-symbol in which the first half is HI and the second half is
LO. A CD1 is encoded as a bit-symbol in which the first half is LO and the
second half is HI. Examples of Manchester coding are shown in Figure 9.22.
The line condition IDL is also used as an encoded signal. An IDL always
starts with a HI signal level. Since IDL always starts with a HI signal, an
additional transition will be added to the data stream if the last bit sent was a
zero. This transition cannot be confused with clocked data CD0 or CD1.
since the transition will occur at the start of a bit cell. There will be no
transition in the middle of the bit cell. The IDL condition, as sent by a driver,
shall be maintained for a minimum of 2 bit times. The IDL condition shall be
detected within 1.6 bit times at the receiving device.
a. System jitter considerations make detection of IDL etd, end transmission delimiter. earlier than 1.3 bit times impractical. The specific
implementation of the phase-locked loop or equivalent clock recovery
mechanism determines the lower bound on the actual IDL detection
time. Adequate margin between lower bound and 1.6 bit times should
be considered.
243
b. Recovery of timing implicit in the data is easily accomplished at the

receiving side of the interface because of the wealth of binary transitions guaranteed to be in the encoded waveform, independent of the
data sequence. A phase-locked loop or equivalent mechanism maintains continuous tracking of the phase of the information on the Data
circuit.
Control Encoding. A simpler encoding mechanism is used for control signaling than for data signaling. The encoded symbols used in this signaling
mechanism are CS0, CS1, and IDL. The CS0 signal is a signal stream of
frequency equal to the bit rate BR.. The CS1 signal is a signal stream of
frequency equal to half of the bit rate BRr2.. If the interface supports more
than one bit rate, the bit rate in use on the data circuits is the one to which
the control signals are referenced. The IDL signal used on the control
circuits is the same as the IDL signal defined for data circuits. The Control
Out circuit is optional O. as is the one message on Control In CI..
The frequency tolerance of the CS1 and CS0 signals on the CO circuit is
"5%, and that of the CS1 signal on the CI circuit is "15%. The duty cycle
of the above signals is nominally 50%r50% and is no worse than 60%r40%.
The CS0 signal on the CI circuits has a frequency tolerance of BR q 25%,
y15% with the pulse widths no less than 35 ns and no greater than 70 ns at
the zero crossing points.
The meanings of the signals on the Control Out circuit DTE to
MAUmedium attachment unit. are as follows:
Signal
IDL
CS1
CS0
Message
Description
normal
mau request O.
isolate O.
Instructs the MAU to enter remain in. normal mode

Requests that the MAU should be made available
Instructs the MAU to enter remain in. monitor mode
The meaning of the signals on the Control In circuit MAU to DTE. are
as follows:
Signal
Message
Description
CS1
mau aailable
mau not aailable
Indicates MAU is not ready to output data
CS0
signal quality error
Indicates MAU has detected an error output data
IDL
SIGNALING RATE.
Indicates MAU is ready to output data
The CSMArCD specification covers multiple signaling rates.

The signaling specified for these sections is 10 Mbps " 0.01%.
It is intended that a given MDI medium-dependent interface . operate at
a single data rate. It is not precluded that specific DTE and MAU designs be
manually switched or set to alternate rates. A given local network operates at
244
a single signaling rate. To facilitate the configuration of operational systems,

DTE and MAU devices are labeled with the actual signaling rate used with
that device.
9.6.6.3 Medium Attachment Unit (MAU) 10BASE5. The MAU has the
following general characteristics:
a. Enables coupling the PLS by way of the AUI to the explicit baseband
coaxial transmission system defined in this clause of the standard
Ref. 3..
b. Supports message traffic at a data rate of 10 Mbps alternative data
rates may be considered in future additions to the standard ..
c. Provides for driving up to 500 m 1640 ft. of coaxial trunk cable
without the use of a repeater.
d. Permits the DTE to test the MAU and the medium itself.
e. Supports system configurations using the CSMArCD access mechanism defined with baseband signaling.
f. Supports a bus topology interconnection means.
g. Provides a communication channel with a mean BER at the PLS of
better than 1 = 10y8.
r CD, 10BROAD36Overview. Commonly,
9.6.6.4 Broadband CSMAr
broadband LANs use community antenna television CATV. techniques to
provide a multichannel LAN. In this case, each channel is governed by the
CSMArCD access method. The medium, of course, comprises CATV-type
cable, taps, connectors, and broadband amplifiers. A coaxial broadband
system permits the assignment of different frequency bands to multiple
applications.
The purpose of the MAU in this application . is to provide a means of
attaching devices to a broadband local network medium. The physical tap of
the coaxial cable. is a passive directional device such that the MAU transmission is directed toward the head-end location reverse direction.. On a
single-cable system the transmission from the MAU is at a carrier frequency
f1. A frequency translator or remodulator. located at the head-end up-converts to a carrier frequency f2, which is sent in the forward direction to the
taps receiver inputs .. On a dual-cable system the transmit and receive
carrier frequencies are identical both f1. and the MAU connects to the
medium via two taps, one on the receive cable and the other on the transmit
cable. The transmit and receive cables are connected to each other at the
head-end location. Figure 9.23 is a conceptual block diagram of broadband
single- and dual-cable systems.
The broadband MAU operates by accepting data from the attached data
termination equipment DTE. and transmitting a modulated radio-frequency
RF. data signal in a data band on the broadband coaxial cable system. All
MAUs attached to the cable system receive and demodulate this RF signal to
recover the DTE data. The broadband MAU emulates a baseband MAU
245
Figure 9.23. Broadband cable systems.
except for delay between transmission and reception, which is inherent in the
broadband cable system. A transmitting MAU logically compares the beginning of the received data with the data transmitted. Any difference between
them, which may be due to either a. errors caused by colliding transmissions,
b. reception of an earlier transmission from another MAU, or c. a bit error
on the channel, is interpreted as a collision.
246
When a collision is recognized, the MAU stops transmission in the data

band and begins transmission of an RF collision enforcement CE. signal in a
separate CE band adjacent to the data band. The CE signal is detected by all
MAUs and informs them that a collision has occurred. All MAUs signal to
their attached MAC controllers the presence of a collision. The transmitting
MACs then begin the collision handling process.
Collision enforcement is necessary because RF data signals from different
MAUs on the broadband cable system may be received at different power
levels. During a collision between RF data signals at different levels, the
MAU with the higher received power may see no errors in the detected data
stream. However, the MAU with the lower RF signal will see a difference
between transmitted and received data; this MAU then transmits the CE
signal to force recognition of the collision by all transmitting MAUs.
Broadband MAU Functional Requirements. The MAU component provides
the means by which signals on the physically separate AUI signal circuits to
and from the DTE and their associated interlayer messages are coupled to
the broadband coaxial medium. To achieve this basic objective, the MAU
component contains the following capabilities to handle message flow between the DTE and the broadband medium:
a. Transmit Function. The ability to transmit serial data bit streams
originating at the local DTE in a band-limited modulated RF carrier
form, to one or more remote DTEs on the same network.
b. Receie Function. The ability to receive a modulated RF data signal in
the band of interest from the broadband coaxial medium and demodulate it into a serial bit stream.
c. Collision Presence Function. The ability to detect the presence of two
or more stations concurrent transmissions.
d. Jabber Function. The ability of the MAU itself to interrupt the Transmit function and inhibit an abnormally long output data stream.
MAU TRANSMIT FUNCTION REQUIREMENTS.
The transmit function requirements

are shown in Figure 9.24. These requirements include the following capabilities:
a. Receive Manchester-encoded data sent by the local DTE to the
attached MAU on circuit DO transmit data pair..
b. Decode the Manchester-encoded data received on circuit DO to
produce NRZ data and recovered clock signal.
c. Scramble NRZ data using CCITT Rec. V.29 Chapter 8. scrambler
with seed changed on each transmitted packet.
d. Transform the incoming bits prior to modulation. to provide an
unscrambled alternating zeroone pattern terminated by an Unscrambled Mode Delimiter UMD., scramble the remainder of the incoming
247
Figure 9.24. Transmit function requirements block diagram.
preamble, Start Frame Delimiter SFD., and data frame; and append
an unscrambled post-amble Broadband End of Frame Delimiter
wBEOFDx..
e. Differentially encode the packet generated previously.
f. Produce a band-limited double-sideband, suppressed carrier, binary
PSK modulated RF signal representing the generated differentially
encoded packet.
g. Drive the coaxial cable with the modulated RF signal.
MAU RECEIVE FUNCTION REQUIREMENTS.
The receive functions include the
following:
a. Receive the differentially encoded binary PSK modulated RF signal
from the broadband coaxial medium.
b. Receive the data band RF signals and reject signals in bands other
than the data band rejection of signals in the adjacent collision
enforcement band is optional..
c. Demodulate and differentially decode the incoming RF data signal
from the coaxial medium to provide a receive bit stream that represents the scrambled bit stream at the transmitter.
d. Descramble the receive bit stream using a self-synchronizing descrambler.
e. Manchester encode the descrambled bit stream.
f. Send to the DTE, using Circuit DI receive data pair., an additional,
locally generated, Manchester-encoded preamble equal to the number
of preamble bits lost in the receive data path plus or minus one bit.,
followed by Manchester-encoded bit stream. No more than 6 preamble
bits may be lost from the preamble presented to Circuit DO at the
transmitting MAU.
g. Detect end of frame, using the postamble BEOFD., and ensure that
no extraneous bits are sent to the DTE on Circuit DI.
248
h. Receive signals in the collision enforcement band and reject signals in

the data band and all other bands on the broadband medium.
MAU COLLISION DETECTION FUNCTION REQUIREMENTS.
The MAU shall perform

the following functions to meet the collision detection requirements:
a. Store the scrambled bits not differentially encoded. in the transmit
section through to the last bit in the source address.
b. Detect the UMD in the transmit and receive paths.
c. Compare received scrambled bits after the received UMD with transmitted scrambled bits after the transmit UMD through to the last bit
in the source address.
d. A Receive UMD Timer function shall be performed by the MAU. The
timer shall be as long as the time required from initial detection of RF
data signal presence to detection of a UMD in a normally received no
collision. packet.
e. Enter a LOCAL COLLISION DETection state if one of the following
occurs:
1. A bit error is found in the bit compare process through the last bit
in the source address.
2. The Receive UMD Timer expires before a UMD is detected in the
received bit stream.
3. The MAU receives the output that is, transmit. signal from the
AUI AFTER having received an RF signal from the coaxial cable.
f. Upon entering the LOCAL COLLISION DET state, cease transmission in the data band and commence transmission in the collision
enforcement band for as long as the DTE continues to send data to
the MAU.
g. Upon entering the LOCAL COLLISION DET state send the
signal quality error SQE. message on Circuit CI collision presence
pair. using the CS0 signal for as long as RF signals are detected on the
broadband coaxial medium in either the data or collision enforcement
bands.
h. Detect power in the collision enforcement band and send the SQE
message on Circuit CI using the CS0 signal. Send the SQE message for
as long as energy is detected in the collision enforcement band.
i. Ensure that during collisions, due to phase cancellations of the colliding carriers, Circuit DI does not become inactive before Circuit CI
becomes active.
j. Test the collision detection circuitry following every transmission that
does not encounter a collision. This test consists of transmitting a
burst of collision enforcement RF signal after the end of the postamble transmission and detecting this burst on the receive side. If the
burst is detected, the CS0 BR. signal is sent on Circuit CI of the
transmitting MAU.
249
TABLE 9.7 Broadband Cable System Electrical Requirements

Impedance
Return loss
Transmit level
Receive level
Maximum receive noise level
Loss variation a (per 18-MHz band)
Path loss (between any transmit port and receive
port, including loss variation)
Group delay variation
around data carrier
over 18 MHz band
a
75
14 dB min
q50 dBmV " 2 dB
q6 dBmV " 10 dB
y30 dBmV r 14 MHz
2 dB min, 52 dB max
36 dB min, 52 dB max
20 ns r 10 MHz max
34 ns max
Not including head-end.
Source: Table 11.5-1, page 246, Ref. 3.
Figure 9.25. Packet format and timing diagram, AUI to coaxial cable interface. (From Figure
11-6, page 233, IEEE Std. 802.3, Ref. 3.)
Selected Coaxial Cable Electrical Interface Requirements. These requirements are listed in Table 9.7.
Figure 9.25 shows the packet format and timing diagram at the coaxial
cable interface. Tables 9.8 and 9.9 show the single cable and dual-cable
frequency plans.
9.6.7 System Considerations for Multisegment 10-Mbps
Baseband Networks
9.6.7.1 Overview. Two transmission models are presented in Section 13 of
Ref. 3: Model 1 and Model 2. These models require that the network size be
limited to control round-trip propagation delay, and that the number of
250
TABLE 9.8 Single-Cable Frequency Allocations (MHz) a

Receiver
Translation
156.25 MHz
Transmitter
Translation
192.25 Mhz
Data
Carrier
Coll Enf b
Center Freq
Transmit
Band
Head-End
Local Osc
Receive
Band
Head-End
Local Osc
Receive
Band
43
49
55
61c
67
73
52
58
64
70
76
82
35.7553.75
41.7559.75
47.7565.75
53.7571.75
59.7577.75
65.7583.75
245.75
257.75
269.75
281.75
293.75
305.75
192210
198216
204222
210228
216234
222240
192.25
192.25
192.25
192.25
192.25
192.25
228246
234252
240258
246264
252270
258276
Some of these optional bands are overlapping. Frequency tolerance of the data carrier and
head-end local oscillator shall each be "25 kHz.
b
Coll Enf, collision enforcement.
c
Preferred frequency allocation.
TABLE 9.9 Dual-Cable Frequency Allocations (MHz)

Data
Carrier a
Coll Enf b
Center Freq
Data
Band
Coll Enf b
Band
43
49
55
61c
67
73
235.25
241.25
247.25
253.25 c
259.25
265.25
52
58
64
70
76
82
244.25
250.25
256.25
262.25
268.25
274.25
3650
4256
4862
5468
6974
6680
228242
234248
240254
246260
252266
258272
5054
5660
6266
6872
7478
8084
242246
248252
254258
260264
266270
272276
Some of these optional bands are overlapping. Frequency tolerance of the data carrier shall be
"25 kHz.
b
Coll Enf, collision enforcement.
c
Preferred frequency allocations.
repeaters between any two DTEs be limited in order to limit the shrinkage of
the interpacket gap as it travels through the network. Model 1 is the more
conservative that meets the two basic design criteria set forth above. The
Model 2 transmission system model validates an additional broad set of
topologies that are fully functional and do not fit within the simpler but more
restrictive rules of Model 1.
The physical size of a CSMArCD network is limited by the characteristics of individual network components. These characteristics include the
TABLE 9.10
251
Delays for Network Media Segments
Media Type
Mixing segment
10BASE5
10BASE2
10BASE-FP
Link segment
FOIRL
10BASE-T
10BASE-FB
10BASE-FL
AUI d 1 DTE r 1 MAU
Maximum Number of Maximum Segment

MAUs per Segment
Length (m)
100
30
33a
2
2
2
2
Maximum Medium
Delay per Segment (ns)
500
185
1000b
2165
950
5000
1000
100c
1000
2000
50
5000
1000
10 000
10 000
257
Actual number depends on the passive-star characteristics.

In addition, a MAU to passive-star link will not exceed 500 m.
Actual maximum segment length depends on cable characteristics.
d
AUI is not a segment.
b
c
Source: Table 13-1, page 284, Ref. 3.
following:
a.
b.
c.
d.
e.
f.
Media lengths and their associated propagation time delay

Delay of repeater units start-up and steady-state .
Delay of MAUs start-up and steady-state .
Interpacket gap shrinkage due to repeater units
Delays within the DTE associated with the CSMArCD access method
Collision detect and deassertion times associated with MAUs
Table 9.10 summarizes the delays for the various network media segments.
For more specific delays of the various MAUs, the reader should consult the
reference specification, Sections 818.
9.6.7.2 Shrinkage of the Interpacket Gap (IPG). The worst-case variabilities of transmission elements in the network plus some of the signal
reconstruction facilities required in the 10-Mbps baseband repeater specification combine in such a way that the gap between two packets travelling
across the network may be reduced below the interFrameGap specified in
Tables 9.3 through 9.6. This parameter limits the equipment i.e., number of
repeaters . between any two DTEs. Again the limit applies to all combinations of DTEs on any network, but the worst case is apparent from an
inspection of a map or schematic representation of the topology in question.
9.6.7.3 Transmission System Model 1. The following network topology
constraints apply to networks Transmission System Model 1. If no segment
length constraints are given for a segment type, the maximum segment
length, as defined in the relevant MAU clause, applies.
252
a. Repeater sets are required for all segment interconnection.

b. MAUs that are part of repeater sets count toward the maximum
number of MAUs on a segment.
c. The transmission path permitted between any two DTEs may consist
of up to five segments, four repeater sets including optional AUIs.,
two MAUs, and two AUIs.
d. AUI cables for 10BASE-FP and 10BASE-FL shall not exceed 25 m.
Since two MAUs per segment are required, 25 m per MAU results in
a total AUI cable length of 50 m per segment..
e. When a transmission path consists of four repeater sets and five
segments, up to three of the segments may be mixing and the remainder must be link segments Figures 9.26, 9.27, and 9.30.. When five
segments are present, each fiber-optic link segment FOIRL, 10BASEFB, or 10BASE-FL. shall not exceed 500 m, and each 10BASE-FP
segment shall not exceed 300 m.
f. When a transmission path consists of three repeater sets and four
segments Figures 9.28 and 9.29., the following restrictions apply:
1. The maximum allowable length of any inter-repeater fiber segment
shall not exceed 1000 m for FOIRL, 10BASE-FB, and 10BASE-FL
segments and shall not exceed 700 m for 10BASE-FP segments.
2. The maximum allowable length of any repeater to DTE fiber
segment shall not exceed 400 m for 10BASE-FL segments and
shall not exceed 300 m for 10BASE-FP segments and 400 m for
segments terminated in a 10BASE-FL MAU.
3. There is no restriction on the number of mixing segments in this
case.
9.6.7.4 Typical Topologies Applicable to Transmission Model 1. Figures 9.26 through 9.30 show several of the network topologies consistent with
Transmission Model 1.
9.7 100-Mbps CSMA r CD BASEBAND NETWORKS

9.7.1 OverviewKey Points
Figure 9.31 shows the relationship between 100BASE-T, the existing IEEE
802.3 CSMArCD MAC. and the ISO OSI reference model.
100BASE-T uses existing IEEE 802.3 MAC layer interface, connected
through a media-independent interface layer to a physical layer entity PHY.
sublayer such as 100BASE-T4, 100BASE-TX, or 100BASE-FX.
100-Mbps CSMArCD BASEBAND NETWORKS
253
254
Figure 9.27. Example of maximum transmission path using coaxial cable segments,
10BASE-T link segments, and fiber-optic link segments.
100BASE-T extends the IEEE 802.3 MAC to 100-Mbps operation. The

following key points should be considered:
The bit rate is faster.

Bit times are shorter.
Packet transmission times are reduced.
Cable delay budgets are smaller.
All of these factors are in proportion to the change in bit rate i.e., 10 Mbps
vs. 100 Mbps.. This means that the ratio of packet duration to network
propagation delay for 100BASE-T is the same as 10BASE-T.
9.7.1.1 Reconciliation Sublayer (RS) and Media-Independent Interface
(MII) Introduction. The media-independent interface MII. provides an
interconnection between the media access control MAC. sublayer and the
physical layer entities PHY. and between PHY layer and station management STA. entities. This MII is capable of supporting both 10-Mbps and
100-Mbps data rates through 4-bit wide nibble wide. transmit and receive
paths. The reconciliation sublayer provides a mapping between signals provided at the MII and the MACrPLS service definition.
255
Figure 9.28. Example of maximum transmission path with three repeater sets, four link
segments (two are 100-m 10BASE-T and two are 1-km fiber).
9.7.1.2 Physical Layer Signaling Systems. The MII attaches to any one
of a number of physical layer implementations as expressed in the following
table:
Designation
100BASE-T4
100BASE-FX
100BASE-TX
100BASE-T2
Physical Layer Description

Four pairs Cat. 3, 4, or 5 balanced cabling or 150-ohm
shielded balanced cabling
Two multimode fibers FDDIrANSI X3T12.
See 100BASE-FX, both defined under 100BASE-X
Two pairs of Cat. 3, 4, or 5 balanced cabling
Note: Category 3, 4, and 5 cabling is described in Chapter 14.
256
Figure 9.29. Example of maximum transmission path with three repeater sets, four segments
(one 1-km 10BASE-FB, one 1-km FOIRL, one 400-m 10BASE-FL, and one 300-m 10BASE-FP).
9.7.1.3 Repeater. Repeater sets are an integral part of any 100BASE-T

network with more than two DTEs in a collision domain. They extend the
physical system topology by coupling two or more segments. Multiple repeaters are permitted within a single collision domain to provide the maximum path length.
9.7.2 Reconciliation Sublayer (RS) and Media Independent-Interface
(MII) Description Details
Figure 9.32 illustrates the relationship of the reconciliation sublayer and MII
to the OSI reference model.
257
Figure 9.30. Example of maximum transmission path with four repeater sets, five segments
(two 500-m 10BASE-FB, one 500-m FOIRL, one 500-m 10BASE-FL, and one 300-m
10BASE-FP).
Summary of Major Concepts

a. Each direction of data transfer is serviced with 7 making a total of 14.
signals: data a 4-bit bundle ., delimiter error, and check.
b. Two media status signals are provided. One indicates the presence of
carrier, and the other indicates the occurrence of a collision.
c. A management interface comprised of two signals provides access to
management parameters and services.
d. The reconciliation sublayer maps the signal set provided at the MII to
the PLS service definition specified in Section 6 of the reference
publication.
258
259
260
Figure 9.33. MII frame format. Sfd s start frame delimiter; efd s end frame delimiter.
The interface through the MII connector is used to provide media independence for various forms of unshielded twisted-pair wiring, shielded
twisted-pair wiring, fiber-optic cabling, and potentially other media, so that
identical media access controllers may be used with any of these media.
Rates of Operation. The MII can support two specific data rates, 10 Mbps
and 100 Mbps. The functionality is identical at both data rates, as are the
signal timing relationships. The only difference between 10-Mbps and 100Mbps operation is the nominal clock frequency.
Relationship of MII and GMII (gigabit media-independent interface). The
GMII is similar to the MII. The GMII uses MII management interface and
register set specified in Section 22.2.4 of the reference document. These
common elements of operation allow station management to determine PHY
capabilities for any supported speed of operation and configure the station
based on those capabilities. In a station supporting both MII and GMII
operation, configuration of the station would include enabling either the MII
or GMII operation as appropriate for the data rate of the selected PHY.
Most of the MII and GMII signals use the same names, but the width of
the RXD and TXD data bundles and the semantics of the associated control
signals differ between MII and GMII operation. The GMII transmit path
clocking differs significantly from MII clocking. MII operation of these
signals and clocks is specified in Section 22 and GMII operation is specified
in Section 35 of the reference publication.
9.7.2.1 MII Frame Structure. The data frames transmitted through the
MII have the format as shown in Figure 9.33.
For MII transmission and reception of each octet of data is done a nibble*
at a time with the order of nibble transmission and reception as shown in
Figure 9.34. The bits of each octet are transmitted and received as two
nibbles: Bits 0 through 3 correspond to bits 0 through 3 of the first nibble
transmitted and received, and bits 4 through 7 of the octet correspond to bits
0 through 3 of the second nibble.
Interframe. The interframe period provides an observation window for an
unspecified amount of time during which no data activity occurs on the MII.
The absence of data activity is indicated by the de-assertion of the RX DV
receive data valid. signal on the receive path, and the de-assertion of the
TX EN transmit enable. signal on the transmit path. The MAC interFrameSpacing parameter defined in Tables 9.3 through 9.6 is measured from
*A nibble is a sequence of 4 bits.
261
Figure 9.34. Octet r nibble transmit and receive order. (From Figure 22-11, page 500, IEEE
Std. 802.3, Ref. 3.)
the de-assertion of the CRS carrier sense. signal to the assertion of the CRS
signal.
Preamble. The preamble begins a frame transmission and consists of at least
seven octets of alternating 1s and 0s transmitted serially.
Start of Frame Delimiter (sfd). The sdf indicates the start of a frame and
follows the preamble. It consists of the octet:
10101011
Data. The data in a well-formed frame consists of N octets of data transmitted as 2 N nibbles. For each octet of data, the transmit order of each nibble is
as specified in Figure 9.34. Data in a collision fragment may consist of an odd
number of nibbles.
9.7.3 100BASE-T4, Its PMA and PCS
9.7.3.1 Introduction. 100BASE-T4 and baseband medium specifications
are aimed at users who want 100-Mbps performance, but would like to retain
the benefits of using voice-grade twisted-pair cable. 100BASE-T4 requires
four pairs of Category 3 or better cables.
9.7.3.2 Objectives. The following are the objectives of 100BASE-T4:
a. To support the CSMArCD MAC in the half duplex mode of operation.
b. To support the 100BASE-T MII, repeater, and optional autonegotiation.
c. To provide 100-Mbps data rate at the MII.
262
d. To provide for operating over unshielded twisted pairs of Category 3,

4, or 5 cable, installed as horizontal runs in accordance with ISOrIEC
11801: 1995, as specified in 23.6 of the reference document, at distances up to 100 m 328 ft..
e. To allow for a nominal network extent of 200 m, including:
1. Unshielded twisted-pair links of 100 m.
2. Two-repeater networks of approximately a 200-m span.
f. To provide a communication channel with a mean ternary symbol
error rate, at the PMA service interface, of less than one part in 10 8.
9.7.3.3 Description. The PCS transmit function accepts data nibbles from
the MII. The PCS transmit function encodes these nibbles using an 8B6T
coding scheme described below. It passes the resulting ternary symbols to the
PMA. In the reverse direction, the PMA conveys received ternary symbols to
the PCS receive function. The PCS receive function decodes them into
octets, and then it passes the octets one nibble at a time up to the MII. The
PCS also contains a PCS carrier sense function, a PCS error sense function, a
Figure 9.35. Division of responsibilities between 100BASE-T4 PCS and PMA. (From Figure
23-2, page 537, IEEE Std. 802.3, Ref. 3.)
263
Figure 9.36. Use of wire pairs. (From Figure 23-3, page 537, IEEE Std. 802.3, Ref. 3.)
PCS collision presence function, and a management interface. Figure 9.35

shows the division of responsibilities between PCS, PMA and MDI layers.
Physical level communication between PHY entities takes place over four
twisted pairs using Category 3, 4, or 5 UTPs. Figure 9.36 shows how the PHY
manages the four twisted pairs at its disposal.
The 100BASE-T4 transmission algorithm always leaves one pair open for
detecting carrier from the far end see Figure 9.36.. Leaving one pair open
for carrier detection in each direction greatly simplifies media access control.
All collision detection functions are accomplished using only the unidirectional pairs TX D1 and RX D2, in a manner similar to 10BASE-T. This
collision detection strategy leaves three pairs in each direction free for data
transmission, which uses an 8B6T block code, schematically represented in
Figure 9.37.
8B6T coding, as used with 100BASE-T4 signaling, maps data octets into
ternary symbols. Each octet is mapped to a pattern of 6 ternary symbols,
called a 6T code group. The 6T code groups are fanned out to three
independent serial channels. The effective data rate carried on each pair is
one third of 100 Mbps, which is 33.333 . . . Mbps. The ternary symbol transmission rate on each pair is 6r8 times 33.33 Mbps, or precisely 25.000 MHz.
Figure 9.37. 8B6T coding.
264
9.7.3.4 Summary of Physical Medium Attachment (PMA) Specification. The PMA couples messages from the PMA service interface onto the
twisted-pair physical medium. The PMA provides communications, at 100
Mbps, over four pairs of twisted-pair wiring up to 100 m in length.
The PMA transmit function, shown in Figure 9.35, comprises three independent ternary data transmitters. Upon receipt of a PMA UNITDATA.request message, the PMA synthesizes one ternary symbol on each of the three
output channels TX D1, BI D3, and BI D4.. Each output driver has a
ternary output, meaning that the output waveform can assume any of three
values, corresponding to the transmission of ternary symbols CS0, CS1, or
CS-1. A CS0 is a ternary symbol 0 with an output voltage of 0 volts, CS1 is a
ternary symbol 1 with a nominal peak voltage of q3.5 V, and a CS-1 conveys
the ternary symbol y1 and has a nominal peak voltage of y3.5 volts., on
each of the twisted pairs.
The PMA receive function comprises three independent ternary data
receivers. The receivers are responsible for acquiring clock, decoding the
start of stream delimiter SSD. on each channel, and providing data to the
PCS in the synchronous fashion defined by the PMA UNIDATA.indicate
message. The PMA also contains functions for PMA carrier sense and link
integrity.
9.7.4 Physical Coding Sublayer (PCS) and Physical Medium
Attachment (PMA) Sublayer, Type 100BASE-X
9.7.4.1 Objectives. The following are the objectives of 100BASE-X:
a. Support the CSMArCD MAC in the half-duplex and the full-duplex
modes of operation.
b. Support the 100BASE-T MII, repeater, and optional autonegotiation.
c. Provide 100-Mbps data rate at the MII.
d. Support cable plants using Category 5 UTP, 150 STP, or optical
fiber, compliant with ISOrIEC 11801.
e. Allow for a nominal network extent of 200400 m, including:
1. Unshielded twisted-pair links of 100 m
2. Two repeater networks of approximately 200 m span
3. One repeater network of approximately 300 m span using fiber.
4. DTErDTE links of approximately 400 m half-duplex mode using
fiber. and 2 km full-duplex mode using multimode fiber.
f. Preserve full-duplex behavior of underlying PMD channels.
9.7.4.2 Functional Requirements. The PCS comprises the transmit, receive, and carrier sense functions for 100BASE-T. In addition, the collisionDetect signal required by the MAC COL on the MII. is derived from the
265
PMA code-bit stream. The PCS shields the reconciliation sublayer and
MAC. from the specific nature of the underlying channel. Specifically for
receiving, the 100BASE-X PCS passes to the MII a sequence of data nibbles
derived from incoming code groups, each comprised of five code bits,
received from the medium. Code-group alignment and MAC packet delimiting is performed by embedding special nondata code groups. The MII uses a
nibble-wide, synchronous data path, with packet delimiting being provided by
separate TX EN and RX DV signals. The PCS provides the functions
necessary to map these two views of the exchanged data. The process is
reversed for transmit.
The following provides a detailed specification of the functions performed
by the PCS, which comprise five parallel processes transmit, transmit bits,
receive, receive bits, and carrier sense.. Figure 9.38 includes a functional
block diagram of the PCS.
The receive bits process accepts continuous code bits via the
PMA UNITDATA.indicate primitive. Receive monitors these bits and generates RXD 3 : 0:, RX DV and RX ER on the MII, and the internal flag,
receiving, used by the carrier sense and transmit processes.
The transmit process generates continuous code groups based upon the
TXD 3 : 0:, TX EN, and TX ER signals on the MII. These code groups
are transmitted by transmit bits via the PMA UNITDATA.request primitive.
The transmit process generates the MII signal COL based on whether a
reception is occurring simultaneously with transmission. Additionally, it generates the internal flag, transmitting, for use by the carrier sense process.
The carrier sense process asserts the MII signal CRS when either transmitting or receiving is TRUE. Both the transmit and receive processes
monitor link status via the PMA LINK.indicate primitive, to account for
potential link failure conditions.
Code Groups. The PCS maps 4-bit nibbles from the MII into 5-bit code
groups, and vice versa, using a 4Br5B block coding scheme. A code group is
a consecutive sequence of five code-bits interpreted and mapped by the PCS.
Implicit in the definition of a code-group is an establishment of code group
boundaries by an alignment function within the PCS receive process. It is
important to note that, with the sole exception of the SSD, which is used to
achieve alignment, code groups are undetectable and have no meaning
outside the 100BASE-X physical protocol data unit, called a stream.
The coding method used provides the following:
a. Adequate codes 32. to provide for all data code groups 16. plus
necessary control code-groups
b. Appropriate coding efficiency 4 data bits per 5 code bits; 80%. to
effect a 100-Mbps physical layer interface on a 125-Mbps physical
channel as provided by FDDI PMDs
266
Figure 9.38. Functional block diagram, 100BASE-X. (From Figure 24-4, page 615, IEEE Std.
802.3, Ref. 3.)
c. Sufficient transition density to facilitate clock recovery when not

scrambled.
Table 9.11 specifies the interpretation assigned to each 5-bit code group,
including the mapping to the nibble-wide TXD or RXD. Data signals on
267
268
the MII. The 32 code groups are divided into four categories, as shown in
Table 9.11.
DATA CODE GROUPS.
A data code group conveys one nibble of arbitrary data

between the MII and the PCS. The sequence of data code groups is arbitrary,
where any data code group can be followed by any other data code group.
Data code groups are coded and decoded but not interpreted by the PCS.
Successful decoding of data code groups depends on proper receipt of the
start-of-stream delimiter sequence as defined in Table 9.11.
IDLE CODE GROUPS.
The idle code group rIr. is transferred between

streams. It provides a continuous fill pattern to establish and maintain clock
synchronization. Idle code groups are emitted from, and interpreted by,
the PCS.
CONTROL CODE GROUPS.
The control code groups are used in pairs rJrKr,

.
rTrRr to delimit MAC packets. Control code groups are emitted from,
and interpreted by, the PCS.
START-OF-STREAM DELIMITER
rJ r K r .. A start-of-stream delimiter SSD. is

used to delineate the boundary of a data transmission sequence and to
authenticate carrier events. The SSD is unique in that it may be recognized
independently of previously established code-group boundaries. The receive
function within the PCS uses the SSD to establish code group boundaries.
An SSD consists of the sequence rJrKr.
END-OF-STREAM DELIMITER
rT r R r .. An end-of-stream delimiter ESD.

terminates all normal data transmissions. Unlike the SSD, an ESD cannot be
recognized independent of previously established code group boundaries. An
ESD consists of the sequence rTrRr.
INVALID CODE GROUPS. The rHr code group indicates that the PCSs client
wishes to indicate a transmit error to its peer entity. The normal use of this
indicator is for repeaters to propagate received errors. Transmit error code
groups are emitted from the PCS, at the request of the PCSs client through
the use of the TX ER signal.
The presence of any invalid code group on the medium, including rHr,
denotes a collision artifact or an error condition. Invalid code groups are not
intentionally transmitted onto the medium by DTEs. The PCS indicates the
reception of an invalid code group on the MII through the use of the
RX ER signal.
Encapsulation. The 100BASE-X PCS accepts frames from the MAC through
the reconciliation sublayer and MII. Due to the continuously signaled nature
of the underlying PMA, and the encoding performed by the PCS, the
269
Figure 9.39. PCS encapsulation. (From Figure 24-5, page 619, IEEE Std. 802.3, Ref. 3.)
100BASE-X PCS encapsulates the MAC frame 100BASE-X service data

unit, SDU. into a physical layer stream 100BASE-X protocol data unit,
PDU..
Except for the two code-group SSD, data nibbles within the SDU including the non-SSD portions of the MAC preamble and SFD. are not interpreted by the 100BASE-X PHY. The conversion from a MAC frame to a
physical layer stream and back to a MAC frame is transparent to the MAC.
Figure 9.39 depicts the mapping between MAC frames and physical layer
streams.
A properly formed stream can be viewed as comprising three elements:
a. Start-of-Stream Delimiter SSD .. The start of a physical layer stream is
indicated by an SSD, as defined in Section 9.7.4.2. The SSD replaces
the first octet of the preamble from the MAC frame and vice versa.
b. Data Code Groups. Between delimiters SSD and ESD., the PCS
conveys data code groups corresponding to the data nibbled of the
MII. Data code groups comprise the 100BASE-X service data unit
SDU.. Data nibbles within the SDU including those corresponding
to the MAC preamble and SFD. are not interpreted by the 100BASE-X
PCS.
c. End-of-Stream Delimiter ESD .. The end of a properly formed stream
is indicated by an ESD, as defined in Section 9.7.4.2. The ESD is
transmitted by the PCS following the de-assertion of TX EN on the
MII, which corresponds to the last data nibble composing the FCS
270
Figure 9.40. PCS reference diagram. (From Figure 24-6, page 620, IEEE Std. 802.3, Ref. 3.)
from the MAC. It is transmitted during the period considered by the

MAC to be the interframe gap IFG.. On reception, ESD is interpreted by the PCS as terminating the SDU.
Between streams, IDLE code groups are conveyed between the PCS and
PMA.
Data Delay. The PCS maps a nonaligned code bit data path from the PMA
to be an aligned, nibble-wide data path on the MII, for both transmit and
receive. Logically, received bits must be buffered to facilitate SSD detection
and alignment, code translation, and ESD detection. These functions necessitate even longer delays of the incoming code-bit stream.
When the MII is present as an exposed interface, the MII signals TX CLK
and RX CLK, not depicted in the following state diagrams, are generated by
the PCS in accordance with Section 22 of the reference document.
Mapping Between MII and PMA. Figure 9.40 depicts the mapping of the
nibble-wide data path of the MII to the 5-bit-wide code groups internal to
the PCS. and the code-bit path of the PMA interface.
Upon receipt of a nibble from the MII, the PCS encodes it into a 5-bit
code group in accordance with the code group description paragraphs. Code
groups are serialized into code bits and passed to the PMA for transmission
on the underlying medium, according to Figure 9.40. The first transmitted
code bit of a code group is bit 4, and the last code bit transmitted is bit 0.
There is no numerical significance ascribed to the bits within a code group;
that is, the code group is simply a 5-bit pattern that has some predefined
interpretation.
271
Similarly, the PCS deserializes code bits received from the PMA, in
accordance with Figure 9.40. After alignment is achieved, based on SSD
detection, the PCS converts code groups into MII data nibbles, according to
the code group description paragraphs.

100BASE-T Networks
This section provides a brief review on building 100BASE-T networks. The
100BASE-T technology is designed to be deployed in both homogeneous
100-Mbps networks and heterogeneous 10r100-Mbps mixed CSMArCD
networks. Network topologies can be developed within a single 100BASE-T
collision domain, but maximum flexibility is achieved by designing multiplecollision-domain networks that are joined by bridges andror routers configured to provide a range of service levels to DTEs. For example, a combined
100BASE-Tr10BASE-T system built with repeaters and bridges can deliver
dedicated 100-Mbps, shared 100-Mbps, dedicated 10-Mbps, and shared 10Mbps service to DTEs, as shown in Figure 9.41. The effective bit rate
capacity of shared service is controlled by the number of DTEs that share the
service.
Linking multiple 100BASE-T collision domains with bridges maximizes
flexibility. Bridged topology design can provide single bit rate capacity
as shown in Figure 9.41 or multiple bit rate capacity services as shown in
Figure 9.42.
Individual collision domains can be linked by single devices, as shown in
Figures 9.41 and 9.42, or by multiple devices from any of several transmission
systems. The design of multiple-collision-domain networks is governed by the
rules defining each of the transmission systems incorporated into the design.
The design of shared bit rate capacity 10-Mbps collision domains is
defined in Section 9.6.7 of this document. The design of shared bit rate
capacity 100-Mbps CSMArCD collision domains is defined in the present
section.
9.7.5.1 Physical Size of the Network. The physical size of a CSMArCD
network is limited by the characteristics of individual network components.
These characteristics include the following:
a.
b.
c.
d.
e.
f.
Media lengths and their associated propagation delay

Delay of repeater units start-up, steady-state, and end of event.
Delay of MAUs and PHYs start-up, steady-state, and end of event.
Interpacket gap shrinkage due to repeater units
Delays within the DTE associated with the CSMArCD access method
Collision detect and de-assertion times associated with the MAUs and
PHYs
272
Figure 9.41. 100-Mbps multiple-collision-domain topology using multiport bridge. (From

Figure 29-1, page 735, IEEE Std. 802.3, Ref. 3.)
Table 9.12 summarizes the delays for 100BASE-T media segments.

9.7.5.2 Transmission Models. Two different transmission system models
are used to determine maximum network size and complexity:
Transmission System Model 1. The following network topology constraints
apply to networks using Transmission System Model 1.
a. All balanced cable copper. segments less than or equal to 100 m each.
b. Fiber segments less than or equal to 412 m each.
c. MII cables for 100BASE-T shall not exceed 0.5 m each. When evaluating system topology, MII cable delays need not be accounted for
separately. Delays attributable to the MII are incorporated into DTE
and repeater component delays.
273
Figure 9.42. Multiple bit rate capacity, multiple-collision-domain topology using a multiport
bridge.
TABLE 9.12
Delays for Network Media Segments Model 1
Media Type
Balanced Cable Link Segment 100BASE-T
Fiber Link Segment
a
Maximum
Number of Maximum
Maximum Medium
PHYs per
Segment Round-Trip Delay per
Segment Length (m)
Segment (BT) a
2
2
100
412
114
412
BT, bit times.
Transmission System Model 2. The physical size and number of topological

elements in a 100BASE-T network is limited primarily by round-trip collision
delay. A network configuration must be validated against collision delay using
a network model. Since there are a limited number of topology models for
any 100BASE-T collision domain, the modeling process is quite straightforward and can easily be done either manually or with a spreadsheet.
This model derives from Model 2 presented in Section 9.6.7. Modifications
have been made to accommodate adjustments for DTE, repeater, and cable
speeds.
9.7.5.3 Round-Trip Collision DelayKey to Network Operation and
Maximum Size. For a network to be valid, it must be possible for any two
DTEs on the network to contend for the network at the same time. Each
station attempting to transmit must be notified of the contention by the
274
TABLE 9.13
Network Component Delays, Transmission System Model 2
Component
Two TX r FX DTEs
Two T4 DTEs
Two T2 DTEs
One T2 or T4 and one TX r FX DTE a
Cat. 3 cabling segment
STP cabling segment
Fiber-optic cabling segment
Class I repeater
Class II repeater with all ports TX r FX
Class II repeater with any port T4
Class II repeater with any port T2
Round-Trip Delay
in Bit Times per Meter
Maximum Round-Trip
Delay in Bit Times
1.14
1.14
1.112
1.112
1.0
100
138
96
127
114 (100 m)
114 (100 m)
111.2 (100 m)
111.2 (100 m)
412 (412 m)
140
92
67
90
Worst-case values are used (TX r FX values for MAC transmit start and MDI input to collision
detect; T4 value for MDI input to MDI output).
returned collision signal within the collision window. Additionally,

the maximum length fragment created must contain less than 512 bits after
the start-of-frame delimiter SFD.. These requirements limit the physical
diameter maximum distance between DTEs. of a network. The maximum
round-trip delay must be qualified between all pairs of DTEs in the network.
In practice this means that the qualification must be done between those
that, by inspection of the topology, are candidates for the longest delay. The
following network modeling methodology is provided to assist that calculation.
Table 9.13 is presented for guidelines on various system component
delays, Transmission System Model 2.
9.8 1000-Mbps CSMA r CD NETWORKS

9.8.1 Overview
Gigabit Ethernet CSMArCD. uses the extended IEEE 802.3 MAC layer
interface, connected through a gigabit media-independent interface layer to
physical layer entities PHY sublayers . such as 1000BASE-LX, 1000BASE-SX,
1000BASE-CX, and 1000BASE-T. The architectural positioning of gigabit
Ethernet with the total CSMArCD LAN structure and the OSI reference
model is shown in Figure 9.43.
Gigabit Ethernet extends the IEEE 802-3 MAC beyond 100 Mbps to 1000
Mbps. The bit rate is faster, and the bit times are shorterboth in proportion to the change in bandwidth. In full-duplex mode, the minimum packet
1000-Mbps CSMArCD NETWORKS
275
276
transmission time has been reduced by a factor of ten. Achievable topologies

for 1000-Mbps full-duplex operation are comparable to those found in
100BASE-T full-duplex mode. In half-duplex mode, the minimum packet
transmission time has been reduced, but not by a factor of ten. Cable delay
budgets are similar to those in 100BASE-T. The resulting achievable topologies for the half-duplex 1000-Mbps CSMArCD MAC are similar to those
found in half-duplex 100BASE-T.
9.8.2 Reconciliation Sublayer (RS) and Gigabit
Media-Independent Interface (GMII)
9.8.2.1 Overview. Figure 9.43 shows the relationship of the reconciliation
sublayer and the GMII with the OSI reference model. The purpose of this
interface is to provide a simple, inexpensive, and easy-to-implement interconnection between media access control MAC. sublayer and PHYs and between PHYs and station management. This interface has the following
characteristics:
a.
b.
c.
d.
e.
It is capable of supporting 1000-Mbps operation.

Data and delimiters are synchronous to clock references.
It provides independent 8-bit-wide transmit and receive data paths.
It provides a simple management interface.
It uses signal levels, compatible with common CMOS digital ASIC
processes and some bipolar processes.
f. It provides for full-duplex operation.
9.8.2.2 GMII Data Stream. Data frames transmitted through the GMII
are transferred within the data stream as shown in Figure 9.44.
For the GMII, transmission and reception of each octet of data is as
Interframe interframe : . The interframe interframe : period on a GMII
transmit or receive path is an interval during which no data activity occurs on
the path. Between bursts or single frame transmission, the absence of data
activity on the receive path is indicated by the de-assertion of both RX DV
receive data valid. and RX ER receive error. or the de-assertion of the
Figure 9.44. GMII data stream.
277
Figure 9.45. Relationship of data bundles to MAC serial bit stream. (From Figure 35-16,
page 909, IEEE Std. 802.3, Ref. 3.)
RX DV signal with the RXD7 : 0:* value of 100 hexadecimal. On the

transmit path the absence of data activity is indicated by the de-assertion of
both TX EN transmit enable. and TX ER. Between frames within a burst,
the interframe period is signaled as carrier extend on the GMII.
9.8.3 Physical Coding Sublayer (PCS) and Physical Medium

Attachment (PMA) Sublayer, 1000BASE-X
9.8.3.1 Overview. There are currently three embodiments within this family: 1000BASE-CX, 1000BASE-LX, and 1000BASE-SX. 1000BASE-CX covers operation over a single copper medium: two pairs of 150- balanced
copper cabling. 1000BASE-LX covers operation over a pair of optical fibers
using long-wavelength optical transmission. 1000BASE-SX covers operation
over a pair of optical fibers using short-wavelength optical transmission. The
use of the term 1000BASE-X is to cover issues common to any of these
subvariants.
1000BASE-X is based on the physical layer standards developed by ANSI
X3.230-1994 fiber channel physical and signaling interface .. In particular,
this standard uses the same 8Br10B coding as fiber channel, a PMA sublayer
compatible with speed-enhanced versions of the ANSI 10-bit serializer chip,
and similar optical and electrical specifications.
1000BASE-X PCS and PMA sublayers map the interface characteristics of
the PMD layer including MDI. to the services expected by the reconciliation
sublayer. 1000BASE-X can be extended to support any other full-duplex
medium requiring only that the medium be compliant at the PMD level.
*RXD s a bundle of 8 data signals RXD7 : 0:..
278
9.8.3.2 Objectives. The following are the objectives of 1000BASE-X:

a.
b.
c.
d.
e.
To support the CSMArCD MAC

To support the 1000 Mbps repeater
To provide for autonegotiation among like 1000-Mbps PMDs
To provide 1000-Mbps data rate at the GMII
To support cable plants using 150- balanced copper cabling, or
optical fiber compliant with ISOrIEC 11801: 1995
f. To allow for a nominal network extent of up to 3 km, including
1. 150- balanced links of 25-m span
2. one-repeater networks of 50-m span using all 150- balanced
copper cabling.
3. one-repeater networks of 200-m span using fiber.
4. DTErDTE links of 3000 m using fiber.
g. To preserve full-duplex behavior of underlying PMD channels
h. To support a BER objective of 10y12
9.8.3.3 Summary of 1000BASE-X Sublayers*

Physical Coding Sublayer. The PCS interface is the gigabit media-independent interface GMII. that provides a uniform interface to the reconciliation
sublayer for all 1000-Mbps PHY implementations e.g., not only 1000BASE-X
but also other possible types of gigabit PHY entities .. 1000BASE-X provides
services to the GMII in a manner analogous to how 100BASE-X provides
services to the 100-Mbps MII.
The 1000BASE-X PCS provides all services required by the GMII, including
a. Encoding decoding. of GMII data octets to from. 10-bit code groups
8Br10B. for communication with the underlying PMA
b. Generating carrier sense and collision detect indications for use by
PHYs half-duplex clients
c. Managing the autonegotiation process, and informing the management
entity via the GMII when the PHY is ready for use
Physical Medium Attachment (PMA) Sublayer. The PMA provides a
medium-independent means for the PCS to support the use of a range of
serial-bit-oriented physical media. The 1000BASE-X PMA performs the
*1000BASE-X PHY consists of that portion of the physical layer between the MDI and GMII
consisting of the PCS, PMA, and PMD sublayers. 1000BASE-X PHY is roughly analogous to the
100BASE-X PHY.
279
following functions:
a. Mapping of transmit and receive code groups between the PCS and
PMA via the PMA service interface
b. Serialization deserialization . of code groups for transmission reception. on the underlying serial PMD
c. Recovery of clock from the 8Br10B-coded data supplied by the PMD
d. Mapping of transmit and receive bits between the PMA and PMD via
the PMD service interface
e. Data loopback at the PMD service interface
Physical Medium-Dependent (PMD Sublayer. 1000BASE-X physical layer
signaling for fiber and copper media is adapted from ANSI X.320-1994 FC
and PH., Clauses 6 and 7, respectively. These clauses define 1062.5-Mbps,
full-duplex signaling systems that accommodate single-mode optical fiber,
multimode optical fiber, and 150- balanced copper cabling. 1000BASE-X
adapts these basic physical layer specifications for use with the PMD sublayer
and mediums specified in Sections 38 and 39 of the reference document.
The MDI,* logically subsumed within each PMD subsection, is the actual
medium attachment, including connectors, for the various supported media.
Figure 9.46 depicts the relationship between 1000BASE-X and its associated PMD sublayers. Figure 9.47 is a functional block diagram of the
1000BASE-X PHY.
9.8.3.4 Physical Coding Sublayer (PCS). The PCS service interface allows the 1000BASE-X PCS to transfer information to and from a PCS client.
PCS clients include the MAC via the reconciliation sublayer. and repeater.
In the reference specification the GMII variables are set to true or false.
This is equivalent to, respectively, asserting or de-asserting them.
Functions Within the PCS. The PCS comprises the PCS transmit, carrier
sense, synchronization, PCS receive, and autonegotiation processes for
1000BASE-X. The PCS shields the reconciliation sublayer and MAC. from
the specific nature of the underlying channel. When communicating with the
GMII, the PCS uses an octet-wide, synchronous data path, with packet
delimiting being provided by separate transmit control signals TX EN and
TX ER. and receive control signals RX DV and RX ER.. When communicating with the PMA, the PCS uses a 10-bit-wide, synchronous data path,
which conveys 10-bit code groups. At the PMA service interface, code-group
alignment and MAC packet delimiting are made possible by embedding
special nondata code groups in the transmitted code-group stream. The PCS
provides the functions necessary to map packets between the GMII format
and the PMA service interface format.
*MDI, medium-dependent interface.
280
281
Figure 9.47. Functional block diagram of the 1000BASE-X PHY. (From Figure 36-2, page
926, IEEE Std. 802.3, Ref. 3.)
The PCS transmit process continuously generates code groups based upon
the TXD7 : 0:, TX EN, and TX ER signals on the GMII, sending them
immediately to the PMA service interface via the PMA UNITDATA.request
primitive. The PCS transmit process generates the GMII signal COL based
on whether a reception is occurring simultaneously with transmission. Additionally, it generates the internal flag, transmitting, for use by the carrier
sense process. The PCS transmit process monitors the autonegotiation process xmit flag to determine whether to transmit data or reconfigure the link.
The carrier sense process controls the GMII signal CRS see Figure 36-8
of the reference document..
282
The PCS synchronization process continuously accepts code groups via the
PMA UNITDATA.indicate primitive and conveys received code groups to
the PCS receive process via the SYNC UNITDATA.indicate primitive. The
PCS synchronization process sets the sync status flag to indicate whether the
PMA is functioning dependably as well as can be determined without
exhaustive error-rate analysis..
The PCS receive process continuously accepts code groups via the
SYNC UNITDATA.indicate primitive. The PCS receive process monitors
these code groups and generates RXD7 : 0:, RX DV, and RX ER on the
GMII and generates the internal flag, receiving, used by the carrier sense and
transmit processes.
The PCS autonegotiation process sets the xmit flag to inform the PCS
transmit process to either transmit normal idles interspersed with packets as
requested by the GMII or to reconfigure the link. The PCS autonegotiation
process is specified in Clause 37 of the reference document.
8Br10B Transmission Code. The PCS uses a transmission code to improve
the transmission characteristics of information to be transferred across the
link. The encodings defined by the transmission code ensure that sufficient
transitions are present in the PHY bit stream to make clock recovery possible
at the receiver. Such encoding also greatly increases the likelihood of
detecting any single or multiple bit errors that may occur during transmission
and reception of information. In addition, some of the special code groups of
the transmission code contain a distinct and easily recognizable bit pattern
that assists a receiver in achieving code-group alignment on the incoming
PHY bit stream. The 8Br10B transmission code specified for use in this
standard has a high transition density, is a run-length-limited code, and is
dc-balanced. The transition density of the 8Br10B symbols ranges from 3 to
8 transitions per symbol.
The definition of the 8Br10B transmission code is identical to that
specified in ANSI X3.230-1994, Clause 11. The relationship of code-group bit
positions to PMA and other PCS constructs is illustrated in Figure 9.48.
Notation Conventions. 8Br10B transmission code uses letter notation for
describing the bits of an unencoded information octet and a single control
variable. Each bit of the unencoded information octet contains either a
binary zero or a binary one. A control variable, Z, has either the value D or
the value K. When the control variable associated with an unencoded
information octet contains the value D, the associated encoded code group is
referred to as a data code group. When the control variable associated with
an unencoded information octet contains the value K, the associated encoded code group is referred to as a special code group.
The bit notation of A, B, C, D, E, F, G, H for an unencoded information
octet is used in the description of the 8Br10B transmission code. The bits
A, B, C, D, E, F, G, H are translated to bits a, b, c, d, e, f, g, h, j of 10-bit transmission code groups. 8Br10B code-group bit assignments are illustrated in
283
Figure 9.48. PCS reference diagram. (From Figure 36-3, page 928, IEEE Std. 802.3, Ref. 3.)
Figure 9.48. Each valid code group has been given a name using the following
convention: rDx. yr for the 256 valid data code groups and rKx. yr for
special control code groups, where x is the decimal value of bits EDCBA and
y is the decimal value of bits HGF.
Transmission Order. Code group bit transmission order is shown in Figure
9.48.
Code groups within multi-code-group ordered sets as specified in Table
36-3 of the reference document. are transmitted sequentially beginning with
the special code group used to distinguish the ordered set e.g., rK28.5r.
and proceeding code group by code group from left to right within the
definition of the ordered set until all code groups of the ordered set are
transmitted.
1000-Mbps Networks
9.8.4.1 Overview. The 1000-Mbps technology is designed to be deployed
in both homogeneous 1000-Mbps networks and 10r100r1000 Mbps mixed
networks using bridges andror routers. Network topologies can be developed
within a single 1000-Mbps collision domain, but maximum flexibility is
achieved by designing multiple-collision-domain networks that are joined by
bridges andror routers configured to provide a range of service levels to
284
Figure 9.49. Multiple data rate, multiple-collision-domain topology using multiport bridge.
(From Figure 42-2, page 1067, IEEE Std. 802.3, Ref. 3.)
DTEs. For example, a combined 1000BASE-Tr100BASE-Tr10BASE-T system built with repeaters and bridges can deliver dedicated or shared service
to DTEs at 1000 Mbps, 100 Mbps, or 10 Mbps.
Figure 9.41 outlines a 100-Mbps multiple-collision-domain topology using
a multiport bridge. This figure serves equally as well for a 1000-Mbps
multiple-collision-domain topology, and it shows how individual collision
domains can be linked by single devices or by multiple devices from any of
several transmission systems. Also see Figure 9.49.. The design of multiplecollision-domain networks is governed by the rules defining each of the
transmission systems incorporated into the design.
9.8.4.2 Single-Collision-Domain Multisegment Networks. The proper
operation of a CSMArCD network requires the physical size of the collision
domain to be limited in order to meet the round-trip propagation delay
requirements defined by the slot time discussed in Section 9.6.5.1. It also
requires that the number of repeaters be limited so as not to exceed the
interframe shrinkage given in Tables 9.3 through 9.6 and discussed in Section
9.6.7.2.
As in the previous section, there are two transmission models: Transmission System Model 1 and Model 2. Model 1 is the more conservative,
whereas Model 2 provides a set of calculation aids that allow those configuring a network to test a proposed configuration against a simple set of criteria
that allows it to be qualified. Transmission System Model 2 validates an
TABLE 9.14
Delays for Network Media SegmentsModel 1
Media Type
Category 5 UTP link segment (1000BASE-T)
Shielded jumper cable link segment (1000BASE-CX)
Optical fiber link segment (1000BASE-SX,
1000BASE-LX)
a
285
Maximum
Maximum
Number of Maximum Medium RoundPHYs per Segment Trip Delay per
Segment Length (m) Segment (BT)
2
2
2
100
25
316 a
1112
253
3192
May be limited by the maximum transmission distance of the link.
additional broad set of topologies that are fully functional and do not fit
within the simpler, but more restrictive rules of Model 1.
The physical size of a CSMArCD network is discussed in Section 9.7.5.1.
Table 9.14 summarizes the delays, measured in bit times BTs., for 1000-Mbps
media segments.
Repeater Usage. Repeaters are the means used to connect segments of a
network medium together into single collision domain. Different physical
signaling systems e.g., 1000BASE-CX, 1000BASE-SX, 1000BASE-LX,
1000BASE-T. can be joined into a common collision domain using a repeater. Bridges can also be used to connect different signaling systems.
However, if a bridge is so used, each LAN connected to the bridge will
comprise a separate collision domain.
9.8.4.3 Transmission System Model 1. The following network topology
constraints apply to networks using Transmission System Model 1:
a. Single repeater topology maximum
b. Link distances not to exceed the lesser of 316 m or the maximum
transmission distance of the link
9.8.4.4 Transmission System Model 2. Transmission System Model 2 is
a single repeater topology with the physical size limited primarily by round-trip
collision delay. A network configuration must be validated against collision
delay using a network model for a 1000-Mbps collision domain. The modeling process is quite straightforward and can easily be done either manually or
with a spreadsheet.
Key elements in the calculation are two: round-trip collision delay and
interpacket gap shrinkage. The validation for each of these must be done
separately. The model consists of a series of segments consisting of a left-end
segment, mid-segments, and a right-end segment. The model consists of the
worst-case path through the network for that particular calculation.
Round-Trip Collision Delay. For a network to be valid, it must be possible
for any two DTEs on that network to properly arbitrate for the network.
When two or more stations attempt to transmit within the slot time interval,
286
Figure 9.50. System model 2: single repeater.
each station must be notified of the contention by the returned collision

signal within the collision window.* Additionally, the maximum length
fragment created on a 1000-Mbps network must contain less than 512 bytes
after the start frame delimiter SFD.. These requirements limit the physical
diameter maximum distance between DTEs. of a network. The maximum
round-trip delay must be qualified between all pairs of DTEs in the network.
In practice, this means that the qualification must be done between those
that, by inspection of the topology, are candidates for the longest delay.
The following network modeling methodology is provided to assist that
calculation.
WORST-CASE PATH DELAY VALUE PDV. SELECTION. The worst-case path through
a network to be validated is identified by examination of the aggregate DTE
delays, cable delays, and repeater delay. The worst case consists of the path
between the two DTEs at opposite ends of the network that have the longest
round-trip time. Figure 9.50 shows a schematic representation of a one-repeater path.
WORST-CASE PDV CALCULATION. Once a set of paths is selected for calculation,
each is checked for validity against the following formula:
PDV s link delays LSDV . q repeater delay

q DTE delays q safety margin
Values for the formula variables are determined by the following method:
a. Determine the delay for each link segment link segment delay value,
or LSDV., using the formula
LSDV s 2 for round-trip delay . = segment length
= cable delay for this segment
*A given station can experience a collision during the initial part of its transmission collision
window. before its transmitted signal has had time to propagate to all stations on the
CSMArCD medium.
287
TABLE 9.15
Network Component Delays, Transmission System Model 2

Component
Round-Trip Delay in Bit

Times per Meter (BTrm)
Maximum Round-Trip
Delay in Bit Times (BT)
11.12
10.10
10.10
864
1112 (100 m)
253 (25 m)
1111 (110 m)
976
Two DTEs
Category 5 UTP cable segment
Shielded jumper cable segment
Optical fiber cable segment
Repeater
TABLE 9.16
Conversion Table for Cable Delays
Speed Relative to c
nsrm
BTrm
Speed Relative to c
nsrm
BTrm
0.4
0.5
0.51
0.52
0.53
0.54
0.55
0.56
0.57
0.58
0.5852
0.59
0.6
0.61
8.34
6.67
6.54
6.41
6.29
6.18
6.06
5.96
5.85
5.75
5.70
5.65
5.56
5.47
8.34
6.67
6.54
6.41
6.29
6.18
6.06
5.96
5.85
5.75
5.70
5.65
5.56
5.47
0.62
0.63
0.64
0.65
0.654
0.66
0.666
0.67
0.68
0.69
0.7
0.8
0.9
5.38
5.29
5.21
5.13
5.10
5.05
5.01
4.98
4.91
4.83
4.77
4.17
3.71
5.38
5.29
5.21
5.13
5.10
5.05
5.01
4.98
4.91
4.83
4.77
4.17
3.71
Source: Table 42-4, pages 1072 and 1073, Ref. 3.
Note 1: Length is the sum of the cable lengths between the PHY
interfaces at the repeater and PHY interfaces at the farthest DTE. All
measurements are in meters.
Note 2: Cable delay is the delay specified by the manufacturer or the
maximum value for the type of cable used as shown in Table 9.15. For
this calculation, cable delay must be specified in bit times per meter
BTrm.. Table 9.16 can be used to convert values specified relative to
the speed of light %c. or nanoseconds per meter nsrm..
Note 3: When actual cable lengths or propagation delays are not
known, use the Max delay in bit times as specified in Table 9.15 for
copper cables. Delays for fiber should be calculated, as the value
found in Table 9.15 will be too large for most applications.
Note 4: The value found in Table 9.15 for shielded jumper cable is the
maximum delay for cable with solid dielectric. Cables with foam
dielectric may have a significantly smaller delay.
b. Sum together the LSDVs for all segments in the path.
288
c. Determine the delay for the repeater. If model specific data is not
available from the manufacturer, enter the appropriate default value
from Table 9.15.
d. Use the DTE delay value shown in Table 9.15 unless the equipment
manufacturer defines a different value. If the manufacturers supplied
values are used, the DTE delays of both ends of the worst-case path
should be summed together.
e. Decide on appropriate safety margin0 to 40 bit timesfor the PDV
calculation. Safety margin is used to provide additional margin to
accommodate unanticipated delay elements, such as extra-long connecting cable runs between wall jacks and DTEs. A safety margin of
32 BT is recommended..
f. Insert the values obtained through these calculations into the following formula to calculate the PDV. Some configurations may not use
all the elements of the formula..
PDV s link delays LSDV . q repeater delay
q DTE delay q safety margin
g. If the PDV is less than 4096, the path is qualified in terms of
worst-case delay.
h. Late collisions andror CRC errors may be indications that path delays
exceed 4096 BT.
REFERENCES
1. The IEEE Standard Dictionary of Electrical and Electronic Terms, 6th ed., IEEE Std.
100-1996, IEEE, New York, 1996.
2. Information Processing SystemsLocal Area Networks, Part 2, Logical Link Control,
ANSIrIEEE Std. 802.2, 1998 edition, IEEE, New York, 1998.
3. Information TechnologyTelecommunication and Information Exchange Between
SystemsLocal and Metropolitan Area NetworksSpecific Requirements, Part 3:
Carrier Sense Multiple Access with Collision Detection (CSMA r CD) Access Method
and Physical Layer Specification, IEEE Std. 802.3, 1998 edition, IEEE, New York,
1998.
4. Information TechnologyTelecommunications and Information Exchange Between
SystemsLocal and Metropolitan Area NetworksSpecific Requirements, Part 6:
Distributed Queue Dual Bus (DQDB) Access Method and Physical Layer Specifications, ANSIrIEEE 802.6, 1994 edition, IEEE, New York, 1994.
10
DATA COMMUNICATIONS IN THE
OFFICE ENVIRONMENT, PART 2
10.1
MEDIUM ACCESS CONTROLTOKEN-PASSING SCHEMES
10.1.1
Introduction
Whereas CSMArCD Section 9.5. was a random accessrcontention method

of local area network LAN. access, token-passing LAN methods are disciplined access techniques. The following lists the four major points of the
token-passing method:
1. A token controls the right of access to the physical medium. The LAN
station that holds possesses . the token has control over the medium for
a short time period.
2. The token is passed by LAN stations residing on the medium. As the
token is passed from station to station, a logical ring is formed.
3. Steady-state operation consists of a data transfer phase and a tokenpassing phase.
4. Ring maintenance functions within the LAN stations provide for ring
initialization, lost token recovery, new station addition to the logical
ring, and general housekeeping of the logical ring. The ring maintenance functions are replicated among all the token-using LAN stations
on the network.
Shared media generally can be categorized into two major types: broadcast
and sequential. The token-bus overview that follows deals with the broadcast
type. On a broadcast medium, every station may receive signals transmitted.
Media of the broadcast type are usually configured as a physical bus.
This section covers three token-passing methods:
Token bus
Token ring
Fiber distributed data interface FDDI.
289
290
Figure 10.1. Logical ring on a physical bus.
The token-bus method is covered by only a brief overview because its market
penetration has dropped to just several percent of the total LAN market.*
Token ring and FDDI are discussed in greater detail. The IEEE 802.4 i.e.,
token-bus. committee has a current specification available. Token ring was
backed by IBM and is covered by the IEEE 802.5 committee. FDDI is an
ANSI initiative.
10.1.2
Token-Passing Bus
Figure 1.1 shows a simple example of a bus network. Figure 10.1 illustrates a
logical ring on a physical bus. It should be noted that the token medium
access method is always sequential in a logical sense. That is, during normal
steady-state operation, the right to access the medium passes from station to
station in a sequential manner i.e., station 1 to station 2, to station
3 . . . to station n in the logical ring.. In addition, the physical connectivity
has little impact on the order of the logical ring, and stations can respond to
the token holder even without being part of the logical ring. For example,
Stations H and F in Figure 10.1 can receive frames and could respond but
cannot initiate a transmission because they will never be sent the token.
The MAC sublayer provides sequential access to the shared bus medium
by passing control of the medium from station to station in a logically circular
fashion. The MAC sublayer determines when the LAN station has the right
to access the shared medium by recognizing and accepting the token from the
predecessor station, and it determines when the token will be passed to the
successor station.
The MAC sublayer carries out the following general functions:
1.
2.
3.
4.
5.
Lost token timer

Distributed initialization
Token-holding timer
Limited data buffering
Node address recognition
*In contrast, Ethernet Section 9.5. has over 75% market share year 2000..
MEDIUM ACCESS CONTROL TOKEN-PASSING SCHEMES
291
Figure 10.2. Token-passing bus architecture relates to OSI layers.
6. Frame encapsulation including token preparation .

7. Frame check sequences FCS. generation and checking
Figure 10.2 relates the token bus architecture to the OSI reference model.
Token-passing bus is a 1-, 5-, 10-, or 20-Mbps system.
10.1.2.1 Token-Bus Frame Formats. All frames sent or received by the
MAC sublayer conform to the format shown in Figure 10.3. In a MAC frame
the number of octets between SD and ED is 8191 or less. The components of
a frame are detailed in the legend under the figure.
Figure 10.4 shows the format of a token. Here the destination address
DA. has the value of the next station in sequence following the station
issuing the token. The frame control FC. field has the unique binary
sequence: 00001000.
Figure 10.3. MAC sublayer general frame format.
Figure 10.4. The token format.
292
10.1.2.2 Operational Aspects

SLOT TIME.
Slot time is the maximum time any station need wait for an
immediate MAC level response from another station. Slot time is measured
in octet times and is defined as
slot time s INTEGER 2)transission path delay q station delay
qsafety margin . rMAC-symbol time q 7 4 r8 .
The slot time, along with the stations address and several other management parameters, is known to the station before it attempts to transmit on
the network. If all stations on the network are not using the same value for
slot time, the MAC protocol may not operate properly.
RIGHT TO TRANSMIT.
The token right to transmit. is passed from station to

station in descending order of station address. When a station which belongs
to the logical ring hears a token frame addressed to itself, it has the token
and may transmit data frames. When a station has completed transmitting
data frames, it passes the token to the next station in the logical ring.
When a LAN station has the token, it may temporarily delegate to another
station its right to transmit by sending a request with response data frame.
When a station hears a request with response data frame addressed to
itself, it responds with a response data frame, if the request with response
option is implemented. The response data frame causes the right to transmit
to revert back to the station which sent the request with response data
frame.
TOKEN PASSING. After each station has completed transmitting any data
frames it may have and has completed other maintenance functions, the
station passes the token to its successor by sending a token MAC control
frame.
After sending the token frame, the station listens for evidence that its
successor has heard the token frame and is active. If the sender hears a valid
frame following the token within one slot time, it assumes that its successor
has the token and is transmitting. Otherwise the token-sending station
attempts to assess the state of the network.
If the token-sending station hears a noise burst e.g., an unidentifiable
sequence or a frame with an incorrect FCS., it cannot be sure which station
sent the transmission. The MAC protocol treats this condition in a way which
minimizes the chance of the station causing a serious error.
Because a station on a broadband network should always hear its own
frames, if the token-sending station hears a single noise burst without
hearing its own token frame, the station assumes that it heard its own token
293
that had been garbled and continues to listen. If a second noise burst is
heard or if a noise burst is heard after the station hears its own token frame,
the token-sending station continues to listen in the CHECK TOKEN PASS
state for up to four more slot times. If nothing more is heard, the station
assumes that the noise burst it heard was not a garbled frame from the
successor station and so it repeats the token transmission. If anything is
heard during the four-slot time delay, the station assumes that its successor
successfully received the token.
RESPONSE WINDOW. New stations are added to the logical ring through a
controlled contention process using response windows. A response window
is a controlled interval of time equal to one slot time. after transmission of
a MAC control frame in which the station sending the frame pauses and
listens for a response. If a station hears a transmission start during the
response window, the station continues to listen to the transmission, even
after the response window time expires, until the transmission is complete.
Thus, the response windows define the time interval during which a station
will hear the beginning of a response from another station Ref. 1..
10.1.3
Token-Passing Ring
10.1.3.1 Overview. A token ring consists of a set of LAN stations serially

connected by a transmission medium as shown in Figure 10.5, with the last
station and first station folded back connecting one to the other. Thus a ring
is formed. Information is transferred sequentially bit by bit from one LAN
station to the next around the ring. Each LAN station regenerates and
repeats each bit and serves as the means of attaching one or more data
devices terminals, workstations, computers, print servers, etc.. to the ring for
Figure 10.5. Typical token-passing ring configuration.
294
the purpose of communicating with other devices on the network. A given

LAN station has access to the medium for a given time token-holding time..
During this time a station may transfer information onto the ring, where the
information circulates from one station to the next. The addressed destination stations. copies the information as it passes. Finally, the station that
transmitted the information effectively removes the information from the
ring as it is returned around to that station again.
A LAN station gains the right to transmit its information onto the medium
when it detects a token passing on the medium. The token is a control signal
comprising a unique binary sequence that circulates on the medium following
each information transfer. Any LAN station detecting the token may capture
the token by modifying it to a start-of-frame sequence SFS. and by appending the appropriate control and status fields, address fields, information field,
frame check sequence FCS., and end-of-frame sequence EFS.. On completion of its information transfer and after appropriate checking for proper
operation, the LAN station initiates a new token. This provides other stations
the opportunity to gain access to the ring.
A LAN station may only hold the token for a maximum amount of time
before it must pass the token onwards.
There are multiple levels of priority available for independent and dynamic assignment depending on the relative class of service required for any
given message, such as synchronous real-time voice., asynchronous interactive., or immediate network recovery.. The allocation of priorities is by
mutual agreement among users of the network.
Error detection and recovery mechanisms are provided to restore network
operation in the event that transmission errors or medium transients cause
the access method to deviate from normal operation. Detection and recovery
for these cases utilize a network monitoring function that is performed in a
specific station with backup capability in all other stations that are attached
to the ring.
Figure 10.6 shows a token ring with server stations servers. that are the
means through which the system manager manages the stations in such a
system. Servers are data collection and distribution points on each ring where
reports from data stations are gathered. Servers then communicate the
necessary information to the system manager for the purpose of managing a
LAN token ring system.
Data stations communicate with servers by reporting errors that are
detected, such as lost token, FCS error, or lost frames, and requesting
operating parameters when inserting into the ring, reporting changes in
configuration due to insertion or removal of stations, responding to requests
for various status information, and removal from the ring when requested.
Figure 10.7 shows the token rings relationship with the open system
interconnection OSI. reference model. Essentially the LLC and MAC sublayers relate to OSI layer 2, the data-link layer, and PHY the physical layer.
relates to OSI layer 1.
295
Figure 10.6. Relationship between data stations, servers, and system manager.
Figure 10.7. Relating the OSI reference model to the LAN model.
10.1.3.2 Token-Ring Frame Formats. Figure 10.8 shows the token format, and Figure 10.9 illustrates the frame format. The frame format is used
for transmitting both MAC and LLC messages to the destination stations.. A
frame may or may not have an information field.
Abort Sequence. The abort sequence consists of two fields: the starting
delimiter 1 octet. and the ending delimiter 1 octet.. This sequence is not
used for the purpose of terminating the transmission of a frame prematurely.
Figure 10.8. Token format.
296
Figure 10.9. Token-ring frame format. *Maximum length only limited by the token-holding
time.
The abort sequence may occur anywhere in the bit stream; that is, receiving
stations are able to detect an abort sequence even if it does not occur on
octet boundaries.
Fill. When a LAN station is transmitting as opposed to repeating., it
transmits fill preceding or following frames, tokens, or abort sequences to
avoid what would otherwise be an inactive or indeterminate transmitter state.
Fill may be either 1s or 0s or any combination thereof and may be any
number of bits in length.
Starting Delimiter (SD). A frame or token is started with these eight symbols.
If otherwise, the sequence is not considered valid.
J K 0 J K 0 0 0
where
J s nondata J
K s nondata K
0 s zero bit
Access Control (AC). The access control sequence is one octet and consists
of the following bits:
PPP T M RRR
where
PPP s priority bits
T s token bit
M s monitor bit
RRR s reservation bits
297
PRIORITY BITS.
Priority bits indicate the priority of a token and, therefore,

which stations are allowed to use the token. In a multiple priority system,
stations use different priorities depending on the priority of the PDU
protocol data unit. to be transmitted.
There are eight levels of priority increasing from the lowest 000. to the
highest 111. priority. For the purposes of comparing priority values, the
priority is transmitted most significant bit first. For example, 110 has higher
priority than 011, with leftmost bit transmitted first.
TOKEN BIT.
The token bit is a 0 in a token and 1 in a frame. When a station

with a PDU to transmit detects a token that has a priority equal to or less
than the PDU to be transmitted, it may change the token to a start-of-frame
sequence SFS. and transmit the PDU.
MONITOR M. BIT.
The M bit is used to prevent a token that has a priority

greater than 0 or any frame from continuously circulating on the ring. If an
active monitor detects a frame or a high-priority token with the M bit equal
to 1, the frame or token is aborted. The M bit is transmitted as 0 in all frames
and tokens. The active monitor inspects and modifies this bit. All other LAN
stations repeat this bit as received.
RESERVATION BITS. These allow stations with high-priority PDUs to request
in frames or tokens as they are repeated . that the next token be issued at
the requested priority. The eight levels of reservation increase from 000 to
111, and these are operated upon in a similar fashion as the priority bits.
Frame Control (FC). The FC field defines the type of the frame and certain
MAC and information frame functions. It consists of one octet as follows:
FF ZZZZZZ
where
FF s frame-type bits
ZZZZZZ s control bits
Frame-type bits indicate the type of frame as follows:
00 s MAC frame i.e., contains a MAC PDU .
01 s LLC frame i.e., contains an LLC PDU .
1X s undefined format reserved for future use .
MAC FRAMES. If the frame-type bits indicate a MAC frame, all stations on the
ring shall interpret and, based on the finite state of the system, act on the
ZZZZZZ control bits.
298
LLC FRAMES. If the frame-type bits indicate an LLC frame, the ZZZZZZ bits
are designated as rrrYYY. The rrr bits are reserved and are transmitted as 0s
in all designated frames and ignored upon reception. The YYY bits may be
used to carry the priority Pm. of the PDU from the source LLC entity to the
target LLC entity or entities. Note that P, the priority of the AC field of a
frame, is less than or equal to Pm when the frame is transmitted onto
the ring.
UNDEFINED FORMAT.
The value of 1x is reserved for frame types that may

be defined in the future. However, although currently undefined, any future
frame formats will adhere to the following conditions:
1. The format is delimited by the 2-octet SFS start-of-frame sequence.
field and the 2-octet EFS end-of-frame sequence. field, as they are
defined above. Additional fields may follow the EFS field.
2. The position of the FC frame control. field is unchanged.
3. The SFS and EFS of the format are separated by an integral number of
octets. This number is at least 1 i.e., the FC field., and the maximum
length is subject to the constraints of the token-holding timer THT..
4. All symbols between the SFS and EFS are 0 and 1 bits.
5. All stations on the ring are to check for data symbols and an integral
number of octets between the SFS and EFS fields. The error-detected
E. bit of formats that are repeated shall be set to 1 when a nondata
symbol or nonintegral number of octets is detected between the SFS
and EFS fields.
6. All bit errors having a Hamming distance* of less than four must be
detectable by stations using the format and shall not be accepted by any
other stations conforming to this standard.
Destination and Source Address Fields. Each frame contains two address
fields: the destination LAN station. address and the source LAN station.
address, in that order. Addresses may be either 2 or 6 octets in length.
However, all stations of a specific LAN will have addresses of uniform length.
Destination address DA. field identifies the LAN stations. for which the
information field of the frame is intended. Included in the DA is a bit to
indicate whether the DA is an individual or group address; and, for 6-octet
addresses only, the second bit indicates whether it is a universally or locally
administered address.
*Hamming distance is described in Section 4.5.
299
INDIVIDUAL AND GROUP ADDRESSES. The first bit transmitted of the DA

distinguishes individual from group addresses as follows:
0 s individual address
1 s group address
Individual addresses identify a particular station on the LAN and are distinct
from all other individual station addresses on the same LAN in the case of
local administration ., or from individual addresses of other LAN stations on
a global basis in the case of universal administration ..
A group address is used to address a frame to multiple destination
stations. Group addresses may be associated with zero or more stations on a
given LAN. In general, a group address is an address associated by convention with a group of logically related stations.
BROADCAST ADDRESS.
The group address consisting of 16 or 48 1s for 2- or

6-octet addressing, respectively . constitutes a broadcast address, denoting the
set of stations on a given LAN. Stations using 48-bit addressing must also
recognize XC000FFFFFFFF as a broadcast address in MAC frames.
NULL ADDRESS.
An address of 16 or 48 0s for 2- or 6-octet addressing,

.
respectively is considered a null address. It means that a frame is not
addressed to any particular station.
Source address SA. field identifies the station originating the frame and
has the same format and length as the DA in a given frame. The
individualrgroup bit is 0.
Information (INFO) Field. The INFO field contains zero, one, or more octets
that are intended for MAC, SMT, or LLC. Although there is no maximum
length specified for the information field, the time required to transmit a
frame may be no greater than the token-holding time that has been established for that LAN station. The format of the INFO field is indicated in the
frame-type bits of the FC field. There are two frame types: MAC and LLC.
MAC FRAME FORMAT. Figure 10.10 shows the format of the information field,
when present, for MAC frames.
The vector is the fundamental unit of MAC and SMT information. A
vector contains its length, an identifier of its function, and zero or more
subvectors. Only one vector is permitted per MAC frame.
LLC FRAME FORMAT. This is not specified in the IEEE 802.5 standard Ref.
2.. However, all LAN stations should be capable of receiving frames with
information fields up to and including 133 octets in length.
300
Figure 10.10. MAC frame information field structure.
FRAME CHECK SEQUENCE FCS.. The FCS is a 32-bit sequence based on the
following standard generating polynomial of degree 32:
G X . s X 32 q x 26 q X 23 q X 22 q X 16 q X 12 q X 11 q X 10 q X 8 q X 7
q X5 q X4 q X2 q X q 1
The FCS is the 1s complement of the sum modulo 2. of the following: a. the
remainder of X k = X 31 q X 30 q X 29 q qX 2 q X q 1. divided modulo
2. by G X ., where k is the number of bits in the FC, DA, SA, and INFO
fields, and b. the remainder after multiplication by X 32 and then division
modulo 2. by G X . of the content treated as a polynomial. of the FC, DA,
SA, and INFO fields. The FCS is transmitted commencing with the coefficient of the highest term.
Ending Delimiter (ED)
J K 1 J K 1 I E
where
J s nondata J
K s nondata K
1 s binary 1
I s intermediate frame bit
E s error-detected bit
The transmitting station transmits the delimiter as shown. Receiving stations
consider the ED valid if the first six symbols J K 1 J K 1 are received
correctly.
INTERMEDIATE FRAME BIT I BIT.. If the I flag is used to determine the end of
a stations transmission, the I bit is transmitted as a 1 in intermediate or
first. frames of a multiple-frame transmission. The I bit in the last or only
frame of the transmission is sent as a 0.
301
ERROR-DETECTED BIT E BIT..
The E bit is transmitted as 0 by the station that

originates the token, abort sequence or frame. All stations on the ring, as
they repeat frames, check tokens and frames for errors. The E bit of tokens
and frames that are repeated are set to 1 when a frame with error is
detected; otherwise the E bit is repeated as received.
Frame Status (FS). The frame status FS. field consists of one octet as
follows:
A C r r A C r r
where
A s address-recognized bits
C s frame-copied bits
r s reserved bits
The reserved bits are reserved for future standardization. They are transmitted as 0s; however, their value is ignored by receivers.
ADDRESS-RECOGNIZED A. BITS AND FRAME-COPIED C. BITS. The A and C bits

are transmitted as 0 by the LAN station originating the frame. If another
station recognizes the DA as its own address or relevant group address, it
sets the A bits to 1. If it copies the frame into its receiver buffer, it also sets
the C bits to 1. This allows the originating station to differentiate among
three conditions:
1. Station nonexisting or nonactive on the ring

2. Station exists but frame not copied
3. Frame copied
Notes. After transmission of frames. has been completed, the station
checks to see if the stations address has returned in the SA field, as
indicated by the MA FLAG. If it has not been seen, the station transmits fill
until the MA FLAG is set, at which time the station transmits a token.
After transmission of the token, the station will remain in the transmit
state until all the frames that the station originated are removed from the
ring. This is called stripping and is done to avoid unnecessary recovery action
that would be caused if a frame were allowed to continuously circulate on
the ring.
10.1.3.3 Token-Ring Symbol Transmission Characteristics
PHY Serial Stream of Symbols. Each symbol is one of the following:
0 s binary zero
1 s binary one
J s nondata J
K s nondata K
302
Figure 10.11. Symbol encoding using differential Manchester coding.
The symbols are transmitted to the medium in the form of differential

Manchester encoding that is characterized by the transmission of two signal
elements per symbol as shown in Figure 10.11.
In the case of two data symbols, binary 1 and binary 0, a signal element of
one polarity is transmitted for one-half the duration of the symbol to be
transmitted, followed by the contiguous transmission of a signal element of
the opposite polarity for the remainder of the symbol duration. This provides
two distinct advantages:
1. The resulting signal has no DC component and can readily be inductively or capacitively coupled.
2. The forced mid-bit transition conveys inherent timing information on
the channel.
In the case of differential Manchester coding, the sequence of signal
element polarities is completely dependent on the polarity of the trailing
signal element of the previously transmitted data or nondata bit symbol. If
the symbol to be transmitted is a binary zero, the polarity of the leading
signal element of the sequence is opposite to that of the trailing element of
the previous symbol, and, consequently, a transition occurs at the bit symbol
boundary as well as at mid-bit. If the symbol to be transmitted is a binary
303
one, the algorithm is reversed and the polarity of the leading signal element
is the same as that of the trailing signal element of the previous bit. Here
there is no transition at the bit symbol boundary.
The nondata symbols J and K depart from this rule in that a signal
element of the same polarity is transmitted for both signal elements of the
symbol and there is, therefore, no mid-bit transition. A J symbol has the same
polarity as the trailing element of the preceding symbol. The transmission of
only one nondata symbol introduces a DC component, and thus nondata
symbols are normally transmitted as a pair of J or K symbols. By its nature, a
K symbol is opposite polarity of the preceding symbol..
Data Signaling Rates. The data signaling rates should be within "0.01% of
4 and 16 Mbps.
Symbol Timing. The PHY recovers the symbol timing information inherent
in the transitions between levels of the received signal. It minimizes the
phase jitter in this recovered timing signal to provide suitable timing at the
data signaling rate for internal use and for the transmission of symbols on
the ring. The rate at which symbols are transmitted is adjusted continuously
in order to remain in phase with the received signal.
In normal operation there is one station on the ring that is the active
monitor. All other stations on the ring are frequency- and phase-locked to
this station. They extract timing from received data by means of a phaselocked loop design based on the requirement to accommodate a combined
total of at least 250 stations and repeaters on the ring.
10.1.3.4 Symbol Decoder. Received symbols are decoded using an algorithm that is the inverse of that used for symbol encoding Section 10.1.3.3..
The station monitors the received data for the signal element pattern
corresponding to a starting delimiter SD. and ending delimiter ED. as
Starting delimiters SD. and ending delimiters ED. shall be detected even
if they do not occur on a symbol boundary. If an ending delimiter is detected
that does not occur on the symbol boundary established by the previous SD,
then an ED alignment error is indicated to the MAC the MAC uses this
signal to verify that the ending delimiter falls exactly on an octet boundary..
Once a delimiter has been detected, it shall establish the symbol and octet
boundary for the decoding of the received signal elements until another
delimiter is detected.
10.1.3.5 Latency. Latency is the time, expressed in number of symbols, it
takes for a signal element to pass through a ring component. Two latency
buffers are specified: a fixed buffer and an elastic buffer, which are inserted
in the ring signal path as requested by the MAC protocol. To ensure that
stations do not introduce excessive latency, the specification recommends
304
Figure 10.12. Starting and ending delimiter patterns. (From Figure 28, page 110, ANSIrIEEE
Std. 802.5, Ref. 2.)
that the average PHY latency should not exceed 6.25 ms to prevent the
expiration of the timer, return to repeat TRR ., which is set between 4.0 and
4.1 ms. It is further recommended that this time should not exceed 2.5 ms for
performanceraccess reasons.
Ring Latency. Ring latency, defined as the time it takes for a signal element
to travel around the entire ring, is equal to the cumulative latency of the ring
plus the latency of the active monitor. Cumulative latency is the time it takes
for a signal element to travel from the active monitors transmitter output to
its receiver input. The latency of the active monitor consists of two buffers,
elastic buffer and fixed latency buffer, in addition to the normal station
latency. The active monitor station uses the elastic buffer to compensate for
variations in cumulative latency and uses the fixed latency buffer to provide
the required latency to ensure ring operation even when only one station is
on the ring.
When the ring is in a normal operating state, the MAC protocol ensures
that only one station, the active monitor, will have Crystal transmit asserted,
which inserts the fixed latency and elastic latency buffers in the ring path.
The active monitor transmits data to the ring timed from its local clock,
which provides the master timing for the ring. Although the mean data
signaling rate around the ring is established by the active monitor, segments
of the ring can, for short periods of time, operate at a slightly higher or lower
305
signaling rate than the signaling rate of the active monitor. The cumulative
effect of these signaling rate variations may cause variations in the cumulative latency.
The elastic buffer in the active monitor compensates for the variation in
the cumulative latency as follows. When the frequency of the received signal
at the active monitor is slightly faster than the crystal clock, the elastic buffer
will expand, as required, to maintain a constant ring latency. Conversely, if
the frequency of the received signal is slightly slower, the elastic buffer will
contract to maintain a constant ring latency. Constant ring latency is a
requirement to avoid adding or dropping signal elements from the data
stream. The elastic buffer compensates for dynamic variations in latency due
to jitter. The elastic buffer does not compensate for some changes such as
step changes in latency caused by stations becoming or ceasing to be stacking
stations.
The fixed latency buffer is provided by the active monitor to provide the
latency required for token circulation.
In order for the token to continuously circulate around the ring, the ring
must have a minimum ring latency of at least 24 symbols the number of
symbols in a token.. To ensure this minimum ring latency, a fixed latency
buffer of at least 24 symbols is inserted into the data path of the station when
MAC asserts Crystal transmit.
Cumulative latency variations are handled by the elastic buffer. These
variations are due to jitter. The elastic buffer accommodates without error a
minimum cumulative latency variation of positive or negative B symbols
where
Data rate
s 4
16
Mbps
s 3
15
symbols
The elastic buffer in the active monitor is initialized to the center of the
elastic latency range whenever the MAC indicates Token received station is
repeating token..
10.1.3.6 Station Organization and Basic PMC Requirements. Figure
10.13 shows station functional organization and data flow. The PMC physical medium components. feeds the TCU trunk coupling unit. concentrator.
Two transmission parameters are specified for both active and passive
channels. These are attenuation and differential impedance. Figure 10.14
gives an example of station-to-TCU connection.
The channel attenuation of a passive channel consists of the loss of all
channel components between a ring station medium interface connector
MIC. and the MIC of the next downstream station. The channel attenuation
of an active channel consists of the loss between a station MIC and its
respective active concentrator port MIC, or the loss between an active RO
306
Figure 10.13. Station functional organization and data flow. (From Figure 2, page 16,
ANSIrIEEE Std. 802.5, Ref. 2.)
307
Figure 10.14. Example of station-to-TCU connection. (From Figure 29, page 145, ANSIrIEEE
Std. 802.5, Ref. 2.)
ring out. port MIC of one concentrator and the active RI ring in. port MIC
of a second concentrator. Table 10.1 gives the maximum attenuation values.
These loss values are based on the channel terminated in 100 ohms for UTP
cable and 150 ohms for STP cable.
The magnitude of the channel differential input and output impedance for
both passive and active channels using STP and UTP media will meet the
specifications given in Table 10.2.
Error Performance. The token-ring transmission system provides a communication channel that consists of a transmitter, receiver and channel with an
equivalent BER of less than or equal to 1 = 10y9 with a ring equivalent BER
of 1 = 10y8.
Section 10.1.3 is based on ANSIrIEEE Standard 802.5, 2nd ed., December 29, 1995, Ref. 2.
308
TABLE 10.1
Channel Attenuation Specifications

Channel Attenuation
Channel Type
Only 4 Mbps
16 r 4 Mbps
Passive
Active
- 19 dB @ 4 MHz
- 19 dB @ 4 MHz
- 19 dB @ 16 MHz
- 16 dB @ 16 MHz
Source: Table 28, page 160, Ref. 2.
TABLE 10.2
Channel Differential Impedance Specifications

Frequency Range
Media Type
Differential
Impedance
Only 4 Mbps
16r4 Mbps
UTP
STP
100 " 15
150 " 15
112 MHz
112 MHz
125 MHz a
125 MHz
Category 3 is only specified to 16 MHz but it is assumed that it can be extrapolated to 25 MHz.
Source: Table 29, page 160, Ref. 2.
10.1.4
Fiber Distributed Data Interface (FDDI)
10.1.4.1 Introduction and Overview. Fiber distributed data interface

FDDI. is a LAN protocol that uses fiber-optic cable or other transmission
medium. Its peak data rate is 100 Mbps and a sustained data rate of 80
Mbps. It uses a 4-out-of-5 code on the medium such that the line modulation
rate is 125 Mbaud. FDDI can support 500 stations linked by up to 100 km of
dual fiber cable. Larger networks can be supported by increasing timer
values.
The FDDI operates in a token-ring format and is an outgrowth of IEEE
802.5 standard Section 10.1.3.. Table 10.3 compares FDDI with IEEE 802.5
standard.
FDDI consists of a set of nodes connected by a counterrotating ring
topology, which is illustrated in Figure 10.15. There are two classes of nodes
TABLE 10.3
Token-Ring Comparison
FDDI
Half-duplex architecture and symbol

(or byte)-level manipulation
Token sent immediately behind
packet
Traffic regulated through timed token
Uses 4-out-of-5 group coding, up to
10% DC component
Ring is decentralized; individual clocks
limit packet size
Fiber-optic medium (wire-pair option)
Token Ring
Full-duplex architecture and bit-level
manipulation
Token sent only after source address
has returned
Traffic regulated through priority and
reservation bits on each packet
Uses differential Manchester coding,
no DC component
Centralized control with active monitor
clock, allowing very long packets
Wire-pair medium
309
Figure 10.15. Typical FDDI topology.
or stations: single attachment and dual attachment. FDDI trunk rings are
composed of dual attachment stations and have two physical-layer mediumdependent PMD. entities. Single attachment stations are connected to the
trunk ring via concentrators.
Two kinds of data service can be provided in a logical ring: packet service
and circuit-connected service. The more common application today is the
packet service, where a given station or node. transmits information on to
the ring as a series of data packets or frames, where each packet circulates
from one station to the next. Each station copies the packet or frame header
as the packet passes. The addressed stations. passes. the packet internally
to the SMT station management. or LLC. Finally, the station that transmitted the packets effectively removes them from the ring. For circuit service,
some of the logical bit rate capacity is allocated to independent channels.
Two or more stations can simultaneously communicate via each channel.
FDDI provides packet service via a token ring in a similar manner as
described in Section 10.1.3. With FDDI a station gains the right to transmit
frames or packets on to the medium when it detects a token passing on the
medium. A token is a unique sequence consisting of four fields, which is
illustrated in Figure 10.16. The token follows each series of transmitted
packets. Any station, upon detection of a token, may capture the token by
removing it from the ring. That station may then transmit one or more data
packets that it has in queue.
310
Figure 10.16. The FDDI token format.
The time a station may hold a token is controlled by the token-holding

timer. At the expiration of its preset time, the station must release the token
and return it to the ring. This permits other stations to gain access to
the ring.
Multiple levels of priority are available for independent and dynamic
assignment depending upon the relative class of service required. The classes
of service may be synchronous, which is typically used for such applications as
real-time packet voice. They may be asynchronous such as for interactive
applications, or they may be immediate, which is used for extraordinary
applications such as ring recovery. The allocation of ring bit rate capacity
occurs by mutual agreement among users of the ring.
Error detection and recovery mechanisms are provided to restore ring
operation in the event that transmission errors or medium transients such as
those resulting from station insertion or removal. cause the access method to
deviate from normal operation. Detection and recovery for these cases utilize
a recovery function that is distributed among the stations attached to
the ring.
In the hybrid mode both token-ring operation and isochronous data
transfer are multiplexed on the same medium. Hybrid operation requires
the existence of a hybrid ring control HRC. entity. HRC passes all packet
data to the packet MAC entity. HRC does not pass isochronous data to the
packet MAC.
10.1.4.2 Ring Transmission. The FDDI MAC transmits symbols. The
symbol is the smallest signal element used by the MAC. A symbol provides
three types of information:
1. Line states, such as the halt line state or the idle line state.
2. Data quartets, where each quartet represents four ordered data bits.
3. Control sequences, such as the starting delimiter SD., ending delimiter
ED., or control indicator sequences.
Table 10.4 provides the FDDI symbol set and line coding. Each data quartet
shown in the table conveys a hexadecimal digit 0F.. There are seven 5-bit
invalid code groups, which are not transmitted by the MAC because they can
TABLE 10.4
311
FDDI Code Symbol Set
Code Group
Decimal
Symbol
Binary
Name
Assignment
00000
00100
11111
Q
H
I
Quiet
Halt
Idle
11000
10001
J
K
First symbol of JK pair

Second symbol of JK pair
00101
Line State Symbols

00
04
31
Starting Delimiter
24
17
Embedded Delimiter
05
Data Quartets
30
09
20
21
10
11
14
15
18
19
22
23
26
27
28
29
Second symbol of IL pair

Hexadecimal
Binary
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
11110
01001
10100
10101
01010
01011
01110
01111
10010
10011
10110
10111
11010
11011
11100
11101
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
01101
Terminate
00111
11001
R
S
Reset (logical Zero or Off)

Set (logical One or On)
Ending Delimiter
13
Control Indicators
07
25
Note the 4Br5B code.
Source: Extracted from Ref. 3. X3.231-1994.
generate patterns that can violate run length or duty cycle requirements. The
invalid code points are 01, 02, 03, 06, 08, 12, and 16.
In FDDI, information is signaled as a stream of fixed-length code groups,
each containing 5 bits see Table 10.4.. Each valid code group represents a
symbol. The first level code performed by PHY is the conversion of symbols
from the data-link layer MAC or HRC. to code groups. Each symbol is
mapped to its corresponding code group and consists of a sequence of five
NRZ code bits. The second level of coding performed by the PHY is the
312
Figure 10.17. FDDI frame format.
conversion of NRZ code bits to NRZI* code bits. Each NRZ code bit 1
generates a transition in polarity of the NRZI signal, while an NRZ code bit
0 maintains the previous NRZI signal polarity. For the incoming pulse
stream, NRZI code bits are first decoded to equivalent NRZ code bits and
then decoded to symbols for delivery to the data-link layer DLL..
10.1.4.3 Protocol Data Units. Two protocol data unit PDU. formats are
used by the FDDI MAC: tokens and frames. Figure 10.16 illustrates the
token and Figure 10.17 shows the FDDI frame. The token is the means by
which a station gains the right to transmit MAC SDUs as opposed to the
normal process of repeating .. The token is passed from one MAC to another.
The frame format is used for transmitting MAC SDUs between peer MAC
entities and for transmitting MAC recovery information.
Frame Length. In the basic mode as opposed to the HRC., the physical
layer PHY. of FDDI requires limiting the maximum frame length to 9000
symbols, including four symbols of the preamble. The hybrid mode does not
have a limit on frame length, but it is recommended that it also limit its
frame length to 9000 symbols.
Frame Format. Refer to Figures 10.16 and 10.17. The preamble PA. of a
frame is transmitted by the frame originator as a minimum of 16 symbols of
idle pattern only 4 symbols in the hybrid mode.. This sequence allows for
station synchronization and frame alignment. PHYs of subsequent repeating
stations may change the length of the idle pattern consistent with PHY
clocking requirements. Thus, repeating stations may see a variable-length
preamble that may be shorter or longer than the originally transmitted
preamble. Stations in the basic mode are not required to be capable of
copying frames with preambles of less than 12 symbols.
The starting delimiter SD. consists of a J and K symbol see Table 10.4..
No frame or token is considered valid unless it starts with this explicit
*NRZI stands for non-return-to-zero, invert on ones.
SDU stands for service data unit.
313
sequence. After the SD there is a frame control FC. field which defines the
C bit, an
type of frame and associated control functions. It contains a class "
address length L. bit, format bits FF., and four control bits ZZZZ.. For
example, the C bit tells us whether the operation is asynchronous or synchronous; the address length L. bit tells us whether the address is 16 or 48
bits in length; frame format bits indicate whether a token is restricted or
nonrestricted, whether the frame is a station management frame, a MAC
frame, or an LLC frame, among others.
Destination and source addresses DA and SA. follow the FC. These
addresses may be 16 or 48 bits long. However, all stations must have a 16-bit
address capability as a minimum. A station with only a 16-bit address
capacity must be capable of functioning in a ring with stations concurrently
operating with 48-bit addresses.
The routing information RI. field is included in the frame when there is a
48-bit address field. In this case the RII bit is set to 1. The RI field contains 2
to 30 octets symbol pairs. whose format and meaning are specified in the
ISOrIEC 10038 standard on bridging.
The information INFO. field contains zero, one, or more data symbol
pairs whose meaning is determined by the FC field and whose interpretation
is made by the destination entity e.g., MAC, LLC, or SMT*.. The length of
the INFO field is variable. It does have a maximum length restriction based
on a maximum frame length of 9000 symbols, including the four symbols of
the preamble.
The frame check sequence FCS. is used to detect bits in error within the
frame as well as erroneous addition or deletion of bits to the frame. The
fields covered by the FCS include FC, DA, SA, INFO, and FCS
fields. The FCS generating polynomial is the ANSI 32-bit standard algorithm
given in Section 10.1.3.2.
The ending delimiter ED. is the symbol T, which ends tokens and frames.
Ending delimiters and optional control indicators form a balanced symbol
sequence. These are transmitted in pairs so as to maintain octet boundaries.
This is accomplished by adding a trailing symbol T as required. The ending
delimiter of a token consists of two consecutive T symbols; the ending
delimiter of a frame has a single T symbol.
The frame status FS. field consists of an arbitrary length sequence of
control indicator symbols R and S.. The FS field follows the ending delimiter of a frame. It ends if any symbol other than R and S is received. A
trailing T symbol, if present, is repeated as part of the FS field. The first
three control indicators of the FS field are mandatory. These indicators
indicate error detected E., address recognized A., and frame copied C..
The use of additional trailing control indicators in the FS field is optional
and may be defined by the user.
*SMT stands for station management.
314
10.1.4.4 FDDI Timers and Timing Budget. There are three timers used
to regulate the operation of the FDDI ring: token-holding timer THT., valid
transmission timer TVX., and the token rotation timer TRT.. The values of
these timers are locally administered. They may vary from station to station
on the ring, provided that the applicable ring limits are not violated.
Definitions. Token time is the time required to transmit a token 6 symbols.
and a nominal length preamble 16 symbols.. Token time token length. in
basic mode: 0.00088 ms; in hybrid mode, 0.00089 ms to 0.11458 ms.
D max is the maximum latency circulation delay. for a starting delimiter
to travel around the logical ring. It includes node delay and propagation
delay.
D max s 1.7745 ms default . s maximum ring latency with hybrid mode
disabled.
D max s 2.8325 ms default . s maximum ring latency with hybrid mode
enabled.
10.1.4.5 FDDI Ring Operation. Access to the ring is controlled by passing
the token around the ring as in standard token-ring practice. The token gives
the downstream station the receiving station relative to the station passing the token. the opportunity to transmit a frame or sequence of frames. If a
station has frames. in queue for transmission, it strips the token from the
ring before the frame control field of the token is repeated. After the
captured token is completely received, the station begins transmitting
the eligible queued frames. After completion of transmission, the station
issues a new token for use by a downstream station.
Stations that have nothing to transmit at a particular time merely repeat
the incoming symbol stream. While in the process of repeating the incoming
symbol stream, the station determines whether the information is intended
for that station. This is done by matching the downstream address to its own
address or a relevant group address. If a match occurs, subsequent received
symbols up to the FCS are processed by the MAC or sent to the LLC.
Frame stripping is an important concept. Each transmitting station is
responsible for stripping frames from the ring that it originated. This is
accomplished by stripping the remainder of each frame whose source address
matches the stations own address from the ring and replacing it with idle
symbols symbol I Table 10.4..
It will be noted that the process of stripping leaves remnants of frames
consisting of the PA, SD, FC, DA, and SA fields, followed by idle symbols.
This happens because the decision to strip a frame is based on the recognition of the stations own address in the SA field. This cannot occur until after
the initial part of the frame has already been repeated and passed on to the
next downstream station. These remnants cause no ill effects. This is because
of the various specified criteria, including recognition of an ending delimiter
315
ED., which must be met to indicate that a frame is valid. To the level of
accuracy required by statistical purposes, the remnants cannot be distinguished from errored or lost frames because they are always followed by an
idle I. symbol pattern. Such remnants are removed from the ring when they
encounter the first transmitting station.
The FDDI specification distinguishes asynchronous from synchronous
transmission. In this context, asynchronous transmission is a class of data
transmission service whereby all requests for service contend for a pool of
dynamically allocated ring bandwidth and response time. Synchronous transmission is a class of data service whereby each requester is preallocated a
maximum bandwidth and guaranteed a response time not to exceed a specific
delay.
Clocking. A local clock is used to synchronize both the internal operation of
the PHY and its interface to the data-link layer DLL.. The clock is derived
from a fixed frequency reference. This reference may be created internally
within the PHY implementation or supplied to the PHY. Often a crystal
oscillator is used for this purpose.
Characteristics of the local clock are as follows:
a.
b.
c.
d.
e.
f.
Nominal symbol time UI. s 40 ns 1rUI s 25 MHz.

Nominal code bit cell time UI. s 8 ns 1rUI s 125 MHz.
Frequency accuracy - "0.005% "50 ppm.
Harmonic content above 125.02 MHz. - y20 dB
Phase jitter above 20 kHz. - "8 0.044 UI pp.
Phase jitter below 20 kHz in hybrid mode. - "270 1.5 UI pp.
The receive function derives a clock by recovering the timing information

from the incoming serial bit stream. This clock is locked in frequency and
phase to the transmit clock of the upstream node. The maximum difference
between the received bit frequency and the local bit frequency is 0.01% of
the nominal frequency. The received frequency can be either slower or faster
than the local frequency, resulting in an excess or a deficiency of bits unless
some compensation is included. The elasticity buffer function provides
this compensation by adding or dropping Idle bits in the preamble between
DLL PDUs.
The operation of the elasticity buffer function produces variations in the
lengths of the preambles between DLL PDUs as they circulate around the
logical ring. The cumulative effect on preamble size of PDU propagation
through many elasticity buffers can result in excessive preamble erosion and,
in hybrid mode, excessive cycle clock jitter. The smoothing function serves to
filter out these undesirable effects.
Section 10.1.4 consists of abstracts of Refs. 3, 4, and 5.
316
10.1.5
Wireless LANS (WLANs)
10.1.5.1 Introduction. A wireless LAN WLAN. offers a flexible alternative to the wired LAN. In many cases a WLAN is an extension to a wired
LAN, augmenting the resources of an enterprise network. Using radio
frequency RF. technology, wireless LANs transmit and receive data by
means of radiated RF or light signals, minimizing or negating the need for
wired connections. WLANs are particularly attractive in the factory and
office environment where repeated rearrangement of communication facilities can be expected.
The following are some of the advantages of wireless LANs:
Ideal when many rearrangements are expected moves and changes..

Mobility. Allows a LAN user to set up anywhere within range of
radiorwireless link.
Installation speed. Rapid installation of system. No requirement to
stretch wires.
Economic. Less expensive than its wire counterpart.
Scalability. Changes in topology; added users.
A wireless LAN consists of interconnected access

together by means of either frequency-hop or direct
trum radio on 2.5 GHz ISM band, or by means of
connectivity. The bit rate of these systems is either 1
based on the IEEE 802.11 standard.
points which are tied

sequence spread specinfrared radiated light
Mbps or 2 Mbps when
10.1.5.2 Definitions
Access Point (AP). Any entity that has station functionality and provides
access to the distribution services via the wireless medium WM. for
associated stations.
Association. The service used to establish APrSTA mapping and enable
STA invocation of the distribution system services.
Basic Service Area (BSA). The conceptual area within which members of a
basic service set may communicate.
Basic Service Set (BSS). A set of stations controlled by a single coordination
function.
Coordination Function (CF). The logical function that determines when a
station operating within a BSS is permitted to transmit and may be able to
receive PDUs via the wireless medium. The CF within a BSS may have
one PCF and shall have one DCF.
Distributed Coordination Function (DCF). A class of coordination function
where the same coordination function logic is active in every station in the
BSS when ever the network is in operation.
317
Distribution. The service which by using the association information. delivers MSDUs within the DS.
Distribution System (DS). A system used to interconnect a set of basic
service sets and integrated LANs to create an extended service set.
Distribution System Medium (DSM). The medium or set of media used by a
distribution system for communications between access points and portals
of an ESS.
Distribution System Services (DSS). The set of services provided by the
distribution system which enable the MAC to transport MSDUs between
stations that are not in direction communication with each other over a
single instance of the WM. These services include transport of MSDUs
between the APs of BSSs within an ESS, transport of MSDUs between
portals and BSSs within an ESS, and the transport of MSDUs between
stations in the same BSS in cases where the MSDU has a multicast or
broadcast destination address or where the destination is an individual
address, but the station sending the MSDU chooses to involve DSS. DS
services are provided between pairs of 802.11 MACs.
Extended Service Area (ESA). The conceptual area within which members of
an extended service set may communicate. An extended service area is
larger or equal to a basic service area and may involve BSSs in overlapping, disjoint, or both configurations.
Extended Service Set (ESS). A set of one or more interconnected basic
service sets and integrated LANs which appear as a single basic service set
to the LLC layer at any station associated with one of those BSSs.
Infrastructure. The infrastructure includes the distribution system medium,
access point, and portal entities, as well as being the logical location of
distribution and integration service functions of an ESS. An infrastructure
contains one or more access points and zero or more portals in addition to
the distribution system.
MAC Management Protocol Data Unit (MMPDU). The unit of data exchanged between two peer MAC entities using the services of the PHY.
MAC Protocol Data Unit (MPDU). The unit of data exchanged between two
peer MAC entities using the services of the PHY.
MAC Service Data Unit (MSDU). The MAC service data unit is information
that is delivered as a unit between MAC service access points.
Net Allocation Vector (NAV). An indicator, maintained by each station, of
time periods when transmission onto the WM shall not be initiated by the
station whether or not the stations CCA function senses the WM as being
busy.
Point Coordination Function (PCF). A class of possible coordination functions where the coordination function logic is active in only one station in
a BSS at any given time that the network is in operation.
318
Figure 10.18. WLAN MAC frame format. (From Figure 12, page 34, ANSIrIEEE Std. 802.11,
Ref. 8.)
Figure 10.19. Frame control field. (From Figure 13, page 35, IEEE Std. 802.11-1997, Ref. 6.)
Portal. The logical point at which MSDUs from an integrated, non-802.11

LAN enter the distribution system of an ESS.
Wireless Medium (WM). The medium used to implement the transfer of
PDUs between peer PHY entities of a wireless LAN.
10.1.5.3 Frame Formats. Figure 10.18 shows the general MAC frame
format. In the figure the following fields are only present for certain frame
types: Address 2, Address 3, Sequence Control, Address 4, and Frame Body.
A selected group of fields is briefly described below.
Frame Control Field. The format of the frame control field is illustrated in
Figure 10.19.
PROTOCOL VERSION.
The Protocol Version subfield consists of 2 bits. For the

present version of IEEE 802.11, its value is 0. All other values are reserved.
The revision level will be incremented only when a fundamental incompatibility exists between a new revision and this revision of the standard.
TYPE AND SUBTYPE FIELDS. The Type field consists of 2 bits, and the Subtype
field is 4 bits long. These two fields identify the function of the frame. There
are three frame types: control, data, and management. Each of the frame
types has several defined subtypes. Type value binary 00 indicates a management frame, type value binary 01 indicates the frame is a control frame, and
binary value 10 for Type indicates that the frame is a data frame. For the
Subtype field, an example is the binary value 0100 which indicates a probe
request.
319
MORE FRAGMENTS. The More Fragments field is 1 bit in length and is set to
binary 1 in all frames which have another fragment of the current MSDU or
current MMPDU to follow. It is set to binary 0 in all other frames.
RETRY. The Retry subfield is 1 bit in length and is set to binary 1 in a Data or
Management type frame that is a retransmission of an earlier frame. A
receiving station uses this indication to aid in the process of eliminating
duplicate frames.
WEP FIELD.
The WEP Wired Equivalent Privacy. field is 1 bit in length. It is

set to binary 1 if the Frame Body field contains information that has been
processed by the WEP algorithm. The WEP field is only set to 1 within
frames of Type Data and frames of Type Management, Subtype Authentication. The WEP field is set to binary 0 in all other frames.
ORDER FIELD.
The Order field is 1 bit in length and is set to 1 in any Data

Type frame that contains an MSDU, or fragment thereof, which is being
transferred using the Strictly Ordered service class. This field is set to 0 in all
other frames.
ADDRESS FIELDS. There are four address fields in the MAC frame format.
These fields are used to indicate the BSSID Basic Service Set Identification .
source address, destination address, transmitting station address and receiving station address. The usage of the four address fields in each frame type is
indicated by the abbreviations BSSID, DA, SA, RA, and TA indicating BSS
Identifier, Destination Address, Source Address, Receiver Address, and
Transmitter Address, respectively. Certain frames may not contain some of
the address fields.
Certain address field usage is specified by the relative position of the
address field 14. within the MAC header, independent of the type of
address present in that field. For example, receiver address matching is
always performed on the contents of the Address 1 field in received frames,
and the receiver address of CTS clear to send. and ACK frames is always
obtained from the Address 2 field in the RTS request to send. frame, or
from the frame being acknowledged.
Each Address field contains a 48-bit address.
Address Designation. A MAC sublayer address is one of the following

two types:
a. Indiidual Address. An address associated with a particular station on
the network.
b. Group Address. A multidestination address, associated with one or
more stations on a given network.
320
The two kinds of group addresses are as follows:

1. Multicast-Group Address. An address associated by higher-level convention with a group of logically related stations.
2. Broadcast Address. A distinguished, predefined multicast address that
always denotes the set of all stations on a given LAN. All 1s in the
Destination Address field are interpreted to be the broadcast address.
This group is predefined for each communication medium to consist of
all stations actively connected to that medium; it is used to broadcast to
all active stations on that medium. All stations are able to recognize
the broadcast address. It is not necessary that a station be capable of
generating the broadcast address.
The address space is also partitioned into locally administered and universal globally administered . addresses. The nature of a body and the procedures by which it administers these universal globally administered . addresses is beyond the scope of the reference standard.
BSSID Field. The BSSID field is a 48-bit field of the same format as an
IEEE 802 MAC address. This field uniquely identifies each BSS basic
service set.. The value of this field, in an infrastructure BSS, is the MAC
currently in use by the STA station. in the AP access point. of the BSS.
The value of this field in an IBSS independent basic service set. is a
locally administered IEEE MAC address formed from a 46-bit random
number generated according to the procedure defined in Section 11.1.3 of
the reference document. The individualrgroup bit of the address is set to 0.
The universalrlocal bit of the address is set to 1. This mechanism is used to
provide a high probability of selecting a unique BSSID.
The value of all 1s is used to indicate the broadcast BSSID. A broadcast
BSSID may only be used in the BSSID field of management frames of
subtype probe request.
Destination Address (DA) Field. The DA field contains an IEEE MAC
individual or group address that identifies the MAC entity or entities intended as the final recipient s. of the MSDU or fragment thereof. contained
in the frame body.
Source Address (SA) Field. The SA field contains an IEEE MAC individual address that identifies the MAC entity from which the transfer of the
MSDU or fragment thereof. contained in the frame body field was initiated.
The individualrgroup bit is always transmitted as a zero in the source
address.
321
Figure 10.20. Sequence control field.
Receier Address (RA) Field. The RA field contains an IEEE MAC individual or group address that identifies the intended immediate recipient wSTSs.,
on the WM wireless medium.x for the information contained in the frame
body field.
Transmitter Address (TA) Field. The TA field contains an IEEE MAC
individual address that identifies the STA that has transmitted, onto the
WM, the MPDU contained in the frame body field. The IndividualrGroup
bit is always transmitted as a zero in the transmitter address.
SEQUENCE CONTROL AND SEQUENCE NUMBERS FIELDS.
The Sequence Control

field is 16 bits long and consists of two subfields: the Sequence Number and
the Fragment Number. The format of the Sequence Control field is illustrated in Figure 10.20. The Sequence Number field is a 12-bit field indicating
the sequence number of the MSDU or MMPDU. Each MSDU or MMPDU
transmitted by an STA is assigned a sequence number. Sequence numbers
are assigned from a single modulo 4096 counter, starting at 0 and incrementing by 1 for each MSDU or MMPDU. Each fragment of an MSDU or
MMPDU contains the assigned sequence number. The sequence number
remains constant in all retransmissions of an MSDU, MMPDU, or fragment
thereof.
Fragment Number Field. The Fragment Number field is a 4-bit field
indicating the number of each fragment of an MSDU or MMPDU. The
fragment number is set to zero in the first or only fragment of an MSDU or
MMPDU and is incremented by one for each successive fragment of the
MSDU or MMPDU. The fragment number remains constant in all retransmissions of the fragment.
FRAME BODY FIELD. The frame body is a variable length field that contains
information specific to individual frame types and subtypes. The minimum
frame body is 0 octets. The maximum-length frame body is defined by the
maximum length MSDU q ICV q IV., where ICV and IV are the WEP
fields defined in Section 8.2.5 of the reference publication.
THE FCS FIELD. The FCS field is a 32-bit field containing the standard 32-bit
CRC defined in Section 10.1.3.2.
322
Figure 10.21. RTS frame. (From Figure 16, page 41, ANSIrIEEE Std. 802.11, Ref. 8.)
RTS REQUEST TO SEND. AND CTS CLEAR TO SEND. FRAME FORMATS. The
format for the RTS frame is defined in Figure 10.21.
The RA receiver address . of the RTS frame is the address of the station,
on the wireless medium, that is the intended immediate recipient of the
pending directed Data or Management frame. The TA is the address of the
station transmitting the RTS frame.
The Duration value is the time, in microseconds, required to transmit the
pending Data or Management frame, plus one CTS frame, plus one ACK
frame, plus three SIFS short interframe space. intervals. If the calculated
duration includes a fractional microsecond, that value is rounded up to the
next higher integer.
The format of the CTS frame is shown in Figure 10.22.
The RA of the CTS frame is copied from the Transmitter Address TA.
field of the immediately previous RTS frame to which the CTS is a response.
The Duration value is the value obtained from the Duration field of the
Immediately previous RTS frame, minus the time, in microseconds, required
to transmit the CTS frame and its SIFS interval. If the calculated duration includes a fractional microsecond, it is rounded up to the next higher
integer.
ACK FRAME FORMAT.
The frame format of the ACK frame is given in

Figure 10.23.
The Receiver Address RA. of the ACK frame is copied from the Address
2 field of the immediately previous direct Data, Management, or PS-Poll
Control frame.
Figure 10.22. CTS frame.
323
Figure 10.23. ACK frame.
If the More Fragment bit was set to 0 in the Frame Control field of the
immediately previous directed data or management frame, the duration value
is set to 0. If the More Fragment bit was set to 1 in the Frame Control field
of the immediately previous directed data or management frame, the duration value is the value obtained from the Duration field of the immediately
previous data or management frame, minus the time, in microseconds,
required to transmit the ACK frame and its SIFS interval. If the calculated
duration includes a fractional microsecond, that value is rounded up to the
next higher integer.
10.1.5.4 Overview of the WEP. Eavesdropping is a familiar problem of
other types of wireless technology. IEEE 802.11 specifies a wired LAN
equivalent data confidentiality algorithm. Wired equialent priacy WEP. is
defined as protecting authorized users of a wireless LAN from casual
eavesdropping. This service is intended to provide functionality for the
wireless LAN equivalent to that provided by the physical security attributes
inherent to a wired medium. The use of the WEP algorithm is optional.
10.1.5.5 MAC Access
MAC Architecture. One way of illustrating the WLAN MAC architecture is
illustrated in Figure 10.24. It is shown as providing the PCF point coordination function. through the services of the DCF distributed coordination
function..
DISTRIBUTED COORDINATION FUNCTION DCF..
The principal underlying access

method is a DCF known as carrier sense multiple access with collision aoidance CSMArCA.. The DCF is implemented at all stations STAs. for use
without both the IBSS and infrastructure network configurations.
For a station to transmit, it uses the medium to determine if another
station is transmitting. If the medium is not determined to be busy see
Carrier Sense Mechanism,., the transmission may proceed. The CSMArCA
distributed algorithm mandates that a gap of a minimum specified duration
exist between contiguous frame sequences. A transmitting station ensures
that the medium is idle for this required duration before attempting to
transmit. If the medium is determined to be busy, the station defers until the
324
Figure 10.24. MAC architecture. (From Figure 47, page 70, ANSIrIEEE Std. 802.11, Ref. 8.)
end of the current transmission. After deferral, or prior to attempting to

transmit again immediately after a successful transmission, the station selects
a random backoff interval and decrements the backoff interval counter while
the medium is idle. A refinement of the method may be used under various
circumstances to further minimize collisions; here the transmitting and receiving stations exchange short control sequences wrequest to send RTS. and
clear to send CTS. framesx after determining that the medium is idle and
after deferrals or backoffs, prior to data transmission. Deferrals and
Backoffs are described in the following text.
PCF.. The IEEE 802.11 MAC may also
incorporate an optional access method called a PCF, which is only usable on
infrastructure network configurations. This access method uses a point coordinator PC., which operates at the access point of the BSS to determine
which station currently has the right to transmit. The operation is essentially
that of polling, with the PC performing the roll of polling master. The
operation of the PCF may require additional coordination, not specified in
the reference standard, to permit efficient operation in cases where multiple
point-coordinated BSSs are operating on the same channel in overlapping
physical space.
The PCF uses a virtual carrier-sense mechanism aided by an access
priority mechanism. The PCF distributes information within Beacon management frames to gain control of the medium by setting the network allocation
vector NAV. in stations. In addition, all frame transmission under the PCF
may use an interframe space IFS. that is smaller than the IFS for frames
transmitted via the DCF. The use of a smaller IFS implies that point-coordinated traffic has priority access to the medium over stations in overlapping
BSSs operating under the DCF access method.
POINT COORDINATION FUNCTION
325
The access priority provided by a PCF may be utilized to create a

contention-free CF. access method. The PC controls the frame transmissions
of the stations so as to eliminate contention for a limited period of time.
COEXISTENCE OF DCF AND PCF.
The DCF and PCF coexist in a manner that

permits both to operate concurrently within the same BSS. When a PC is
operating in a BSS, the two access methods alternate, with a contention-free
period CFP. followed by a contention period CP..
FRAGMENTATION r DEFRAGMENTATION OVERVIEW. The process of partitioning

a MAC service data unit MSDU. or a MAC management protocol data unit
MMPDU. into smaller MAC level frames, MAC protocol data units
MPDUs., is called fragmentation. Fragmentation creates MPDUs smaller
than the original MSDU or MMPDU length to increase reliability, by
increasing the probability of successful transmission of the MSDU or MMPDU in cases where channel characteristics limit reception reliability for
longer frames. Fragmentation is accomplished at each immediate transmitter.
The process of recombining MPDUs into a single MSDU or MMPDU is
defined as defragmentation. Defragmentation is accomplished at each immediate recipient.
Only MPDUs with a unicast receiver address shall be fragmented. Broadcastrmulticast frames shall not be fragmented even if their length exceeds
aFragmentationThreshold.
When a directed MSDU is received from the LLC or a directed MMPDU
is received from the MAC sublayer management entity MLME. with a
length greater than aFragmentationThreshold, the MSDU or MMPDU shall
be fragmented. The MSDU or MMPDU is divided into MPDUs. Each
fragment is a frame no larger than aFragmentationThreshold. It is possible
that any fragment may be a frame smaller than aFragmentationThreshold.
An illustration of fragmentation is shown in Figure 10.25.
The MPDUs resulting from the fragmentation of an MSDU or MMPDU
are sent as independent transmissions, each of which is separately acknowledged. This permits transmission retries to occur per fragment, rather than
per MSDU or MMPDU. Unless interrupted due to medium occupation
limitations for a given PHY, the fragments of a single MSDU or MMPDU
Figure 10.25. Fragmentation. (From Figure 48, page 71, IEEE Std. 802.11-1997, Ref. 6.)
326
are sent as a burst during the CP, using a single invocation of the DCF
medium access procedure. The fragments of a single MSDU or MMPDU are
sent during a CFP as individual frames obeying the rules of the PC medium
access procedure.
MAC DATA SERVICE. The MAC data service translates MAC service requests
from an LLC into input signals utilized by the MAC state machines. The
MAC data services also translates output signals from the MAC state
machines into service indications to an LLC. The translations are given in the
MAC data service state machine defined in Annex C of the reference
document.
DCF. The basic medium access protocol is a DCF that allows for automatic
medium sharing between compatible PHYs through the use of CSMArCA
and a random backoff time following a busy medium condition. In addition,
all directed traffic uses immediate positive acknowledgment ACK frame.
where retransmission is scheduled by the sender if no ACK is received.
The CSMArCA protocol is designed to reduce the collision probability
between multiple STAs accessing a medium, at the point where collisions
would most likely occur. Just after the medium becomes idle following a busy
medium as indicated by the CS function. is when the highest probability of a
collision exists. This is because multiple STAs could have been waiting for
the medium to become available again. This is the situation that necessitates
a random backoff procedure to resolve medium contention conflicts.
Carrier sense shall be performed both through physical and virtual
mechanisms.
The virtual carrier-sense mechanism is achieved by distributing reservation
information announcing the impending use of the medium. The exchange of
RTS and CTS frames prior to the actual data frame is one means of
distribution of this medium reservation information. The RTS and CTS
frames contain a DurationrID field that defines the period of time that the
medium is to be reserved to transmit the actual data frame and the returning
ACK frame. All STAs within the reception range of either the originating
STA which transmits the RTS. or the destination STA which transmits the
CTS. shall learn of the medium reservation. Thus an STA can be unable to
receive from the originating STA, yet still know about the impending use of
the medium to transmit a data frame.
Another means of distributing the medium reservation information is the
DurationrID field in directed frames. This field gives the time that the
medium is reserved, either to the end of the immediately following ACK, or
in the case of a fragment sequence, to the end of the ACK following the next
fragment.
The RTSrCTS exchange also performs both a type of fast collision
inference and a transmission path check. If the return CTS is not detected by
the STA originating the RTS, the originating STA may repeat the process
after observing the other medium-use rules. more quickly than if the long
327
data frame had been transmitted and a return ACK frame had not been
detected.
Another advantage of the RTSrCTS mechanism occurs where multiple
BSSs utilizing the same channel overlap. The medium reservation mechanism
works across the BSA boundaries. The RTSrCTS mechanism may also
improve operation in a typical situation where all STAs can receive from the
AP, but cannot receive from all other STAs in the BSA.
The RTSrCTS mechanism cannot be used for MPDUs with broadcast and
multicast immediate address because there are multiple destinations for the
RTS, and thus potentially multiple concurrent senders of the CTS in response. The RTSrCTS mechanism need not be used for every data frame
transmission. Because the additional RTS and CTS frames add overhead inefficiency, the mechanism is not always justified, especially for short
data frames.
The use of the RTSrCTS mechanism is under control of the
dot11RTSThreshold attribute. This attribute may be set on a per-STA basis.
This mechanism allows STAs to be configured to use RTSrCTS either
always, never, or only on frames longer than a specified length.
An STA configured not to initiate the RTSrCTS mechanism shall still
update its virtual carrier-sense mechanism with the duration information
contained in a received RTS or CTS frame, and shall always respond to an
RTS addressed to it with a CTS.
The medium access protocol allows for STAs to support different sets of
data rates. All STAs shall receive all the data rates in aBasicRateSet and
transmit at one or more of the aBasicRateSet data rates. To support the
proper operation of the RTSrCTS and the virtual carrier-sense mechanism,
all STAs shall be able to detect the RTS and CTS frames. For this reason the
RTS and CTS frames shall be transmitted at one of the aBasicRateSet rates.
See Section 9.6 of the reference document for a description of multirate
operation..
Data frames sent under the DCF shall use the frame type Data and
subtype Data or Null Function. STAs receiving Data type frames shall only
consider the frame body as the basis of a possible indication to LLC.
CARRIER SENSE MECHANISM. Physical and virtual carrier-sense functions are
used to determine the state of the medium. When either function indicates a
busy medium, the medium is considered busy; otherwise, it is considered idle.
A physical carrier-sense mechanism is provided by the PHY. See Section
12 of the reference document for how this information is conveyed to the
MAC. The details of physical carrier sense are provided in the individual
PHY specifications.
A virtual carrier-sense is provided by the MAC. The mechanism is referred to as the network allocation ector NAV.. The NAV maintains a
prediction of future traffic on the medium based on duration information
that is announced in the RTSrCTS frames prior to actual exchange of data.
328
The duration information is also available on the MAC headers of all frames
sent during the CP other than PS-POLL Control frames.
The carrier-sense mechanism combines the NAV state and the stations
transmitter status with the physical carrier sense to determine the busyridle
state of the medium. The NAV may be thought of as a counter which counts
down to zero at a uniform rate. When the counter is zero, the virtual
carrier-sense indication is that the medium is idle; when nonzero, the
indication is busy. The medium is determined to be busy whenever the
station is transmitting.
MAC-LEVEL ACKNOWLEDGMENTS. The reception of some frames requires the
receiving station to respond with an acknowledgment, generally an ACK
frame, if the FCS of the received frame is correct. This technique is known as
positive acknowledgment.
Lack of reception of an expected ACK frame indicates to the source STA
that an error has occurred. Note, however, that the destination station may
have received the frame correctly and that the error may have occurred in
the reception of the ACK frame. To the initiator of the frame exchange, this
condition is indistinguishable from an error occurring in the initial frame.
INTERFRAME SPACE IFS.. The time interval between frames is called the IFS.
A station determines that the medium is idle through the use of the
carrier-sense function for the interval specified. Four different IFSs are
defined to provide priority levels for access to the wireless media; they are
listed in order, from the shortest to the longest. Figure 10.26 illustrates some
of these relationships.
1.
2.
3.
4.
SIFS
PIFS
DIFS
EIFS
short interframe space

PCF interframe space
DCF interframe space
extended interframe space
The different IFSs shall be independent of the STA bit rate. The IFS
timings shall be defined as time gaps on the medium, and they shall be fixed
Figure 10.26. Some IFS relationships. (From Figure 49, page 74, ANSIrIEEE Std. 802.11,
Ref. 8.)
329
for each PHY even in multirate-capable PHYs.. The IFS values are determined from attributes specified by the PHY.
Short IFS (SIFS). The SIFS is used for an ACK frame, a CTS frame, and
the second or subsequent MPDU of a fragment burst, as well as by a station
responding to any polling by the PCM. It may also be used by a PC for any
types of frames during the CFP. The SIFS is the time from the end of the last
symbol of the previous frame to the beginning of the first symbol of the
preamble of the subsequent frame as seen at the air interface.
The SIFS timing is achieved when the transmission of the subsequent
frame is started at the TxSIFS Slot boundary as specified in Section 9.2.10 of
the reference document. An IEEE 802.11 implementation does not allow the
space between frames that are defined to be separated by a SIFS time, as
measured on the medium, to vary from the nominal SIFS value by more than
"10% of aSlotTime for the PHY in use.
SIFS is the shortest of the interframe spaces. SIFS is used when stations
have seized the medium and need to keep it for the duration of the frame
exchange sequence to be performed. Using the smallest gap between transmissions within the frame exchange sequence prevents other stations which
are required to wait for the medium to be idle for a longer gap. from
attempting to use the medium, thus giving priority to completion of the frame
exchange sequence in progress.
PCF IFS (PIFS). The PIFS is used only by stations under the PCF to gain
priority access to the medium at the start of the CFP. A station using the
PCF is allowed to transmit contention-free traffic after its carrier-sense
mechanism determines that the medium is idle at the TxPIFS slot boundary
as defined in Section 9.2.10 of the reference document. Section 9.3 of this
document describes the use of the PIFS by stations operating under the PCF.
DCF IFS (DIFS). The DIFS is used by stations operating under the DCF
to transmit data frames MPDUs. and management frames MMPDUs.. A
station using the DCF is allowed to transmit if its carrier-sense mechanism
determines the medium is idle at the TxDIFS slot boundary as defined in
Section 9.2.10 of the reference document after a correctly received frame,
and its backoff time has expired. A station using the DCF does not transmit
within an EIFS after it determines that the medium is idle following reception of a frame for which the PHYRXEND.indication primitive contained an
error or a frame for which the MAC FCS value was not correct. A station
may transmit after subsequent reception of an error-free frame, resynchronizing the station. This allows the station to transmit using the DIFS following
that frame.
Extended IFS (EIFS). The EIFS is used by the DCF whenever the PHY
has indicated to the MAC that a frame transmission has begun that did not
result in the correct reception of a complete MAC frame with a correct FCS
value. The duration of an EIFS is defined in Section 9.2.10 of the reference
330
document. The EIFS interval begins following indication by the PHY that the
medium is idle after detection of the erroneous frame, without regard to the
virtual carrier-sense mechanism. The EIFS is defined to provide enough time
for another station to acknowledge what was, to this station, an incorrectly
received frame before this station commences transmission. Reception of an
error-free frame during the EIFS resynchronizes the station to the actual
busyridle state of the medium, so the EIFS is terminated and normal
medium access using DIFS and, if necessary, backoff. continues following
reception of that frame.
RANDOM BACKOFF TIME.
A station desiring to initiate transfer of data MPDUs

andror management MMPDUs shall invoke the carrier-sense mechanism to
determine the busyridle state of the medium. If the medium is busy, the
station shall defer until the medium is determined to be idle without
interruption for a period of time equal to DIFS when the last frame detected
on the medium was received correctly, or after the medium is determined to
be idle without interruption for a period of time equal to EIFS when the last
frame detected on the medium was not received correctly. After this DIFS or
EIFS medium idle time, the station shall then generate a random backoff
period for an additional deferral time before transmitting, unless the backoff
timer already contains a nonzero value, in which case the selection of a
random number is not needed and not performed. This process minimizes
collisions during contention between multiple stations that have been deferring to the same event.
Backoff Time s Random . = aSlotTime
where
Random . s Pseudorandom integer drawn from a uniform distribution over the interval w0, CWx, where CW is an integer
within the range of values of the PHY characteristics
aCWmin and aCWmax, aCWmin F CW F aCWmax. It
is important that designers recognize the need for
statistical independence among the random number
streams among stations
aSlotTime s Value of the correspondingly named PHY characteristic.
The contention window CW. parameter shall take an initial value of
aCWmin. Every STA shall maintain an STA short retry count SSRC. as well
as an STA long retry count SLRC., both of which shall take an initial value
of zero. The SSRC shall be incremented whenever any short retry count
associated with any MSDU is incremented. The SLRC shall be incremented
whenever any long retry count associated with any MSDU is incremented.
The CW shall take the next value in the series every time an unsuccessful
attempt to transmit an MPDU causes either station retry counter to increment, until the CW reaches the value of aCWmax. A retry is defined as the
331
Figure 10.27. An example of exponential increase of CW. (From Figure 50, page 76, IEEE
Std. 802.11-1997, Ref. 6.)
entire sequence of frames sent, separated by SIFS intervals, in an attempt to

deliver an MPDU, as described in Section 9.7 of the reference document.
Once it reaches aCWmax, the CW remains at the value of aCWmax until it is
reset. This improves stability of the access protocol under high-load conditions. See Figure 10.27.
The CW shall be reset to aCWmin after every successful attempt to
transmit an MSDU or MMPDU, when SLRC reaches aLongRetryLimit, or
when SSRC reaches dot11ShortRetryLimit. The SSRC shall be reset to 0
whenever a CTS frame is received in response to an RTS frame, whenever an
ACK frame is received in response to an MPDU or MMPDU transmission,
or whenever a frame with a group address in the Address1 field is transmitted. The SLRC shall be reset to 0 whenever an ACK frame is received in
response to transmission of an MPDU or MMPDU of length greater than
dot11RTSThreshold, or whenever a frame with a group address in the
Address1 field is transmitted.
The set of CW values shall be sequentially ascending integer powers of 2,
minus 1, beginning with a PHY-specific aCWmin value, and continuing up to
and including a PHY-specific aCWmax value.
DCF Basic Access Procedure. The CSMArCA access method is the foundation of the DCF. Operational rules vary slightly between the DCF and
the PC.
Basic access refers to the core mechanism a station uses to determine
whether it may transmit. In general, a station may transmit a pending MPDU
when it is operating under the DCF access method, either in the absence of a
332
PC or in the CP of the PCF access method, when the station determines that
the medium is idle for greater than or equal to a DIFS period, or an EIFS
period if the immediately preceding medium-busy event was caused by
detection of a frame that was not received at this station with a correct MAC
FCS value. If, under these conditions, the medium is determined by the
carrier-sense mechanism to be busy when a station desires to initiate the
initial frame of one of the frame exchanges described in Section 9.7 of
the reference document, exclusive of the CF period, the random backoff
algorithm described in Section 9.2.5.2 of the reference document is followed.
There are conditions, specified in Sections 9.2.5.2 and 9.2.5.5 of the reference
document, where the random backoff algorithm is followed even for the first
attempt to initiate a frame exchange sequence.
In a station having a frequency hop FH. PHY, control of the channel is
lost at the dwell time boundary and the station has to contend for the
channel after that dwell boundary. It is required that the stations having an
FH PHY complete transmission of the entire MPDU and associated acknowledgment if required. before the dwell time boundary. If, when transmitting or retransmitting an MPDU, there is not enough time remaining in
the dwell to allow transmission of the MPDU plus acknowledgment if
required., the station defers the transmission by selecting a random backoff
time, using the present CW without advancing to the next value in the
series.. The short retry counter and long retry counter for the MSDU are not
affected. The basic access mechanism is illustrated in Figure 10.28.
10.1.5.6 Overview of the Air Interface. The data rates presently offered
by the 802.11 specification are 1.0 and 2.0 Mbps.
There are three different modulation plans: Two of these plans use a form
of spread spectrum, and the third uses radiated light in the IR region. The
spread spectrum types are frequency hop FH. and direct sequence DS..
The hopping rate for FH systems is 2.5 or more hopsrs.
Frequency Hopping Spread Spectrum Operation. Section 15.247 of the
FCC Rules and Regulations CFR 47, Ref. 7. states that the dwell time is not
Figure 10.28. Basic access method. (From Figure 51, page 77, IEEE Std. 802.11-1997,
Ref. 6.)
333
greater than 0.4 hopsrs in any 30 hopsrs period. This type of operation is
permitted in two frequency bands: 2400 to 2483.5 MHz and 5725 to 5850
MHz. The IEEE 802.11 specification only allows operation in the former.
The FCC has assigned 79 specific hop frequencies in the 2.4-GHz band
beginning at 2402 MHz and extending to 2483 MHz. These frequencies are
valid for North America and Europe, not including France and Spain.
Maximum transmitter output power FCC Part 15.247. is 1.0 watt q30
dBm.. Transmit antennas may have gains as high as q6 dBi before there is a
power penalty. For example, if a transmit antenna has a gain of q10 dBi, the
transmit output power must be reduced 4 dB. Thus, the maximum EIRP is
q36 dBm or q6 dBW.
If we were to allow a receive threshold of y80 dBm Section 14.6.15.4 of
IEEE 802.11, Ref. 8., what kind of line-of-sight range could we expect?
Based on Radio System Design for Telecommunications, Ref. 9., the free
space loss FSL. equation is
FSL s 36.58 q 20 log D q 20 log F in dB .
where D is in miles and F is in MHz.
Let F s 2483 MHz, what is D?
FSL s 36.58 q 20 log 2483 . q 20 log D
s 36.58 q 67.9 q 20 log D
At 1 mile, FSL s 104.48 dB.
With q36-dBm EIRP, the isotropic receive level IRL. would be
IRL s q36 dBm y 104.48 dB
s y68.48 dBm
If we assume that the gain of the receive antenna is 0 dBi and that the
transmission line to the receiver input is lossless, then there is a margin on
the link of y80 dBM* y y68.48 dBm. or 11.52 dB. This is at 1-mile
distance. At half a mile, the margin will be 6 dB greater, or 17.52 dB; at one
quarter-mile, there is another 6-dB improvement giving a 23.52-dB margin.
Because of the complex propagation conditions inside a building, link margin
is at a premium. Remember, every time we reduce to half of its value, the
free space loss decreases 6 dB Ref. 9..
The modulation employed for the radio systems is a form of frequency
shift keying FSK. called Gaussian FSK. For 1-Mbps systems it is binary
GFSK, and for the 2-Mbps systems it is 4-ary GFSK with a theoretical
channel occupancy of 2 bitsrHz of bandwidth. The channel occupancy for
either 1 or 2 Mbps is 1 MHz.
One concern of a system designer is the maximum permissible input level
to the receiver and its dynamic range. The maximum input signal is y20
dBm, and the dynamic range should extend to y8 dBm.
*Input threshold for a frame error ratio FER. 8 = 10y2 at 2-Mbps operation.
334
Direct Sequence Spread Spectrum (DSSS) Operation. The DSSS system

provides a wireless LAN with both 1-Mbps and 2-Mbps data payload communication capability. Based on US FCC Part 15.247, the DSSS system must
provide a processing gain of at least 10 dB. This is accomplished by chipping
the baseband signal at 11 MHz with an 11-bit PN code. The baseband
modulation employed is differential binary phase shift keying DBPSK. for
1-Mbps operation and differential quadrature phase shift keying DQPSK.
for the 2-Mbps data rate.
The North American and ETSI European, less France and Spain. provide
11 channels starting with 2412 MHz and extending up to 2462 MHz with
5-MHz spacing. ETSI adds two more 5-MHz channels to the group, 2467 and
2472 MHz. Transmit power and antenna gain values are the same as for
FHSS systems given previously.
Transmit RF power control is provided for transmitted power greater than
100 mW. A maximum of four power levels may be provided. At a minimum, a
radio capable of transmission greater than 100 mW must be capable of
switching power back to 100 mW or less.
At the receiver, the frame error ratio FER. must be less than 8 = 10y2 at
an MPDU length of 1024 bytes for an input level of y80 dBm measured at
the antenna connector. The FER is specified for 2-Mbps DQPSK modulation. A similar FER must be achieved with an input level as high as y4 dBm
measured at the antenna.
Infrared (IR) Transmission. The physical layer for the infrared system uses
near-visible light in the 850- to 950-nm range. This is similar to the spectral
usage of both common consumer devices such as infrared remote controls, as
well as other data communications equipment, such as Infrared Data Association IrDA. devices.
Unlike many other infrared devices, however, the IR PHY is not directed.
That is, the receiver and transmitter do not have to be aimed at each other
and do not need a clear line-of-sight. This permits the construction of a true
LAN system, whereas with an aimed system, it would be difficult or impossible to install a LAN because of physical constraints.
A pair of conformant infrared devices would be able to communicate in a
typical environment at a range up to about 10 m. This standard allows
conformant devices to have more sensitive receivers, and this may increase
range up to about 20 m.
The IR PHY relies on both reflected infrared energy as well as line-of-sight
infrared energy for communications. Most designs anticipate that all of the
energy at the receiver is reflected energy. This reliance on reflected infrared
energy is called diffuse infrared transmission.
This standard specifies the transmitter and receiver in such a way that a
conformant design will operate well in most environments where there is no
line-of-sight path from the transmitter to the receiver. However, in an
environment that has few or no reflecting surfaces and where there is no
line-of-sight, an IR PHY system may suffer reduced range.
LAN BRIDGES OVERVIEW
335
The IR PHY will operate only in indoor environments. Infrared radiation

does not pass through walls, and it is significantly attenuated passing through
most exterior windows. This characteristic can be used to contain an IR
PHY in a single physical room, like a classroom or conference room.
Different LANs using the IR PHY can operate in adjacent rooms separated
only by a wall without interference and without the possibility of eavesdropping.
At the time of this standards preparation, the only known regulatory
standards that apply to the use of infrared radiation are safety regulations,
such as IEC 60825-1: 1998 wB2x and ANSI Z136.1-1993 wBx. While a conformant IR PHY device can be designed to also comply with these safety
standards, conformance with this standard does not ensure conformance with
other standards.
Worldwide, there are currently no frequency allocation or bandwidth
allocation regulatory restrictions on infrared emissions.
10.2
10.2.1
REPEATERS, BRIDGES, ROUTERS, AND HUBS

Definitions and Interactions
In these types of applications, a repeater extends a LANs range see

Figure 10.29.. It simply regenerates the baseband signal being carried on the
LAN. A repeater operates on the physical layer. A bridge may be used to
segment a LAN. Users on a segment have a high community of interest. They
do this based on the MAC destination address es. the work-station hardware address .. Routers connect dissimilar LANs or provide LAN-to-WAN
connectivity by operating at OSI layer 3, the network layer. Repeaters,
bridges, and routers are compared in Figure 10.29.
10.3
LAN BRIDGESOVERVIEW
There are a number of types of bridges. Well start with the transparent
bridge, the learning bridge, and then discuss the spanning tree algorithm.
Figure 10.29. Conceptual drawing illustrating the functions of repeaters, bridges, and routers.
336
The transparent bridge simply relays framesrpackets. Such a bridge has

several ports. Framesrpackets received on a particular port are stored and
then forwarded on all other ports. At first blush, it would appear that this
type of bridge is the same as a repeater. However, this is not the case. A
LAN repeater simply repeats an incoming bit stream bit-by-bit. A transparent
bridge, on the other hand, receives an entire frame and delivers the frame to
an outgoing ports.. In the case of CSMArCD, an outgoing port is just
another station on the LAN to which it is connected. The port must wait
until the LAN is free of transmissions before it sends the frame. Such a
frame is sent unaltered. Thus the bridge is transparent.
If a number of LANs were connected to the bridge, traffic peaks could
possibly occur. If the bridge has sufficient storage capacity, there would be no
traffic peak problem. In this case the bridge would hold the frames in queue
and the frames would be transmitted on each port other than the incoming
port. as their turn arises at each port of interest. If the LAN has insufficient
storage capacity, some frames would have to be dropped and would then
be lost.
Adding Intelligence to a Bridge. Suppose a processor is incorporated in the
bridge design so that ports are associated with destination and source fields.
Knowing the source field assures that a frame is not returned to the
originators LAN. Knowledge of destination fields and their association with
ports cuts down on unnecessary traffic on LANs. What we mean here is that
a particular frame will only travel to LANs. containing stations with the
pertinent MAC destination addresses, whereas before, with the dumb
bridges described above, frames went out every port but the port associated
with that incoming frame. All frames received were relayed out all bridge
ports but the port of origin of a particular frame.
Radia Perlman Ref. 11. describes the strategy of such a smart bridge as
follows:
1. The bridge receives every packet transmitted.
2. For each packet received, the bridge reads the source address and
along with it the port on which the packet was received. This information is stored in a cache,* which can be called the station cache.
3. For each received packet, the station cache is reviewed by the bridge
for a framerpacket destination address.
a. If no destination address is found in the cache, the framerpacket is
forwarded to each of the bridges ports less the port on which the
framerpacket was received.
*The IEEE defines cache as follows: Ref. 12. A buffer inserted between one or more
processors and the bus, used to hold currently active copies of blocks from main memory.
2. A small portion of high-speed memory used for temporary storage of frequently used data,
instructions, or operands.
337
b. If the address is located in the station cache, the bridge forwards

the framerpacket only to the port indicated in the cache table. If
the specified interface is the one from which the framerpacket
was received, the packet is discarded filtered in IEEE 802.1
terminology..
4. The bridge ages each entry in the station cache and deletes it after a
period of time a parameter known as aging time. in which no traffic is
received with that address as the source address.
Drawbacks of LAN Bridges: Proliferation and Loops. When transparent
bridges have not learned on which port a particular destination address
resides, an incoming framerpacket is submitted to each port on the bridge
except the port on which the frame was received. Consider the situation
where there is more than one bridge in the network, and none of the bridges
have learned a certain destination port for a particular framerpacket. The
framerpacket tends to proliferate with useless capacity occupation on every
LAN but the real destination LAN.
Loops can occur when a network has more than one path to a LAN station
destination address exists. This is shown in Figure 10.30. Traffic in the loop
can circulate around it indefinitely as specific traffic is routed from one
bridge to the next on the loop so formed. This phenomenon can cause a
network to crash or tie up the paths involved in such a way that they cannot
handle real traffic. To avoid such loops and proliferation, a spanning tree
algorithm is employed along with the use of configuration messages.
Figure 10.30. An example of the formation of a loop. Traffic flooding is assumed.
338
Spanning Tree Algorithm. The spanning tree algorithm allows bridges to

dynamically discover a subset of the network topology that is loop-free and
yet be able to provide sufficient connectivity so that every pair of LAN
stations on the network can communicate. To carry out this process, bridges
exchange configuration messages called configuration bridge protocol data
units or configuration BPDUs.
The configuration message contains enough information such that a single
bridge can be nominated a root bridge. Other bridges on the network can
then calculate the shortest path from themselves to the root bridge.
On each LAN, a designated bridge is selected from the bridges residing on
that LAN. The designated bridge is the one that is closest to the root bridge.
The designated bridge forwards framerpackets on the LAN toward the root
bridge. A port is then selected that gives the best path from themselves to the
root bridge. This port is called the root port.
Ports are selected on a bridge to be included in the spanning tree. The
ports selected include the root port plus any ports on which self has been
selected as designated bridge.
Ports selected for inclusion in the spanning tree forward data traffic
torfrom their connected LANs. Data traffic is never forwarded to ports that
are not selected for inclusion in the spanning tree. This traffic is discarded.
See Section 10.3.1.3 for specifics on the spanning tree algorithm.
10.3.1
IEEE 802.1 Bridges
10.3.1.1 Application of the 802.1 Bridge. IEEE 802 LANs of all types,
including FDDI, may be connected together with MAC bridges. Each individual LAN has its own independent MAC. The bridged local area network
created allows the interconnection of stations attached to separate LANs as
if they were attached to a single LAN. A MAC bridge operates below the
MAC service boundary, and is transparent to protocols operating above this
boundary, in the logical link control LLC. sublayer or the network layer.
A bridged local area network may provide the following:
a. The interconnection of stations attached to 802 LANs of different
MAC types.
b. An effective increase in the physical extent, the number of permissible
attachments, or the total performance of a LAN.
c. Partitioning of the physical LAN support for administrative or maintenance reasons.
This section describes the operation of the IEEE 802.1 MAC bridge and
highlights a selected group of its operational parameters Ref. 10..
10.3.1.2 Principles of Operation. The principal elements of 802.1 bridge
operation are the relay and filtering of frames. It includes the maintenance of
the information required to make frame filtering and relaying decisions.
A typical 802.1 MAC bridge is shown in Figure 10.31.
339
Figure 10.31. An IEEE Std. 802.1 bridge showing active topology.
A MAC bridge relays individual MAC user data frames between the
separate MACs of the bridged local area networks connected to its ports.
The order of frames of given user priority received on one port and transmitted on another is preserved.
An IEEE 802.1 bridge carries out the following functions: It forwards
received frames to other bridge ports. It discards frames in error and discards
frames to ensure that a maximum bridge transit delay is not exceeded. It
selects outbound access priority and discards frames following the application
of filtering information. The bridge maps service data units and recalculates
the frame check sequence of a frame.
The bridge carries out a filtering function. It eliminates the possibility of
end-less routing loops. This is done by eliminating frame duplication. Thus
frames transmitted between a pair of end stations can be confined to LANs
that form a path between those end stations.
340
The bridge calculates and configures a bridged local area network topology. It automatically learns dynamic filtering information through observation of bridged LAN traffic. Filtering information that has been automatically
learned is deleted. It explicitly configures static filtering information and
permanently configures reserved addresses. The bridge ages out filtering
information that has been automatically learned.
The bridge protocol entity and other higher layer protocol users, such as
bridge management, make use of the logical link control procedures. These
procedures are provided separately for each port, and they use the MAC
service provided by the individual MAC entities.
Model of Operation. Frames are accepted for transmission and delivered on
reception to and from processes and entities that model the operation of the
MAC relay entity in a bridge. These are as follows:
1. The forwarding process, which forwards received frames that are to
be relayed to other bridge ports, filtering frames on the basis of information contained in the filtering database and on the state of the
bridge ports.
2. The learning process, which by observing the source addresses of
frames received on each port, updates the filtering database conditionally on the state of the port on which frames are observed.
3. The filtering database, which holds filtering information either explicitly configured by management action or automatically entered by the
learning process, and which supports queries by the forwarding process
as to whether frames with given values of the destination MAC address
field should be forwarded to a given port.
Each bridge port also functions as an end station providing the MAC service
to LLC that supports:
1. The bridge protocol entity, which operates a MAC sublayer configuration protocol between bridges, which determines, in part, the state of
each bridge port and its participation in the active topology of the
bridged local area network.
2. Other users of LLC, such as protocols providing bridge management.
Figure 10.32 illustrates the relaying of a single frame between ports of a
bridge with two ports. Figure 10.33 shows the inclusion of information
carried by a single frame, received on one of the ports of a bridge with two
ports, in the filtering database. The bridged protocol entity operation is
illustrated in Figure 10.34. This entity operates the spanning tree algorithm
and protocol. It includes its modification of port state information as part of
determining the active topology of the bridged local area network.
341
Figure 10.32. Relaying MAC frames. (From Figure 3-4, page 33, ANSIrIEEE Std. 802.1D,
Ref. 10.)
Figure 10.33. Observation of network traffic. (From Figure 3-5, page 33, ANSIrIEEE Std.
802.1D, Ref. 10.)
342
Figure 10.34. Operation of the interbridge protocol. (From Figure 3-6, page 34, ANSLrIEEE
Std. 802.1D, Ref. 10.)
Learning Process. The learning process observes the source addresses of

frames received on each port and update the filtering database conditionally
on the state of the receiving port. Frames are submitted to the learning
process by the individual MAC entities associated with each bridge port.
The learning process may deduce the path through the bridged local area
network to particular stations by inspection of the source address field of
received frames. It creates or updates a dynamic entry in the filtering
database, associating the port on which the frame was received with the
MAC address in the source address field of the frame, if and only if:
1. The port on which the frame was received is in a state that allows
learning.
2. The source address field of the frame denotes a specific end station
i.e., is not a group address ..
3. A static entry for the associated MAC address does not already
exist.
4. The resulting number of entries would not exceed the capacity of the
filtering database.
Figure 10.32 illustrates the operation of the learning process.
It should be noted that static entries are added or removed from the
filtering database under explicit management control. They are not automatically removed by any timeout mechanism.
343
Static entries specify the following:

1. The MAC address for which filtering is specified.
2. For each inbound port on which frames are received, a port map that
specifies for each outbound port on which frames may be transmitted,
whether frames will be filtered or forwarded to that port.
These MAC addresses that can be specified include group and broadcast
addresses.
Dynamic addresses, on the other hand, are created and updated by the
learning process. They are automatically removed after a specified time
timeout. has elapsed since the entry was created or last updated. This timing
out of entries ensures that end stations that have been moved to a different
part of the bridged local area network will not be permanently prevented
from receiving frames. It also takes into account the changes on the active
topology of the bridged local area network which can cause end stations to
appear to move from the point of view of a bridge i.e., the path to those end
stations subsequently lies through a different bridge port..
The timeout value, also called aging time, after which a dynamic entry is
automatically removed, may be set by management. The recommended
default value is 300 s, and the range is 10.0 to 1.0 = 10 6 s.
The spanning tree algorithm discussed in the next subsection . includes a
procedure for notifying all bridges in the bridged local area network of
topology changes and specifies a short value for the timeout value which is
enforced for a period after any topology change.
Dynamic entries specify the following:
1. The MAC address for which filtering is specified
2. A port number
Frames with the specified destination MAC address are forwarded only to
the specified port. A dynamic entry acts like a static entry with a single port
selected in the port map.
It should be pointed out that the filtering database contains a permanent
database that provides fixed storage for static entries. The filtering database
is initialized with the static entries contained in this fixed data store. Static
entries may be added to and removed from the permanent database under
explicit management control.
Addressing. All MAC entities communicating across a bridged local area
network use 48-bit addresses, which may be universally administered or
locally administered addresses or both. Bridges may also use 16-bit locally
administered addresses.
344
UNIQUE IDENTIFICATION OF A BRIDGE. It is necessary for the operation of the

Bridge Protocol that a single unique identifier be associated with each
bridge. This identifier is derived from a unique address for the bridge, known
as the bridge address.
It is recommended that this address be the specific MAC address of the
lowest numbered bridge port Port 1..
10.3.1.3 The Spanning Tree Algorithm

Key Requirements
a. The algorithm configures the active topology of a bridged local area
network or arbitrary topology into a single spanning tree,* such that
there is at most one data route between any two end stations, eliminating data loops.
b. The entire active topology will stabilize in any sized bridged local area
network. It will, with a high probability, stabilize within a short, known
bounded interval in order to minimize the time where service is
unavailable for communication between any pair of end stations.
c. The active topology is predictable and reproducible, and may be
selected by management of the parameters of the algorithm, thus
allowing the application of configuration management, following traffic analysis to meet performance objectives.
d. It will operate transparently to the end stations such that they are
unaware of their attachment to a single LAN or a bridged local area
network regardless of the total number of bridges or LANs.
Overview
THE ACTIVE TOPOLOGY AND ITS COMPUTATION. The spanning tree algorithm
and protocol configure a simply connected active topology from the arbitrarily connected components of a bridged local area network. Frames are
forwarded through some of the bridge ports in the bridged local area network
and not through others, which are held in a blocking state. At any time,
bridges effectively connect just the LANs to which ports in a forwarding state
are attached. Frames are forwarded in both directions through bridge ports
that are in a forwarding state. Ports that are in a blocking state do not
forward frames in either direction but may be included in the active topology
i.e., be put into a forwarding state if components fail, are removed, or are
added.. Figure 10.31 shows the active topology i.e., the logical connectivity.,
of the same bridged local area network following configuration.
*The IEEE Ref. 12. defines Spanning tree as: A bridging technique where a network of
randomly interconnected bridges can automatically build a logical tree structure so as to
guarantee a unique path between any pair of stations on the network. In this scheme, the
transmitter does not have to know how to route the frame to the destination; that is the job of
the bridges.
345
Figure 10.35. Spanning tree (example). (From Figure 4-2, page 46, ANSIrIEEE Std. 802.1D,
Ref. 10.)
One of the bridges of such a configured network is known as the root or

the root bridge. Each individual LAN has a bridge port connected to it that
forwards frames from the direction of the root onto that LAN. This port is
known as the designated port for that LAN, and the bridge of which it is part
is the designated bridge for the LAN. The root is the designated bridge for all
the LANs to which it is connected. The ports on each bridge that are in the
forwarding states are the root port the closest to the root. and the designated ports if there ary any..
In Figure 10.35, Bridge 1 has been selected as the root although one
cannot tell simply by looking at the topology which bridge is the root. and is
the designated bridge for LAN 1 and LAN 2, Bridge 2 is the designated
bridge for LAN 3 and LAN 4, and Bridge 4 is the designated bridge for LAN
5. Figure 10.35 shows the logical tree topology of this configuration of the
bridged local area network.
346
The stable active topology of a bridged local area network is determined

by the following:
1. The unique bridge identifiers with each bridge port
2. The path cost* associated with each bridge port
3. The port identifier associated with each bridge port
The bridge with the highest priority bridge identifier is the root for convenience of calculation, this is the identifier with the lowest numerical value..
Each bridge port in the bridge local area network has a root path cost
associated with it. This is the sum of the path costs for each bridge port
receiving frames forwarded from the root on the least-cost path to the bridge.
The designated port for each LAN is the bridge port for which the value of
the root path cost is lowest: If two or more ports have the same value of root
path cost, then the first bridge identifier of their bridges and then their port
identifiers are used as tie-breakers. Thus, a single bridge port is selected as
the designated port for each LAN, the same computation selects the root
port of a bridge amongst the bridges own ports, and the active topology of
the bridged local area network is completely determined.
A component of the bridge identifier of each bridge, and the path cost and
port identifier of each bridge port, can be managed, thus allowing a manager
to select the active topology of the bridged local area network.
PROPAGATING TOPOLOGY INFORMATION. Bridges send a type of bridge protocol data unit known as a configuration BPDU to each other to communicate
and compute the above information. A MAC frame conveying a BPDU
carries the bridge group address in the destination address field and is
received by all bridges connected to the LAN on which the frame is
transmitted.
BPDUs are not directly forwarded by bridges, but the information in them
may be used by a bridge in calculating its own BPDU to transmit and may
stimulate the transmission. The configuration BPDU, which is conveyed
between bridge ports attached to a single LAN, is distinguished from the
notion of a configuration message, which expresses the propagation of the
information carried throughout the bridged local area network.
Each configuration BPDU contains, among other parameters, the unique
identifier of the bridge that the transmitting bridge believes to be the root,
the cost of the path to the root from the transmitting port, the identifier of
the transmitting bridge, and the identifier of the transmitting port. This
information is sufficient to allow a receiving bridge to a. determine whether
the transmitting port has a better claim to be the designed port on the LAN
on which the configuration BPDU was received than the port currently
*We call cost a routing parameter that determines shortest path route.
347
believed to be the designated port and b. determine whether the receiving

port should become the root port for the bridge if it is not already.
Timely propagation throughout the bridged local area network of the
necessary information to allow all bridge ports to determine their state i.e.,
blocking or forwarding. is achieved through three basic mechanisms:
1. A bridge that believes itself to be the root all bridges start believing
themselves to be the root until they discover otherwise. will originate
configuration messages by transmitting configuration BPDUs. on all
LANs to which it is attached, at regular intervals.
2. A bridge that receives a configuration BPDU on what it decides is its
root port conveying better information i.e., highest priority root identifier, lowest root path cost, highest priority transmitting bridge and port.
will pass that information on to all the LANs for which it believes itself
to be the designated port.
3. A bridge that receives inferior information on a port it considers to be
the designated port on the LAN to which it is attached will transmit its
own information in reply for all other bridges to the LAN to hear.
Hence, spanning tree paths to the bridge with highest priority root
identifier are quickly learned throughout the bridged local area network, with
inferior information about other potential roots and paths being contradicted.
Figure 10.36 shows the various port states of a bridge and the transitions
between the port states.
SELECTED KEY BRIDGE PARAMETERS
Root Path Cost. The cost of the path to the root from this bridge. It is equal
to the sum of the values of the designated cost and path cost parameters
held for the root port. When the bridge is the root, this parameter has the
value zero. The parameter is used to test the value of the root path cost
parameter conveyed in received configuration BPDUs; it is also used as
the value of the root path cost parameter offered in all configuration
BPDUs transmitted by the bridge.
Root Port. The port identifier of the port that offers the lowest cost path to
the root i.e., that port for which the sum of the values of the designated
cost and path cost parameters held for the port is lowest..
Max Age. The maximum age of received protocol information before it is
discarded.
Hello Time. The time interval between the transmission of configuration
BPDUs by a bridge that is attempting to become the root or is the root.
Bridge Hello Time. The value of the hello time parameter when the bridge is
the root or is attempting to become the root. The time interval between
348
Figure 10.36. Port states. (From Figure 4-3, page 48, ANSIrIEEE Std. 802.1D, Ref. 10.)
transmissions of topology change notification BPDUs toward the root

when the bridge is attempting to notify the designated bridge on the LAN
to which its root port is attached of a topology change. This parameter
may be updated by management action.
Port ParameterPath Cost. The contribution of the path through this port,
when the port is the root port, to the total cost of the path to the root for
this bridge. The parameter is used, added to the value of the designated
cost parameter for the root port, as the value of the root path cost
parameter offered in all configuration BPDUs transmitted by the bridge,
when it is not the root. This parameter may be updated by management
action.
Performance. The IEEE 802.1D specification recommends default operational values for performance parameters. These have been specified in order
to avoid the need to set values prior to operation, and they have been
selected with a view of optimizing the ease with which bridged local area
network components interoperate. It specifies absolute maximum values for
performance parameters. The ranges of applicable values specified to assist
in the choice of operational values and to provide guidance to implementors.
349
REQUIREMENTS. For correct operation, the parameters and configuration of

bridges in the bridge local area network ensure the following:
1. Bridges do not initiate reconfiguration if none is needed. This means

that a bridge protocol message is not timed out before its successor
arrives, unless a failure has occurred.
2. Following reconfiguration frames are not forwarded on the new active
topology, while frames that were initially forwarded on the previous
active topology are still in the network. This ensures that frames are not
duplicated.
These requirements are met through placing restrictions on the following:
Maximum bridge diameter expresses the maximum number of bridges
between any two points of attachment of end stations. The recommended
value is 7.
Maximum bridge transit delay in seconds. is the maximum time elapsing
between reception and transmission by a bridge of a forwarded frame, frames
that would otherwise exceed this limit being discarded. It consists of three
components for which we express the recommended value and the maximum
value:
Maximum bridge transit delay:
Maximum BPDU transmission delay:
Maximum message age increment overestimate:
1.0 s
1.0 s
1.0 s
4.0 s
4.0 s
4.0 s
Spanning Tree Algorithm Timer Values:

Bridge hello time:
Bridge max age:
Bridge forward delay:
Hold time:
2.0 s range 1.010.0.

20.0 s range 6.040.0.
15.0 s range 4.030.0.
Fixed value of 1.0
Priority Parameter Values (recommended or default values)

Bridge priority:
Port priority:
32,768
128
range 065535.
range 0255.
Path Cost Parameter Values

Path cost s 100rattached LAN speed in Mbps.
Example:
10 Mbps LAN, path cost s 100
Absolute minimum path cost s 1 range 165535.
for each bridge port
350
Granularities are in 1-s increments.

A bridge must enforce the following relationships:
2 = bridge forward delay y 1.0 s. G Bridge Max Age
Bridge Max Age G 2 = Bridge Hello Time q 1.0 s.
10.3.1.4 Bridge Protocol Data Unit (BPDU). Figure 10.37 illustrates the
format of a configuration BPDU.
Each transmitted configuration contains the following parameters and no
others.:
1. The protocol identifier is encoded in octets 1 and 2 of the BPDU. It
takes the value 0000 0000 0000 0000, which identifies the spanning
tree algorithm and protocol as specified in Section 4 of this standard.
Figure 10.37. Configuration BPDU, parameters and

format. (From Figure 5-1, page 103, ANSIrIEEE Std.
802.1D, Ref. 10.)
351
2. The protocol version identifier is encoded in octet 3 of the BPDU. It

takes the value 0000 0000.
3. The BPDU type is encoded in octet 4 of the BPDU. This field shall
take the value 0000 0000. This denotes a configuration BPDU.
4. The topology change acknowledgment flag is encoded in bit 8 of octet
5 of the BPDU.
5. The topology change flag is encoded in bit 1 of octet 5 of the BPDU.
6. The root identifier is encoded in octets 6 through 13 of the BPDU.
7. The root path cost is encoded in octets 14 through 17 of the BPDU.
8. The bridge identifier is encoded in octets 18 through 25 of the BPDU.
9. The port identifier is encoded in octets 26 and 27 of the BPDU.
10. The message age timer value is encoded in octets 28 and 29 of the
BPDU.
11. The max age timer value is encoded in octets 30 and 31 of the BPDU.
12. The hello time timer value is encoded in octets 32 and 33 of the
BPDU.
13. The forward delay timer value is encoded in octets 34 and 35 of the
BPDU.
ENCODING OF BRIDGE IDENTIFIERS. A bridge identifier is encoded as 8 octets,
taken to represent an unsigned binary number. Two bridge identifiers may be
numerically compared; the lesser number denotes the bridge of the higher
priority.
The two most significant octets of a bridge identifier comprise a settable
priority component that permits the relative priority of bridges to be managed. The six least significant octets ensure uniqueness of the bridge identified. They are derived from the globally unique bridge address according to
the following procedure.
The third most significant octet is derived from the initial octet of the
MAC address, the least significant bit of the octet Bit 1. is assigned the
value of the first bit of the bridge address, the next most significant bit is
assigned the value of the second bit of the bridge address, and so on. In a
bridged local area network utilizing 48-bit MAC address, the fourth through
the eighth octets are similarly assigned the values of the second to the sixth
octets of the bridge address.
In a bridged local area network utilizing 16-bit MAC addresses, the fourth
octet is assigned the value of the second octet of the bridge address, and the
fifth through eight octets are assigned the value 0000 0000.
ENCODING OF ROOT PATH COST. Root path cost is encoded as four octets,
taken to represent an unsigned binary number, a multiple of arbitrary cost
units. Recommendations for these values are given in Section 10.3.1.3 herein.
352
ENCODING OF PORT IDENTIFIERS.
A port identifier is encoded as 2 octets,

taken to represent an unsigned binary number. If two port identifiers are
numerically compared, the lesser number denotes the port of higher priority.
The more significant octet of a port identifier is a settable priority component that permits the relative priority of ports on the same bridge to be
managed. The less significant octet is the port number expressed as an
unsigned binary number. The value 0 is not used as a port number.
ENCODING OF TIMER VALUES.
Timer values are encoded in 2 octets, taken to

represent an unsigned binary number multiplied by a unit of time of 1r256 of
a second. This permits times in the range 0 to, but not including, 256 s to be
represented.
Sections 10.3.1 through 10.3.4 consist of abstracts of IEEE Std. 802.1D, 1993, edition,
Ref. 10.
10.3.2
Source Routing Bridges
The basic concept behind source routing is that a frame header contains
routing information that has been inserted by the source of the frame. How
must a source end station know its route? It does not. So it must discoer a
route. It can do this by transmitting a special kind of frame. This frame
replicates itself at each node where the route splits i.e., where there are two
or more routing choices for onward delivery.. As a frame travels the route
toward its destination, it accumulates routing information such that upon its
arrival, future routing can be selected. This routing information is stored in
cache so it can serve subsequent frames.
Source Routing Header and the Routing Information (RI) Field. The generic
MAC frame for a LAN contains, as a minimum, a destination address field, a
source address field, and a payload, variously called info field, text, data, and
so on. This simplified header is
The object of this exercise is to show how an RI field may be added to a

MAC header. To tell a receiving bridge that the rules have been changed, we
take advantage of the first bit in the source address field see Section 10.3.2..
This bit has been defined as the individualrgroup address field. In standard
MAC frames, this bit is always carried as a binary 0. To tell a receiving bridge
353
that a frame carries an RI field, the bit is set to binary 1 as follows:
The RI field contains the following source routing information Ref. 11.:
Type
Length
Direction
Largest
frame
Route
3 bits
5 bits
1 bit
3 bits
16 bits
Type 3 bits.. The type of source routing may be one of the following:
a. Specifically routed the route appears in the header.
b. All paths explorer the route information accumulates as a frame
traverses the network.
c. Spanning tree explorer the route information accumulates as frame
copies traverse the network just as with all path explorers, but the
frame only travels along branches of the spanning tree.
Length 5 bits.. Specifies the number of octets in the RI field.
Direction 1 bit.. Specifies whether the route should travel from right to
left or left to right.
Largest Frame LF. 3 bits.. A value representing one of a few popular
frame sizes 516, 1500, 2052, 4472, 8144, 11407, 17800, or 65535 octets..
Route. A sequence of 2-octet-long fields which are called route designators, each of which consists of a 12-bit LAN number followed by a 4-bit
bridge number as illustrated:
LAN
Bridge
12 bits
4 bits
Bridge numbers are required to resolve any ambiguity of multiple

frames. Within an extended LAN, each LAN must be assigned a unique
LAN number. An extended LAN is assumed to have multiple bridges.
As a frame travels across the network, it replicates itself at every bridge
outlet port and the number of copies of a frame grows exponentially.
Each bridge is assigned a 4-bit identifying number to distinguish it from
other bridges that connect the same pair of LANs. Thus there is a LAN
number for each port plus a bridge number. The assignment of a bridge
number must follow the rule that no two bridges interconnecting the same
two LANs be allowed to have the same bridge number.
354
That important route is just an alternating sequence of LAN numbers and

bridge numbers, always starting and ending with a LAN number.
From the viewpoint of a bridge, there are four types of frames to be
handled:
1.
2.
3.
4.
Frames without an RI field. These are known as transparent frames.

Specifically routed frames.
All paths explorer frames.
Spanning tree explorer frames.
10.3.3
Remote Bridges
Remote bridges are used to connect two geographically distant LANs. In this
case, a bridge is installed on each LAN and a wide area network WAN.
connects the two bridges. It is generally assumed that these two bridges will
be bought in pairs to avoid interface problems.
The simplest way of implementing a remote bridge configuration is to
entirely encapsulate the LAN data frame with the WAN header and trailer,
often an FCS..* The remote bridge encapsulation concept is shown in
Figure 10.38. Here an entire CSMArCD frame is encapsulated in a frame
relay leader. Note that the CSMArCD preamble has been stripped off. It is
added back on at the remote bridge at the other side of the WAN link.
The spanning tree works well with the point-to-point WAN, and the two
bridges on each side consider the WAN link as another LAN. One of the
two bridges will be the designated bridge on the point-to-point link. The
second bridge, then, considers the port onto that link as its root port.
10.4
10.4.1
HUBS AND SWITCHING HUBS

Rationale and Function
The original idea of a hub was to centralize equipment, usually in the wiring
closet on a floor. We think of a LAN as a set of operational workstations
PCs. connected to some sort of wire or cable stretching around the floor. It
Figure 10.38. Remote bridging concept, WAN encapsulation. CSMA r CD encapsulated in

frame relay.
*Note: RFC 1483 discusses the application of a remote bridge by means of ATM; RFC 1490, by
means of frame relay.
HUBS AND SWITCHING HUBS
355
Figure 10.39. The star topology idea of a hub. The hub, of course, sits in the center. The
lines connecting workstations to the hub represent wire twisted pairs or a continuous length
of coaxial cable tapped at the workstation.
was found to be more practical to steal a page from the book on telephony in
that each telephone is terminated on a floors mainframe located in the
wiring closet. Why couldnt we do the same for each computer or workstation? Thus we ended up with a star topology illustrated in Figure 10.39.
One advantage of the hub concept was centralization; another was that
LAN equipment could be locked up in a room, away from unauthorized
hands. To improve performance, each hub port termination was a repeater.
In the case of CSMArCD, the problem of collisions remained unchanged.
10.4.2
Improvements in Hub Technology
An early improvement in hubs was that of filtering frames based on sensing

of the destination address. Here a frame would only be transmitted out a hub
spoke if that frame was addressed to the LAN station attached to that
spoke. Of course the hub must learn which stations attach to which spoke
based on the source address of frames emanating from a particular spoke.
Store-and-forward techniques were employed. Note that if there are multiple
hubs or hubs plus bridges in the LAN network, the spanning tree algorithm
must be employed to avoid disastrous traffic loops.
With this type of CSMArCD setup, the chance of collision is eliminated
because the active link is isolated i.e., there are only two stations on the link
in the case of unicastthe hub port processorrrepeater and the receiving
LAN station.. Another advantage is that ports on a hub can be at different
speeds e.g., 10 Mbps and 100 Mbps..
The introspective reader will ask, What is the difference between a hub of
this type and a learning bridge? None. In fact, many in the industry call this
type of hub a layer 2 switch.
356
One also may note the delay incurred in the store of store and forward
switching. One cure for the delay is to use cut-through forwarding. All this
means is that the layer 2 switch can forward as soon as it has made a
forwarding decision. This happens sometime after all the bits of the destination address are received. Another name for a layer 2 switch is a smart hub.
10.5
ROUTERS
As illustrated in Figure 10.40, a router bases its routing decisions on the

addressing found in OSI layer 3, the network layer. So much of this discussion deals with addressing, particularly in layer 3. Because the routing is
based on OSI layer 3, well just call it layer 3 switching.
As we have learned in data communications, we can have connection-oriented service or connectionless service. Using real-life analogies: A telephone
call can be related to connection-oriented service. To make a call, a circuit
must first be set up between the call initiator and its recipient. The circuit
setup is based on the telephone number of the called party. Then conversation is exchanged and carried on until one or the other party hangs up; the
call is terminated and the circuit is taken down. The analogy of connectionless service is the mailing of a letter through the postal service. A letter is
placed in the outgoing mail slot in the Post Office. Some days later we hope.
the letter arrives at its destination. It relies on the address on the envelope
for its proper delivery.
A router works on the connectionless service concept. A frame is
launched into the system. We hope that by means of its destination address,
it will arrive at the recipient uncorrupted and in some reasonable length
of time.
When discussing addressing, ask Just what does an address identify?:
a node, an interface, a switch, a LAN station, an end-user, or a service access
point SAP. of an end-user?
10.5.1
Addressing with Hierarchical Significance
As with telephone numbering, the addressing at layer 3 can have geographical or other. significance. In other words, the addressing scheme is hierarchical. All this means that a network address is broken down in identifiable
chunks. This can mean that a chunk or piece of the address tells us where
the address resides.
Worldwide telephone numbering is based on pieces. In North America,
telephone numbering is a mixed bag. Certainly, area codes have geographical
significance. Exchange codes the 3-digit number group following the area
code. do not; nor do the 4-digit subscriber codes, the last 4 digits of a
telephone number that identify an individual subscriber. In other countries,
the exchange code can have geographical significance.
ROUTERS
357
Figure 10.40. The subnet address field in relation to the total IP address.
All network layer addresses are hierarchical. However, for the purpose of
routing, IEEE 802 addresses are not hierarchical, even though they seem so.
The IEEE uses a block of 3 octets in its 802 series addressing for the OUI
organizationally unique identifier .. It is not geographically hierarchical.
Certain network types, such as ISDN and frame relay, mix voice calls with
data calls. They may be governed by one of two ITU-T Recommendations:
E-164, which is more inclined to the voice telephone world, and X.121 for the
data world. For example, X.121 addressing numbering. consists of two
chunks or pieces: the Data Network Identification Code DNIC., where the
first three digits of the DNIC is the DCC data country code.. The second
part consists of the remaining 10 or 11 digits maximum. which contains the
network terminal number or the national number which is part of the
national numbering scheme. Of course, this addressing will be used for wide
area networks.
IP has its own addressing scheme, and there is a scheme for IP4* and
another for IP6. For this discussion we will stick with IP4 or just IP. It has a
source address and a destination address, and each is 32 bits long 4 octets..
Its address, whether source or destination, consists of two pieces or chunks:
network address and host address. Some texts refer to the host address as the
local address. Refer to this simple drawing:
Link
Host
32 bits
The link in the field indicates a link number, and the second subfield
indicates a system on the link. IP people call the second piece or chunk
a host. The actual number of bits in the link and host fields is not actually
fixed Ref. 11.. However, the total address length is 32 bits IP4.. See
Section 11.4.3.
Another way of looking at an IP address is shown in Figure 10.40. This
shows the subnet address structure in relation to the total address field.
Assignments for the local address is left up to the local network administrator. For instance, for the 16 bits assigned to the local address, half may be
assigned to the subnetwork and half to the host; or 12 bits can be assigned to
the subnetwork and only 4 bits assigned to the host.
*IP4 is currently the most widely used version of IP. It is so widely used that many people just
call it IP, where they assume everybody knows it is IP4.
358
10.5.2
Subnet Masks
For each link in an IP network, the network planner decides the value of the
link number and which bits in the address will correspond to the link
number. A link a LAN, for example. has a link number and a mask
indicating which bits correspond to the link number. The purpose of the
mask is to determine which part of the IP address pertains to the subnet and
which pertains to the host.
According to Black Ref. 13., the convention used for subnet masking is to
use a 32-bit field in addition to the IP address. The contents of the mask field
are set as follows:
Binary 1s: Identify the network address portion of the IP address.
Binary 0s: Identify the host address portion of the IP address.
A bitwise AND function is used to extract the fields of the IP address as
follows.
A bitwise AND is performed on the IP address and subnet mask. The
results of this operation are matched against a destination address in a
routing table. If the results are equal, the next hop IP address relative to the
destination address . is used to determine the next hop on the route. Therefore, the last octet all 0s. identifies the host on the subnet.
The mask becomes part of the routing algorithms conditional statement:
If the destination IP address and the subnet mask equal my IP address and
subnet mask, then send the datagram to local network; otherwise, send
datagram to gateway corresponding to destination address. Black comments
that, indeed, the use of masks handles routes to direct conventions, hostspecified routes, and default routes. The reference suggests that the implementor should note the following guidelines for subnetting:
10.6
The IP algorithm must be implemented on all machines in a subnet.

Subnet masks should be the same for all machines.
If one or more machines do not support masks, proxy ARP address
resolution protocol; see Section 11.4.4.1. can be used to achieve
subnetting.
VIRTUAL LOCAL AREA NETWORKS (VLANs)
VLANs are not well-defined by the industry. Reference 11 defines a VLAN

as the territory over which a broadcast or multicast frame packet. is
delivered also known as a broadcast domain.. The difference between a
VLAN and a LAN, if any, is in packaging. A virtual LAN allows the use of
SERVERS AND INTRANETS
359
separate LANs among ports on the same switch. It breaks a switch up into
two or more bridges. where ports 141 belong to VLAN X and ports 4273
to VLAN Y. In this case, the switch would be broken up into two separate
bridges: one that forwarded between the first 41 ports, and one that covers
the remainder of the ports. A router would be employed to forward traffic
from one VLAN to the other. Of course the separate router connecting the
two VLANs requires one port from each bridge. Some switches are capable
of acting as routers between VLANs.
When two switches are involved in a VLAN environment with a connecting link between the switches, frames can belong to either VLAN if there
are only two.. To solve this situation, additional information is provided
called a VLAN tag, which is added to the frame packet. so that switches can
know to which VLAN frame or other frame is intended. A station would be
confused if it received the frame with a VLAN tag. So switches must be
configured to know which ports connect to switches and which ports connect
to stations. A switch removes the VLAN tag from the frame before forwarding the frame to a non-switch-neighbor port.
A VLAN tag contains 2 octets where 3 bits are assigned a priority
function, 12 bits for a VLAN ID, and 1 bit indicating the addresses are in
canonical format. In IEEE 802.3 LANs, the VLAN tag is inserted just before
the length field.
10.7
10.7.1
SERVERS AND INTRANETS

Servers
A serer is a data device that facilitates. The IEEE Ref. 12. calls it a device
or computer system that is dedicated to providing specific facilities to other
devices attached to the network. There are many examples. A network serer
provides special network applications to all users at the network level; that is,
it allows all users to share files more easily and have larger file storage areas
available for their use. Commonly, network servers also carry out a policing
function where all users check in through the server that validates passwords
and establishes firewall. It often is the heart of a LAN, such that when the
server is down, the network is down.
Other servers are more specialized for example, client serer, where the
client is the requesting device, and the server is the supplying device.
Another example, is where the user interface could reside in the client
workstation while the storage and retrieval functions could reside in the
server database. Another case is the terminal serer which is a device that
permits users to access a central mainframe computer. Still another example
360
is the print serer. This is a device that is dedicated to queueing and sending
printer output from the networked computers to a shared printer.
We also have database servers, disk servers, file servers, and mail servers.
10.7.2
Intranet and Extranet
An intranet is an enterprises internal network. Intranets can communicate

with each other via the intranet, which provides the backbone communication. However, an internet does not need an outside connection to the
internet in order to operate. It simply employs TCPrIP protocols and
applications to provide a private internet.
When a business exposes part of its internal network to the outside
community, it is known as an extranet. Some of this internet exposure may be
a companys own web page. Outsiders i.e., the general public. will not have
access to the entire corporate network, but merely that part that the company
wants the public to have access to. The enterprise blocks access in the
networks routers and places firewalls on resources the enterprise may consider sensitive, forcing the public to have access only to a subset of its
intranet.
REFERENCES
1. Information Processing SystemsLocal Area NetworksPart 4: Token-Passing Bus
Access Method and Physical Layer Specification, ANSIrIEEE Std. 802.4-1990,
IEEE, New York, 1990.
SystemsLocal and Metropolitan Area NetworksSpecific RequirementsPart 5:
Token Ring Access Method and Physical Layer Specifications, ANSIrIEEE Std.
802.5, 1995 edition, IEEE, New York, 1995.
3. Fiber Distributed Data Interface FDDI . Physical Layer Protocol PHY-2 ., ANSI
X3.231-1994, ANSI, New York, 1994.
4. Fiber Distributed Data Interface FDDI . Token Ring Physical Layer Medium
Dependent, ANSI X3.166-1990, ANSI, New York, 1990.
5. Fiber Distributed Data Interface FDDI . Token Ring Media Access Control2
MAC-2 ., ANSI X.239-1004, ANSI, New York, 1994.
SystemsLocal and Metropolitan Area NetworksSpecific RequirementsPart 11:
Wireless LAN Medium Access Control MAC . and Physical Layer PHY . Specifications, IEEE Std. 802.11-1997, IEEE, New York, 1997.
7. Code of Federal Regulations: CFR 47Telecommunication, Part 15, U.S. Government Printing Office, Washington DC, October 1996.
8. Information TechnologyTelecommunication and Information Exchange Between
SystemsLocal and Metropolitan Area NetworksSpecific Requirements: Part 11:
REFERENCES
9.
10.
11.
12.
13.
361
Wireless LAN Medium Access Control MAC . and Physical Layer PHY . Specifications, ANSIrIEEE Std. 802.11, IEEE, New York, August 1999.
Roger L. Freeman, Radio System Design for Telecommunications, 2nd edition,
John Wiley, New York, 1997.
Information TechnologyTelecommunications and Information Exchange Between
Systems Local Area Networks Media Access Control MAC . Bridges,
ANSIrIEEE Std. 802.1D, 1st edition, IEEE, New York, July 1993.
Radia Perlman, Interconnections, 2nd ed., Addison-Wesley, Longman, Reading,
MA, 2000.
The IEEE Standard Dictionary of Electrical and Electronic Terms, 6th ed., IEEE
Std. 100, IEEE, New York, 1996.
Uyless Black, TCPrIP and Related Protocols, McGraw-Hill, New York, 1992.
11
WIDE AREA NETWORKS (WANs)
11.1
BACKGROUND AND SCOPE
Early data networks consisted of point-to-point or multipoint circuits. A star

topology with a computer at the stars hub was also widely employed. When
the personal computer PC. gained wide acceptance, distributed processing
became the rule. As a result, data networks and data networking became
much more complex. We now see network architectures such as intricate
tree, multiple star and mesh topologies, and nodes with multiple inlets
and outlets.
Data rates were based on the limitations of the analog network or roughly
limited to 9600 bps ca. 1985.. To efficiently transport data in such an
environment, packet switchingrtransmission came about. Meanwhile, in the
late 1990s, data rates notably increased to 56 kbs or thereabouts on a
subscriber loop, DSL digital subscriber line. lines displayed rates well in
excess of 1 Mbps, and trunk rates were and are. based on the 64-kbps
channel, T1rE1 rates, and very high SONETrSDH data transmission rates.
The preponderance of packet* networks is based on the ITU-T Rec. X.25
protocol. However, X.25 networks have greater acceptance in Europe than in
the United States and Canada. As we will see, packet networks gave a
perception of being slow and tedious, especially considering the demand for
LAN interconnectivity where data rates are well into the megabitrsecond
range, whereas X.25 supported transmission rates in the kilobit region.
This chapter provides design data and brief descriptions of the comparatively mature technology involved in wide area data. networking as a whole.
A brief review is given on the concept of statmultiplexers and the X.25
protocol. Much of the chapter deals with the ever popular TCPrIP. It was
*We must be careful of language here. There seems to be a general underlying carelessness of
language with data communications. Certainly, we can interpret TCPrIP circuits as carrying
packets, although we may call them datagrams. There are other examples.
Statmultiplexer is the shortened term for statistical multiplexer.
363
364
the authors decision to cover TCPrIP under wide area networks. It could
have just as well been included in Chapter 10. IP version 6 is then reviewed,
followed by two transport protocols: TCP and User Datagram Protocol
UDP.. This is followed by a short discussion of CLNP Connectionless
Network Protocol.. The concept of VSAT very small aperture terminal.
satellite data networks is covered, as well as a short discussion of the ITU-T
Hypothetical Reference Connection.
11.2
11.2.1
BASIC APPROACHES
Point-to-Point Links
11.2.1.1 Economy Versus Data Perishability. Perishability means how

soon must the distant end-user have the data. A credit check for a credit card
purchase is very perishable. The purchaser wants an immediate reply. Many
payrolls, on the other hand, can take much longer time i.e., hours or days.,
so long as payday is on time.
The most economical way of shipping data that is not perishable is by U.S.
mail. If overnight is required, U.S. mail, Federal Express, UPS, and others
are there to do the job.
When perishability increases, some form of electrical data communications
is probably in order, say when the data are needed in under 12 hours. The 12
hours drop to seconds for interactive data, such as the credit card credit
check. Often this can be accomplished with a dial-up connection using
an acoustic coupler or other dial-up device. Such a concept is shown in
Figure 11.1.
This form of point-to-point data connectivity can be very cost-effective if
there are only a few transactions per day of fairly short durationon the
order of minutes, not hours. Often the connection is half-duplex and we are
relegated to using stop-and-wait ARQ. This is a cost-effective and efficient
approach if there are not many transactions.
Figure 11.1. A dial-up data connection.
BASIC APPROACHES
365
Figure 11.2. A typical dedicated (leased) point-to-point data circuit. Ir O, inputroutput

device. DPTE, data processing terminal equipment.
As the number of transactions increases and possibly the duration of each

transaction increases, a dedicated circuit may be more effective. This should
be justified by a simple cost analysis. A call to the business office of the local
telephone company would be helpful; however, if the transactions will be
interstate or international, a call to ATTrSprintrMCI or others would be
in order.
11.2.1.2 Dedicated Circuits. Figure 11.2 illustrates the concept of a
point-to-point dedicated connection. Expect it to be a leased facility. It is
nearly always 4-wire. This means that we can expect equal service in either
direction simultaneously. It is ours 24 hours a day, 365 days a year. It can be
quite costly. Can we justify the cost?
The demand for its use will ordinarily be through the conventional
workday, 5 days a week. This investment may lay dormant after 5 P.M. until 8
A.M. and on weekends and holidays. There may be ways of prioritizing service
so that less perishable data can wait until these dormant periods. In many
instances, we may not keep this leased connection occupied even during
normal working hours.
One way to increase utilization is to share the line with others in our own
organization, locally or elsewhere. Convert the service from point-to-point to
multipoint. Now we have to set out some rules for who gets to use the service
and when. A representative multipoint multidrop. arrangement is shown in
Figure 11.3.
Figure 11.3. A multipoint arrangement for a leased (dedicated) line (4-wire).
366
If there are several users on one line, contention might serve using CSMA
carrier sense multiple access, meaning listen before use.. Polling* is
another alternative. Some scheme of error control would also probably be in
order. A link-layer protocol such as HDLC described in Chapter 3 would
solve most of these issues.
Frame relay, which is covered in Chapter 12, might be another alternative
if we were concerned with speeding this up. Such an alternative become
viable if the leased service provides 64 kbps and multiples thereof. channelization. This type of channelization is described in Chapter 7. Care must be
taken here in that T1 DS1. PCM in North America may only provide a clear
56-kbps channel due to the signaling bit appearing in every sixth frame.
Often clear 64-kbps channels must be specially ordered. European E1
systems do not suffer this shortcoming because signaling for an entire
30-channel group appears on a separate channel i.e., Channel 31..
11.2.2
Data Multiplexers and Statmultiplexers
11.2.2.1 Time Division Multiplexing. We are dealing here with multiple

users sharing a single channel. In the case of the analog channel, there are
two approaches. The first, which predates World War II, involves a voice
channel that is broken down into frequency slots each 120 Hz wide. Using
frequency shift keying FSK. modulation, each frequency slot can support
75-bps operation. In theory, 25 such channels can be accommodated in the
analog voice channel 3003400 Hz., but more often only 24 channels are
carried. The aggregate bit rate would be 24 = 75 s 1800 bps.
The other approach is to use time division rather than frequency division.
multiplexing. Here each user would have a fixed time slot. The bit rate that is
divided up among the users is 2400, 4800, or more probably 9600 bits per
second. Assume again 24 users, each with a fixed time slot. Thus 9600r24 s
400 bits, or each user would have a 400-bit time slot each second. Of course
in this simple calculation, no overhead bits are added. Some framing bits
would be needed as well as some way of identifying each users channel so
reconstitution could be carried out at the other end of the circuit. So from a
gross bit rate standpoint, time division multiplex is much more efficient than
multiplexing in the frequency domain. The TDM data concept is illustrated
in Figure 11.4.
11.2.2.2 Statistical Time Division Multiplexing. In a data TDM configuration, a time slot is allocated to each user. In many typical applications, data
traffic is bursty. Here we mean that there is a burst of traffic as in a credit
card verification system, followed by a quiescent period of some duration. In
*Polling is a method of establishing discipline on a multiuser circuit. Each station is polled in
sequence allowing them to have their turn transmitting traffic. Token passing is an outgrowth of
polling.
PACKET NETWORKS FOR DATA COMMUNICATIONS
367
Figure 11.4. Time division multiplexing of data. Only 12 users are shown.
fact, Doll Ref. 1. states that often individual time-slot utilization is not much
more than 10%.
A statistical multiplexer statmux. improves utilization by dynamically
assigning time-slot duration to active users. As one would imagine, some
overhead is required to provide the distant end with status information and
time-slot assignment in nearly real time. A notable increase is required in
buffering to avoid overflows, especially when several incoming traffic bursts
coincide. A queueing regime is also required.
Doll Ref. 1. reports the number of users can be increased from two to
four times compared to the equivalent fixed time-slot TDM device. It is
assumed with such increases that the typical user has a bursty traffic profile.
11.3
11.3.1

Introduction to Packet Data
In Section 11.2 we found that point-to-point and often multipoint links have
periods of inactivity and that they tended to be underutilized. The circuitswitched public telephone network is a prime example of a network that is
underutilized and inefficient. It is sized or dimensioned for the busy hour, a
1-hour period, usually in the morning of a workday, when the network has
the highest traffic activity. After the workday is over, traffic starts to really
368
drop off, and at 2 A.M. it may be 1r500 of what it was during the busy hour.
The design criterion is a 99% probability that a user can successfully
complete a call during the busy hour.
Store-and-forward systems, typically Telex, are very efficient with excellent
utilization factors, even at 2 A.M. However, when a message is submitted
there is no guarantee when it will arrive. Prioritization schemes help. Generally, links connecting store-and-forward nodes can boast nearly 100% utilization, even in off-peak hours.
Packet networks attempt to reach a happy medium: to be efficient with a
high utilization factor and have a good delivery time record, providing nearly
as good service as circuit switching can provide. They were also designed to
better accommodate bursty traffic, typical of a data connectivity.
With the packet concept, data messages are broken down into fixed-length
packets through the process of segmentation on one end of the circuit and
reassembly on the other end. On a particular circuit, packets of many
different users may be transported. Each packet is individually addressed and
transits the network in a store-and-forward fashion. When a packet is
transmitted, the circuit is immediately available for allocation of packets from
other users.
Each node is a packet network has quasi-real-time information on the
status of the network and, in theory, will route a packet on an optimum route
toward its destination. Packets are routed around failed nodes and sectors of
congestion.
Virtual Circuits and Logical Channels. When dealing with a packet network,
we must understand the terms irtual circuit and logical channel. When a
request is received to enter packets into the network, a virtual call is
established. It is something like a dialed switched connection in the telephone network. The connection exists only for the duration of the call. For
that period the network behaves as though there were a fixed path from the
source to the destination. However, in an ideal packet network, each packet
may follow a different route through the network.
For users with high traffic volumes, a permanent virtual circuit PVC. can
be established. This path is similar to a dedicated circuit, eliminating repeated call setup and release procedures.
The concept of a irtual circuit versus a real circuit. we could say derives
from the digital network which is a TDM network. For example, in a DS1
configuration Chapter 7., a caller has use of a circuit for only 1r24th of the
time; for an E1 configuration, 1r32nd of the time. In the analog network, a
user has a physical connectivity 100% of the time.
For packet networks and for ATM networks, logical channel numbers are
used as a quasi-routing tool. Each packet entering the network is assigned a
logical channel number that indicates the session or conersation to which it
belongs. The network can then associate the logical channel number of one
end user and the logical channel number of another. CCITT notes in their
369
definition that a number of logical channels may be derived from a data link
by packet interleaving.
Certainly the most well-known and popular specification for a packet data
network is based on ITU-T Rec. X.25. A general overview of the Recommendation is given in Section 11.3.2.
11.3.2 Packet-Switched and Virtual Connection Based
on ITU-T Rec. X.25
11.3.2.1 Introduction. The X.25 protocol Ref. 2. specifies the rules for
communication between the data user the DTE. and the entry to the public
switched network serving node.. In practice, the electrical interface between
the user and the network is located between the DTE and DCE. The DCE
may be a modem or digital service unit DSU. which connects to the nearest
network node by a transmission line or other media.. The intelligent part of
the X.25 protocol is handled by the DCE and the DSE data switching
equipment. of the node to which the DTE is connected.
The X.25 protocol does not define the internal operation of the data
network i.e., how the packets are routed between various nodes of the
network.. X.25 does precisely define the interface between the network and
the user in such a way as to permit all X.25 DTEs to be connected to any
public data network that conforms to the protocol. X.25 defines three
protocol layers that correspond approximately to the first three layers of the
open system interconnection OSI. reference model. The X.25 architecture is
Figure 11.5. X.25 architecture related to the OSI reference model.
370
Layer 1 is defined by CCITT Rec. X.21 Ref. 3. Section 11.3.2.2. for a

digital network, by X.21 bis for an analog entry point, and by X.31 Ref. 4.
for ISDN connectivity. Bit rates are supported all the way up to 2.048 Mbps.
Layer 2 of X.25 defines the point-to-point transfer of frames between the
DTE and DCE. This is the data-link layer and is specified by the LAPB link
access procedure balanced .. It is almost identical to HDLC as described in
Chapter 3.
The third layer of X.25 concerns the structure of text and control information in packets. This may be called the packet layer in contrast to the OSI
network layer. This packet layer protocol defines the format of the data
field of the frames exchanged at layer 2. Thus, the data text transmitted by
the user is first augmented at layer 3 with a packet header that specifies, for
example, the logical channel number or the type of packet. The packet is
then transferred to layer 2 LAPB., which incorporates it into the information field of a data-link frame. This frame is then transferred to the network
by means of the physical layer layer 1..
11.3.2.2 X.25 Packets Are Contained in LAPB Frames. Figure 11.6
shows the conceptual relationship of X.25 packets embedded in an LAPB
frame. It shows the X.25 packet occupying the LAPB frame information
field. Figure 11.7 shows the actual LAPB frame for modulo-8, modulo-128,
and modulo 32,768 operation. All frame types have 8-bit address fields
Figure 11.6. X.25 packet embedded in the LAPB frame information field.
371
Figure 11.7. LAPB frame formats. (a) Basic (modulo 8) operation, (b) extended (modulo 128)
operation, and (c) super (modulo 32,768) operation. (From Tables 2-1, 2-2, and 2-3r X.25,
page 8, ITU-T Rec. X.25, Ref. 2.)
372
providing an addressing capacity of 2 8 s 256 distinct addresses. It is the

control field that varies for the different operational types. The control field
for basic operation has 1 octet; for extended modulo 128. operation it has 2
octets, and for Super modulo 32768. operation, the control field has 4 octets.
LAPB is almost identical to HDLC as described in Chapter 3.
LAPB Frame Discussion. The frame formats shown in Figure 11.7 are for
single-link procedures. LAPB is basically a subset of HDLC as described in
Chapter 3. If startstop transmission is to be used, there must be eight
information bits between the start element and the stop element. The
information field must be octet-aligned. This means there must be an integer
number of octets. If not, the DCE will pad the information field so that it will
become octet-aligned.
11.3.2.3 Single-Link and Multilink Procedures. The multilink procedure
MLP. exists as an added upper sublayer of the data-link layer, operating
between the packet layer and a multiplicity of single data-link protocol
functions SLPs. in the data-link layer. The multilink concept in X.25 is
illustrated in Figure 11.8. An MLP performs the functions of accepting
packets from the packet layer, distributing those packets across the available
DCE or DTE SLPs for transmission to the DTE or DCE SLPs, respectively,
and resequencing the packets received from the DTE or DCE SLPs for
delivery to the DTE or DCE packet layer, respectively.
Field of Application. A multilink is a group of one or more single links in
parallel between a DCE and a DTE. The multilink procedure provides the
following general features:
1. It achieves economy and reliability of service by providing multiple
SLPs between a DCE and a DTE.
2. It permits addition and deletion of SLPs without interrupting the
service provided by the multiple SLPs.
Figure 11.8. Multilink functional organization. SLP, single-link procedure; MLP, multilink
procedure. (From Figure 2-2r X.25, page 36, ITU-T Rec. X.25, Ref. 2.)
373
Figure 11.9. Multilink frame formats. MLC, multilink control field. (From Table 2-15r X.25,
Figure 11.10. Multilink control field format. (From Table 2-16r X.25, page 37, ITU-T Rec.
X.25, Ref. 2.)
3. It optimizes bandwidth utilization of a group of SLPs through load

sharing.
4. It achieves graceful degradation of service when an SLPs. fails.
5. It provides each multiple SLP group with a single logical data-link layer
appearance to the packet layer.
6. It provides resequencing of the received packets prior to delivering
them to the packet layer.
Multilink Frame Formats. The multilink frame formats are shown in Figure
11.9, and the multilink control field format is given in Figure 11.10.
11.3.2.4 Packet-Level DTE r DCE Interface. Each packet to be transferred across the DTErDCE interface is contained within the data-link layer
information field that delimits its length, and only one packet is contained in
the information field.
374
Figure 11.11. The basic X.25 packet structure.
Basic Structure of X.25 Packets. ITU-T Rec. X.25 describes an interface

between the DTE and DCE for packet data communications. The DTE
resides in the data terminal equipment, whereas the DCE resides on the
network side. The DTE and DCE may be colocated physically or may be
separated by some distance. The DCE, for example, may be in a public
packet data switching center.
Every packet transferred across the DTEDCE interface may consist of
four or five fields: a protocol identifier octet, a general format identifier, a
logical channel group number, a logical channel number, and a packet-type
identifier. Other packet fields are appended as required. The basic packet
structure is illustrated in Figure 11.11. This figure is valid for modulo 8 and
128 operation.
It should be noted that for modulo 8 and 128 operation, the protocol
identifier octet is not present in any packet type. For modulo 32,768 operation, the protocol identifier octet is contained in the first octet of each
packet.
Table 11.1 defines the general format identifier. This is a 4-bit binary
coded field that is provided to indicate the general format of the rest of the
header. For modulo 8 and 128 operation, the general format identifier is
contained in the first octet of each packet. For modulo 32,768 operation, the
general format identifier is contained in the second octet of each packet. The
general format identifier field is located in bit positions 8, 7, 6, and 5 is
the low-order bit as shown in Table 11.1. The A bit, which we discuss later, is
in position 8. The A bit, when applied, deals with addressing. X in the Table
means a bit that may be set to either binary 1 or binary 0.
375
TABLE 11.1
General Format Identifier

Bit Position
8
Call setup packets (Note 1)
General Format Identifier

Sequence numbering scheme modulo 8
Sequence numbering scheme modulo 32,768
X
X
X
X
X
X
0
1
1
1
0
0
Clearing packets (Note 1)

X
X
X
0
0
0
0
1
1
1
0
1
Flow control, interrupt,

reset, restart, and diagnostic
packets

0
0
0
0
0
0
0
1
1
1
0
1
Data packets (Note 1)

X
X
X
X
X
X
0
1
1
1
0
1
Reserved format (Note 2)

a
Undefined.
Note 1. A bit that is indicated as X may be set to either 0 or 1, as discussed in the text.
Note 2. When the general format identifier field is contained in the first octet of a packet, this value
is reserved for other applications. When the first octet of a packet is the protocol identifier octet, then
this value is reserved for general format identifier extension.
Source: From Table 5-1r X.25, page 59, Ref. 2.
Logical ChannelrGroup Number. To enable simultaneous virtual calls

andror permanent virtual circuits, logical groupsrchannels are used. Each
virtual call or permanent virtual circuit is assigned a logical channel group
number less than or equal to 15. and a logical channel number less than or
equal to 255..
The logical channel group number appears in every packet except restart
and diagnostic packets. For modulo 8 and modulo 128 operation, the logical
channel group number is contained in the first octet of each packet. In the
case of modulo 32,786 operation, the logical channel group number is
contained in the second octet of each packet. The logical channel group
number is located in bit position 4, 3, 2, and 1. For each logical channel, this
number has local significance at the DTEDCE interface. This field is binary
coded, and bit 1 is the low-order bit of the logical channel group number. In
restart and diagnostic packets, this field is coded all zeros.
Logical Channel Number. The logical channel number appears in every
packet except restart and diagnostic packets. For modulo 8 and modulo 128
operation, the logical channel number is contained in the second octet of
each packet. For modulo 32,768 operation, the logical channel number is
contained in the third octet of each packet. The logical channel group
number is located in all bit positions of the octet. For each logical channel,
this number has local significance at the DTEDCE interface.
376
For virtual calls, a logical channel group number and a logical channel
number are assigned during the call setup phase. The range of logical
channels used for virtual calls is agreed upon with the administrations or
service companies at the time of subscription of the service. A similar
procedure is used for permanent virtual circuits.
Packet-Type Identifier. The packet-type identifier is carried in octet number
3 for modulo 8 and modulo 128 operation, and in octet number 4 for super
modulo 32,678 operation. Each packet is identified in accordance with
Table 11.2.
TABLE 11.2
Packet-Type Identifier
Packet Type
From DCE to DTE
Bit Position
From DTE to DCE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
0
0
0
1
0
1
1
1
1
1
1
1
1
1
X
0
0
X
0
0
X
1
1
X
0
0
X
0
0
X
0
1
X
1
1
0
1
1
X
0
0
X
0
0
X
0
0
0
0
X
0
0
X
0
0
X
0
0
0
0
X
0
0
X
0
0
X
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
1
1
1
1
1
0
0
0
1
1
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
1
1
1
1
Call Setup and Clearing

Incoming call
Call connected
Clear indication
DCE clear confirmation
Call request
Call accepted
Clear request
DTE clear confirmation
Data and Interrupt

DCE data
DCE interrupt
DCE interrupt confirmation
DTE data
DTE interrupt
DTE interrupt confirmation
Flow Control and Reset

DCE RR (modulo 8)
DCE RR (modulo 128) a
DCE RR (modulo 32,768) a
DCE RNR (modulo 8)
DCE RNR (modulo 128) a
DCE RNR (modulo 32,768) a
DCE REJ (modulo 32,768) a

Reset indication
DCE reset confirmation
DTE RR (modulo 8)
DTE RR (modulo 128) a
DTE RR (modulo 32,768) a
DTE RNR (modulo 8)
DTE RNR (modulo 128) a
DTE RNR (modulo 32,768) a
DTE REJ (modulo 8) a
DTE REJ (modulo 128) a
DTE REJ (modulo 32,768) a
Reset request
DTE reset confirmation
Restart
Restart indication
DCE restart confirmation
Restart request
DTE restart confirmation
Diagnostic
Diagnostic a
a
Not necessarily available on every network.

Note: A bit which is indicated as X may be set to either 0 or 1 as discussed in the text.
Source: Table 5-2r X.25, page 61, Ref. 2.
377
11.3.2.5 A Typical X.25 Packet. Figure 11.12 shows the format of a DTE
and DCE data packet. This is the packet that carries the revenue-bearing
data payload. All the other packet types would be listed as service packets.
Qualifier (Q) Bit. In some operational situations, an indicator bit may be
needed with the user data field to distinguish between two types of informationfor example, user data and control information. When such a mechanism is required, the qualifier Q. bit is used. When the Q bit is not needed,
it is set to binary 0.
Delivery Confirmation Bit. The setting of the delivery confirmation bit D bit.
is used to indicate whether or not the DTE wishes to receive end-to-end
acknowledgment of delivery, for data it is transmitting, by means of the
packet receive sequence number PR..
Sequence Numbering: P(R) and P(S). Each data packet transmitted at the
DTErDCE interface for each direction of transmission in a virtual call or
permanent virtual circuit is sequentially numbered. The sequence numbering
scheme of the packets is performed using modulo 8. The packet sequence
numbers cycle through the entire range of 0 to 7. There are optional
extended packet sequence numbering facility providing a numbering scheme
in modulo 128 and 2 15 in modulo 32,768. For example, for packet send
sequence number P S ., bits 4, 3, and 2 of octet 3, or bits 8 through 2 of octet
3 when extended, or bits 8 through 2 of octet 4 and bits 8 through 1 of octet 5
when superextended are used for indicating the packet send sequence
number P S .. P S . is binary coded and bit 2 is the low-order bit. When
superextended, bit 2 of octet 4 is the low-order bit, and bit 8 of octet 5 is the
high-order bit. Whatever the modulo scheme, the packet sequence numbering scheme is the same in both directions of transmission and is common for
all logical channels at the DTEDCE interface.
Only data packets have the send sequence number P S .. The send
sequence number P S . maintains the integrity of the data because it is a
basic mechanism for the detection of lost packets. The receive sequence
number P R . is used to acknowledge correct receipt of packets and authorize transmission of additional packets.
Flow Control. There are three flow control packets: receive ready RR.,
receive not ready RNR., and reject REJ.. Each of these packets is 3 octets
long in the modulo-8 case; 4 octets long when modulo-128 extended sequence
numbering is used; and six octets long when super modulo 32,768 is used.
The received sequence number P R . acknowledges receipt of packets numbered P R . y 1.
There is also a flow control window defined for each direction of data
transmission of a logical channel used for a virtual call or a permanent virtual
378
Figure 11.12. DTE and DCE data packet formats showing modulo 8 sequence numbering
(upper panel), modulo 128 (middle panel), and modulo 32,678 (lower panel). (From Figure
5-8r X.25, page 75, ITU-T Rec. X.25, Ref. 2.)
379
circuit. The window has a size, W, ranging from 1 to 7. The window size
specifies the number of sequentially numbered packets that a DTE may have
outstanding i.e., unacknowledged .. The standard window size W is 2 for
each direction of transmission. Other window sizes can be negotiated through
the flow control parameter negotiation facility.
When the sequence number P S . of the next data packet to be transmitted is within the window, the DCE is authorized to transmit this data packet
to the DTE. When the sequence number P S . of the next data packet to be
transmitted is outside of the window, the DCE will not transmit the data
packet to the DTE. The DTE follows the same procedure.
When the sequence number P S . of the data packet received by the DCE
is the next in sequence and is within the window, the DCE will accept this
data packet. A received data packet containing a P S . that is out of
sequence i.e., there is a duplicate or gap in the P S . numbering., outside the
window, or not equal to 0 for the first data packet after entering the flow
control ready state d1. is considered by the DCE as a local procedure error.
The DCE will reset the virtual call or permanent virtual circuit. The DTE
follows the same procedure.
The packet receive sequence number P R . conveys across the DTErDCE
interface information from the receiver for the transmission of data packets.
When transmitted across the DTErDCE interface, a P R . becomes the
lower window edge. In this way, additional data packets may be authorized
by the receiver to cross the DTEDCE interface.
The packet receive sequence number P R . is conveyed in data, receive
ready RR., and receive not ready RNR. packets.
The value of P R . received by the DCE must be within the range from
the last P R . received by the DCE up to and including the packet send
sequence number of the next data packet to be transmitted by the DCE.
Otherwise, the DCE will consider the receipt of this P R . as a procedure
error and will reset the virtual call or permanent virtual circuit. The DTE
follows the same procedure.
The receive sequence number P R . is less than or equal to the sequence
number of the next expected data packet and implies that the DTE or DCE
transmitting P R . has accepted at least all data packets numbered up to and
including P R . y 1.
User Data Field Length. The standard maximum user data field length is 128
octets. In addition, other maximum user data field lengths may be offered by
administrations from the following list: 16, 32, 64, 256, 512, 1026, 2048, and
4096 octets. The user data field of data packets transmitted by a DTE or
DCE may contain any number of bits up to the agreed maximum. Rather
than bits, some networks will use octets.
More Data Mark, the M Bit. If a DTE or DCE wishes to indicate a sequence
of more than one packet, it uses a more data mark M bit.. The M bit can be
380
TABLE 11.3 Definition of Two Categories of Data Packets and Network Treatment
of the M and D Bits
Data Packet Sent by Source DTE
Category
Full
Combining with Subsequent

Packet(s) Is Performed by the
Network When Possible
Data packet a Received by

Destination DTE
M
0 or 1
No
No
0 (Note 1)
No
No
No
No
Yes
No
Yes
No
Yes
Yes (Note 2)
Yes
No
0
1
Refers to the delivered data packet whose last bit of user data corresponds to the last bit of user data, if
any, that was present in the data packet sent by the source DTE.
Note 1. The originating network will force the M bit to 0.
Note 2 If the data packet sent by the source DTE is combined with other packets, up to and including a
category B packet, the M and D bit settings in the data packet received by the destination DTE will be
according to that given in the two right-hand columns for the last data packet sent by the source DTE that
was part of the combination.
Source: Table 4-1r X.25, page 51, Ref. 2.
set to 1 in any data packet. When it is set to 1 in a full data packet or in a

partially full data packet also carrying the D bit set to 1, it indicates that
more data are to follow. Recombination with the following data packet may
only be performed within the network when the M bit is set to 1 in a full data
packet which also has the D bit set to 0.
A sequence of data packets with every M bit set to 1 except for the last
one will be delivered as a sequence of data packets with the M bit set to 1
except for the last one when the original packets having the M bit set to 1 are
either full irrespective of the setting of the D bit. or partially full but have
the D bit set to 1.
Two categories of data packets, A and B, have been defined and are
shown in Table 11.3. This table also illustrates the networks treatment of the
M and D bits at both ends of a virtual call or permanent virtual circuit.
Complete Packet Sequence. A complete packet sequence is defined as
being composed of a single category B packet and all contiguous preceding
category A packets if any.. Category A packets have the exact maximum user
data field length with the M bit set to 1 and the D bit set to 0. All other data
packets are category B packets. When transmitted by a source DTE, a
complete packet sequence is always delivered to the destination by a single
complete packet sequence.
X.25 Octet Conventions. Bits of an octet are numbered 8 to 1, where 1 is the
low-order bit and is transmitted first. Octets of a packet are consecutively
numbered starting from 1 and are transmitted in this order.
381
11.3.2.6 Typical X.25 Transmissions Exchange of Packets. Figures

11.13a and 11.13b are X.25 call setup and call clearing state diagrams.
Ready State. If there is no call in existence, a logical channel is in the
ready state.
Call RequestrIncoming Call Packet. The call request is initiated by the
DTE to indicate a request to establish a virtual circuit. The incoming call
packet is identical but is initiated by the DCE. The packet carries the logical
channel number, as well as the calling and called DTE addresses, and
provides a means for carrying parameters used to define specific information
such as billing information about the call. The packet also contains a user
Figure 11.13. X.25 call setup (upper) and call clearing (lower) state diagrams. (From Figure
B.2r X.25, page 123, ITU-T Rec. X.25, Ref. 2.)
382
information field up to 16 octets long. In conjunction with the fast select

facility, it can have 128 octets.
Call-AcceptedrCall-Connected Packet. The call-accepted and call-connected packets are used to indicate that the called DTE has accepted a
request to establish a call. Call accepted is initiated by the called DTE, and
the network delivers a call-connected packet to the calling DTE. Data
transfers can begin after the call is accepted. These packets include the
general format identifier, the logical channel number, and the packet identifier. Address and facilities fields are optional.
Call Clearing. Three packets are used in the call-clearing phase. The clear
request packet is issued by the DTE initiating the clearing call take-down..
It is delivered to the remote DTE as a clear indication packet. The clear
confirmation packet is issued by both the DTE and DCE to acknowledge
receipt of clear packets.
These packets contain the standard 3-octet header. In addition, the
request and indication packets contain a clearing-cause field and an optional
diagnostic code. On some networks, optional address and facilities fields are
permitted.
The clearing-cause field indicates the reason for the clear, such as busy,
network congestion, or invalid call.
11.3.2.7 Addressing in X.25. Figure 11.14 shows the relationship of X.25
address fields in typical X.25 frames. In this case, the frame group selected
deals with call setup and call clearing, specifically the call request and
incoming call packet format.
The call setup and clearing packets contain an address block which can
take on one of two possible formats. The first format, known as the
non TOArNPI address format, can accommodate addresses conforming to
the formats described in ITU-T Recs. X.121 and X.301 whose length including possible prefixes andror escape codes. is not greater than 15 digits. The
second format, known as the TOArNPI address format, can be used by
networks and DTEs to accommodate addresses conforming to the formats
described in ITU-T Recs. X.121 and X.301 whose length is greater than 15
digits and can also be used to carry an alternative address in the called DTE
address field of the call request and clear request packet. The address block of
the TOArNPI format contains in addition to the address itself. fields to
specify the type of address TOA. and numbering plan identification NPI..
The non-TOArNPI address format and the TOArNPI address format are
distinguished by bit 8 A bit. of the general format identifier Table 11.1..
See Figures 11.11, 11.12, and 11.14. When the A bit is set to 0, the
non-TOArNPI address format is used. See Figure 11.15.. When the A bit is
set to 1, the TOArNPI address format is used. See Figure 11.16.. Note that
the A-bit address bit. is valid only for call setup and call clearing packets
and is set to 0 for all other packets.
383
Figure 11.14. Call request and incoming call packet format. The upper diagram deals with
modulo 8 and modulo 128 operation; the lower diagram deals with modulo 32,768 operation.
(From Figure 5-4r X.25, page 67, ITU-T Rec. X.25, Ref. 2.)
The non-TOArNPI address format is supported by all networks. The

TOArNPI address format may be supported by some networks and by
some DTEs.
The use of the TOArNPI address format is only possible if the TOArNPI
address format is supported by the network. When the address format used
by one DTE in a call setup or call clearing packet is not the same as the
address format used by the remote DTE, the network if it supports the
TOArNPI address format. converts from one address format to the other.
384
Figure 11.15. Format of the address block when the A bit is set to 0, non-TOAr NPI
addressing. Note: The figure is drawn assuming the number of semi-octets present in the
called DTE address field is odd and the number of semi-octets in the calling DTE address field
is even. (From Figure 5-1r X.25, page 62, ITU-T Rec. X.25, Ref. 2.)
Figure 11.16. Format of the address block when the A bit is set to 1, TOAr NPI addressing.
Note: The figure is drawn assuming the number of semi-octets present in the called DTE
address field is odd and the number of semi-octets in the calling DTE address field is even.
(From Figure 5-2r X.25, page 63, ITU-T Rec. X.25, Ref. 2.)
In either case, we see that there are four basic elements to the addressing
broken down into two groups: address length fields and address fields. There
is one each for the calling DTE and the called DTE. Address length fields
are each 4 bits long, occupying the first octet in the address block see
Figures 11.15 and 11.16..
In these address formats, the ITU-T Organization has had to turn to using
the term semi-octet, an octet of less than 8 bits, in this case 4 bits. The length
indicators are binary encoded, and bit 1 is the low-order bit of the indicator.
The address fields themselves have three subfields: type of address TOA.
subfield, numbering plan identification NPI. subfield, and the address digits
subfield. The first two subfields are at the beginning of the address and are
binary-encoded. These provide information such as whether the number is
international, network-dependent, national, or alternative address and
whether it is based on ITU-T Rec. E.164 or X.121.
When the type of address subfield indicates an address other than an
alternative address, the other semi-octets of a DTE address are digits, coded
385
Figure 11.17. Logical channel assignment diagram. LCN, logical channel number; LIC,
lowest incoming channel; HIC, highest incoming channel; LTC, lowest two-way channel; HTC,
highest two-way channel; LOC, lowest outgoing channel; HOC, highest outgoing channel.
(From Figure A.1r X.25, page 120, ITU-T Rec. X.25, Ref. 2.)
in binary-coded decimal with bit 5 or 1 being the low-order bit of the digit.
Starting from the high-order digit, the address digits are coded in semioctets. In each octet, the higher-order digit is coded in bits 8, 7, 6, and 5.
When present, the calling DTE address field starts on the first semi-octet
following the end of the called DTE address field. Consequently, when the
number of semi-octets of the called DTE address field is odd, the beginning
of the calling DTE address field, when present, is not octet-aligned.
When the total number of semi-octets in the called and calling DTE
address field is odd, a semi-octet with zeros in bits 4, 3, 2, and 1 is inserted
after the calling DTE address field in order to maintain octet alignment.
11.3.2.8 Assignment of Logical Channels. Logical channel assignment is
based on the coding of the second semi-octet of the first field of the X.25
packet format see Figure 11.12. and the second octet, for a total of 12 binary
digits or 4096 binary digit combinations. The assignment is carried out in a
manner similar to that of telephone trunks in the way it is handled at each
end of a connection. This is shown in Figure 11.17. One end of the
connection selects channels from the low end of the channel assignment,
starting with channel 0, then 1, 2, 3 and so forth. The other end of the circuit
selects starting from the highest-numbered channelin this case channel
number 4095, then 4094, and then 4093 and so on.
386
r INTERNET
11.4 TRANSMISSION CONTROL PROTOCOLr
r IP)
PROTOCOL (TCPr
11.4.1
Background and Application
TCPrIP protocol family was developed by the U.S. Department of Defense

for the ARPANET Advanced Research Projects Agency Network..
ARPANET was one of the first large advanced packet networks. It dates
back to 1968 and was well into existence before CCITT and ISO took interest
in layered protocols.
The TCPrIP suite of protocols Ref. 5. has wide acceptance today. These
protocols are used on both local area networks LANs. and wide area
networks WANs.. They are particularly attractive for their internetworking
capabilities. The internet protocol IP. competes with CCITT Rec. X.75
protocol Ref. 6..
The architectural model of the IP Ref. 7. uses terminology that differs
from the OSI reference model. The IP predates OSI.. Figure 11.18 shows
the relationship between TCPrIP and related Department of Defense DoD.
protocols and the OSI model. It traces data traffic from an originating host,
which runs an applications program, to another host in another network. This
may be a LAN-to-WAN-to-LAN, a LAN-to-LAN, or a WAN-to-WAN connectivity. The host enters its own network by means of a network access
protocol such as LAPB a derivative of HDLC, described in Section 11.3. or
an IEEE 802 series protocol. A common LAN-to-WAN-to-LAN connectivity
is shown in Figure 11.19.
Figure 11.18. How TCPr IP and associated DoD protocols relate to the OSI reference model.
TRANSMISSION CONTROL PROTOCOLr INTERNET PROTOCOL (TCPr IP )
387
Figure 11.19. Connecting one LAN to another via a WAN, employing IP.
In the figure, the IP router will encapsulate* typically an 802.3 MAC

frame derived from the LAN. The encapsulated information is called a
datagram in IP language. IP operates in OSI layer 3 see Figure 11.18.. The
datagram is now passed to the data-link layer layer 2. where it is encapsulated still again. The resulting body of data is now called a packet, which is
transmitted by the appropriate physical layer.
IP is transparent to the information contained in the datagram; it just adds
some control information to the data received from an upper-layer protocol
often TCP or UDP. and then attempts to deliver it to a station on
the network.
The term host is commonly used in the IP world meaning the lowest level
of addressing and refers to a device. That device is often a computer or some
other digital processor.
IP and related documentation are supported by RFCs Request for
Comment.. We must remember that IP was a U.S. Department of Defense
DoD. initiative. RFC is typical language of the DoD. Before releasing a
specification or request for proposal, the DoD would often issue an RFC
document to industry. This allowed industry to shape DoD programs such
that they would be more reasonable to bid, perhaps less costly. The RFC
label stuck for the IP family of protocols Figure 11.18.. RFCs are issued
sequentially, where RFC 791 defines the present IP4. As of this writing, the
issuing agency is beyond RFC 4000.
11.4.2
TCPr
r IP and Data-Link Layers
TCPrIP is transparent to the type of data-link layer involved, and it is also

transparent whether it is operating in a LAN or WAN regime or among
them. There is wide document support available for interfacing a large
*Encapsulateplacing header information around a block or frame of data.
388
Figure 11.20. The incorporation of upper-layer PDUs into a data-link layer frame showing the
relationship of TCP and IP.
number of popular data-link layer protocols. Among these we find the IEEE
802 series, Ethernet, ARCNET for LANs, and X.25 for WANs.
Typical IP encapsulation is shown in Figure 11.20. The figure illustrates
how upper OSI layers are encapsulated with TCP and IP header information
and then incorporated into the data-link layer frame.
For the case of IEEE 802 series LAN protocols, advantage is taken of the
LLC common to all 802 protocols. The LLC extended header contains
the subnetwork access protocol SNAP. such that we have 3 octets for the
LLC header and 5 octets in the SNAP. The LLC header has its fields fixed
as follows:
DSAP s 10101010
destination service access point .
SSAP s 10101010
source service access point .
Control s 00000011
for unnumbered information UI frame .
The 5 octets in the SNAP have three assigned for protocol ID or organizational code and two for EtherType. Typical EtherType assignments are
shown in Table 11.4. EtherType refers to the general class of LANs based on
CSMArCD.
Figure 11.21 shows how OSI relates to TCPrIP and IEEE 802, and Figure
11.22 illustrates an IEEE 802 frame incorporating TCP, IP, and LLC.
Addressing is resolved by a related TCPrIP protocol called address
resolution protocol ARP., which, typically, performs the mapping of the
32-bit IP4 internet address into a 48-bit IEEE 802 address.
Another interface problem may be the recommended minimum and absolute maximum IP datagram lengths. The total length is the length of the
datagram, measured in octets, including internet header and data. The total
length field in the IP header allows the length of a datagram to be up to
TABLE 11.4
389
EtherType Assignments
Ethernet Decimal
Hex
Description
512
513
1,536
2,048
2,049
2,050
2,051
2,052
2,053
2,054
2,055
4,096
21,000
24,577
24,578
24,579
24,580
24,582
24,583
32,773
32,784
32,821
32,824
32,823
0200
0201
0600
0800
0801
0802
0803
0804
0805
0806
0807
1000
5208
6001
6002
6003
6004
6005
6006
8005
8010
8035
8038
8098
XEROX PUP
PUP Address Translation
XEROX NS IDP
DOD Internet Protocol (IP)
X.75 Internet
NBS Internet
ECMA Internet
Chaosnet
X.25 level 3
Address Resolution Protocol (ARP)
XNS Compatibility
Berkeley Trailer
BBN Simnet
DEC MOP Dump r Load
DEC MOP Remote Console
DEC DECnet Phase IV
DEC LAT
DEC
DEC
HP Probe
Excelan
Reverse ARP
DEC LANBridge
Appletalk
Figure 11.21. How TCPr IP working with IEEE 802 relates to OSI. ULP, upper-layer protocol.
390
Figure 11.22. An IEEE 802 frame showing LLC and TCPr IP functions.
65,535 octets. Such long datagrams are impractical for most hosts and
networks. All hosts should be prepared to accept datagrams up to 576 octets
whether they arrive whole or in fragments.. It is recommended that hosts
only send datagrams larger than 576 octets if they have assurance that the
destination is prepared to accept larger datagrams.
The number 576 was selected to allow a reasonable-sized data block to be
transmitted in addition to the required header information. For example, this
size allows a data block of 512 octets plus 64 header octets to fit in a
datagram. The maximum internet header is 60 octets, and a typical internet
header is 20 octets, allowing a margin for headers of higher level protocols.
11.4.3
The IP Routing Function
IP offers typical connectionless service. In OSI, the network layer functions

include routing and switching of a datagram through the telecommunications
subnetwork. IP provides this essential function. It forwards the datagram
based upon the network address contained within the IP header. Each
datagram is independent and has no relationship with other datagrams.
There is no guaranteed delivery of the datagram from the standpoint of IP.
However, the next higher layer, the TCP layer, provides the reliability that IP
lacks. It also carries out segmentation and reassembly functions of a datagram to match the frame sizes of the data-link layer protocols.
Addresses determine routing, and, at the far end, equipment host hardware.. Actual routing derives from the IP address, and equipment addresses
derive from the data-link header, typically the 48-bit CSMArCD MAC
address. Some details of IP routing are covered in Section 10.5.1.
391
User data from upper-layer protocols are passed to the IP layer. The IP
layer examines the network address IP address . for a particular datagram
and determines if the destination node is on its own local area network or
some other network. If it is on the same network, the datagram is forwarded
directly to the destination host. If it is on some other network, it is forwarded
to the local IP router gateway.. The router, in turn, examines the IP address
and forwards the datagram as appropriate. Routing is based on a lookup
table residing in each router or gateway. See Section 11.4.4.2 for a more
detailed discussion of IP routing..
11.4.4
Detailed IP Operation
The IP provides connectionless service, meaning that there is no call setup

phase prior to exchange of traffic. There are no flow control nor error
control capabilities incorporated in IP. These are left to the next higher
layer, the TCP. The IP is transparent to subnetworks connecting at lower
layers, and thus different types of networks can attach to an IP gateway.
Whereas prior to this discussion we have used the term segmentation to
mean the breaking up of a data file into manageable segments, frames,
packets, blocks, and so on, the IP specifications refer to this as fragmentation.
IP messages are called datagrams. As we mentioned in Section 11.4.2, the
minimum datagram length is 576 octets, and the maximum length is 65,535
octets. Fragmentation resolves protocol data unit PDU. sizes of the different
networks with which IP carries out an interface function. For example. X.25
packets typically have data fields of 128 octets, Ethernet limits the size of a
PDU to 1500 octets, and so forth. Of course, IP does a reassembly of the
fragments at the opposite end of its circuit.
11.4.4.1 Frame Format of an IP Datagram; Description of Functions.
The IP datagram frame format is illustrated in Figure 11.23. The format
should be taken in context with Figures 11.21 and 11.22, showing how the IP
datagram relates to TCP and the data-link layer. In Figure 11.23, we move
from left to right and top down in our description.
Ver. Version field 4 bits. gives the release number of the IP version for
which a particular router or gateway. is equipped. IP has versions 4, 5, and
6. Several sources tell us that version 5 does not exist and that version 6 IP6.
is experimental Ref. 9.. Thus the version will be IP4, until IP6 gets the
wrinkles worked out. See RFC 791 for IP4 specification. . It is assigned
decimal 4, and IP6 RFC 1883. is assigned decimal 6.
IHL. Internet header length 4 bits. is measured in units of 32 bits.*
Example: A header without certain options, such as QoS, typically has 20
*It is interesting to note that IP is based on 32-bit boundaries.
392
Figure 11.23. The IP datagram format.
octets. Thus its length is 5 or simply 20r4 32 bits or 4 octets per unit;
5 = 4 s 20..
Type of Service. TOS is rarely used and usually set to 0 8 bits.. The TOS,
when employed, is used to identify several QoS parameters provided by IP.
The 8-bit field is broken down into four active groupings, and the last
two bits are reserved. The first is precedence, which consists of three bits
as follows:
000
010
100
110
Routine
Immediate
Flash override
Internetwork control
001
011
101
111
Priority
Flash
CEITICrECP
Network control
The three bits for precedence are followed by three 1-bit fields:
D delay:
T throughput:
R Reliability:
0 s normal, 1 s low
0 s normal, 1 s high
0 s normal, 1 s high
For example, if there is more than one route to a destination, the router
could read this field to pick a route. This becomes important in open shortest
path first OSPF. routing protocol, which was the first IP protocol to take
advantage of this.
393
Total Length. The total length field specifies the total length of an IP
datagram not packet. in octets. It includes the length of the header. There
are 16 bits in the field. Therefore the maximum length of a datagram is
2 16 y 1 or 65,535 octets; the default length, as we know, is 576 octets. IP6
allows a concept known as the jumbogram to manage very long files.
The IP folks call segmentation fragmentation. It allows for the breaking up
of long files into fragments. The next three fields in the IP header deal
with fragmentation. These are:
Identifier: 16 bits
Flags: 3 bits
Fragment offset: 3 bits
The term fragment suggests part of a whole; the identifier field identifies a
fragment as part of a complete datagram, along with the source. It indicates
which datagram fragments belong together so datagrams do not get mismatched. The receiving IP layer uses this field and the source IP address to
identify which fragments belong together.
Flags 3 bits.. Flags indicate whether more fragments are to arrive or no
more data is to be sent for that particular datagram. They are coded in
sequence as follows:
Bit 0 s reserved
Bit 1 s 0 s fragment; 1 s do not fragment
Bit 2 s M bit . s 0 s last fragment; 1 s more fragments to follow
It should be noted that as a fragment datagram passes through the network it
can be further fragmented. If a router receives a packet that must be
fragmented and the dont fragment bit is set to 1, then it will discard the
packet and send an error message via the ICMP protocol.
Fragment Offset (or just offset). The fragmentation offset field contains a
value that specifies the relative position of the fragment in the original
datagram. The value is initialized at 0 and subsequently set to the proper
number whenrif the router fragments the data. The offset value is measured
in units of 8 octets 64 bits..
Using the total length and the fragment offset fields, IP can reconstruct a
fragmented datagram and deliver it to the upper-layer software. The total
length field indicates the total length of the original packet, and the offset
field indicates to the node that is reassembling the packet the offset from the
beginning of the packet. It is at this point that the data will be placed in the
data segment to reconstruct the packet.
394
Time to Live TTL. 8 bits.. This indicates the amount of time that a
datagram is allowed to stay in the network. Each router in the network
through which the datagram passes is required to check this field and discard
it if the TTL values equals 0. A router is also required to decrement this field
in each datagram it processes. In actual implementations, the TTL field is a
number of hops value. Therefore, when a datagram proceeds through a
router, it has completed one hop and the field value is decremented by one.
Implementation of IP may use a time counter in this field and decrement the
value in 1-s decrements.
The TTL field can be used by routers to prevent endless loops. It can also
be used by hosts to limit the lifetime that segments have in the network.
Network management facilities employing, say, SNMP* might wish to set the
TTL values for diagnostic purposes.
Protocol 8 bits.. The protocol field identifies the next-level protocol above
the IP that is to receive the datagram at the final host destination. It is
similar to the type field found in EthernetrCSMArCD. Table 11.5, taken
from RFC 1060 Ref. 10. shows the protocol decimal numbering scheme and
the corresponding protocol for each number. It identifies the most widely
used upper-layer protocols. The most common protocols we would expect
above the IP layer are TCP and UDP.
Header Checksum 16 bits.. This detects errors in the header by means of
cyclic redundancy check. CRC is described in Sections 4.4 and 4.5. If an error
does occur in the header, the router notifies the originating host via an
ICMP message.
Source Address, Destination Address. Each of these fields is 32 bits long.
There are two types of network addressing schemes used with IP Ref. 9. see
RFC 1918.:
Classless. The full address range all 32 bits. can be used without regard to
bit reservation classes. This type of addressing scheme is basically not
used in direct host assignment. The scheme is directly applied to the
routing tables used on the Internet and ISPs.
Classful. Where the 32-bit address is segmented into specific classes
denoting networks and hosts.
The IP address structure is shown in Figure 11.24. It shows four address
formats in which the lengths of the component fields making up the address
field change with each of the different formats. These component fields are
*SNMP stands for simple network management protocol, a member of the TCPrIP family of
protocols. See Chapter 18.
ICMP stands for internet control message protocol. See Section 11.4.4.3.
TABLE 11.5
Decimal
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
3460
61
62
63
64
65
66
67
68
69
70
71
7275
76
77
78
79
80254
255
IP Protocol Field Numbering (Assigned Internet Protocol Numbers)

Key Word
ICMP
IGMP
GGP
ST
TCP
UCL
EGP
IGP
BBN-MON
NVP-II
PUP
ARGUS
EMCON
XNET
CHAOS
UDP
MUX
DCN-MEAS
HMP
PRM
XNS-IDP
TRUNK-1
TRUNK-2
LEAF-1
LEAF-2
RDP
IRTP
ISO-TP4
NETBLT
MFE-NSP
MERIT-INP
SEP
CFTP
SAT-EXPAK
MIT-SUBN
RVD
IPPC
SAT-MON
IPCV
BRSAT-MON
WB-MON
WB-EXPAK
Protocol
Reserved
Internet Control Message Protocol
Internet Group Management Protocol
Gateway-to-Gateway Protocol
Unassigned
Stream
Transmission Control Protocol
UCL
Exterior Gateway Protocol
Interior Gateway Protocol
BBN-RCC Monitoring
Network Voice Protocol
PUP
ARGUS
EMCON
Cross Net Debugger
Chaos
User Datagram Protocol
Multiplexing
DCN Measurement Subsystems
Host Monitoring Protocol
Packet Radio Monitoring
XEROX NS IDP
Trunk-1
Trunk-2
Leaf-1
Leaf-2
Reliable Data Protocol
Internet Reliable TP
ISO Transport Class 4
Bulk Data Transfer
MFE Network Services
MERIT Internodal Protocol
Sequential Exchange
Unassigned
Any host internal protocol
CFTP
Any local network
SATNET and Backroom EXPAK
MIT Subnet Support
MIT Remote Virtual Disk
Internet Plur. Packet Core
Any distributed file system
SATNET Monitoring
Unassigned
Packet Core Utility
Unassigned
Backroom SATNET Monitoring
Unassigned
Wideband Monitoring
Wideband EXPAK
Unassigned
Reserved
Source: RFC 1060 (Ref. 10 ) and Internet Protocol Transition Workbook (Ref. 11 ).
395
396
Figure 11.24. Internet protocol address formats. (RFC 791). Note that an IP address has
three sections from left to right: Class (1, 2, 3, or 4 bits), Network ID, and Host ID. This holds
for both destination and source addresses. ( Note: This is basis of classful addressing.)
Class A, B, C, or D; occupying the first 1, 2, 3, or 4 bits., network address,

and local address. In Section 10.3.5.1, we break down the local address
into two subfields: subnetwork address and host address.
The Class A addressing is for very large networks, such as ARPANET.
The field starts with binary 0, indicating that it is a Class A format. In this
case the local or host address component field is 24 bits long and has an
address capacity of 2 24 . Class B addressing is for medium-size networks, such
as campus networks. The field begins with 10 to indicate it is a Class B
format; the network component field is 14 bits long, and the host or local
address component field is 16 bits in length. The Class C format is for small
networks with a very large network ID field with an addressing capacity of
2 24 and a considerably small host ID field of only 8 bits 2 8 addressing
capacity .. The address field in this case starts with binary 110. The Class D
format is for multicasting, a form of broadcasting. Its first 4 bits are the
binary sequence 1110.
397
11.4.4.2 IP Routing. An IP router needs only the network ID portion of

the address to perform its routing function. Each router or gateway has a
routing table which consists of destination network addresses and a specified
next-hop router gateway..
Three types of routing are performed with the support of a routing table:
1. Direct routing to locally attached devices hosts..
2. Routing to networks that are reached by one or more routers gateways..
3. Default routing to destination network in case the first types of routing
are unsuccessful.
A typical routing table is shown in Figure 11.25. Each row of the table
contains data for each route that is known to the IP module storing the table.
Destination. Destination is self-explanatory and contains destination address. When the column is coded 0.0.0.0, the route is a default route.
IfIndex Interface Index.. Identifies the local physical port through which
the next hop is routed.
Metric. Columns 3, 4, and 5 are labeled Metric. Actually five columns are
assigned a Metric entry. Each column contains information about the
cost metric used in determining the route. With most systems, the cost
metric is the number of hops necessary to reach the destination.
Ordinarily one metric would be used for route determination. But with
Figure 11.25. IP routing table.
398
new sophisticated algorithms wtypically OSPF open shortest path first.x,

more than one metric could be used in route calculation.
Next Hop Ref. 6.. This contains the IP address which identifies the next
hop for the route selected.
Route Type Ref. 7.. There are four possibilities here: 1, none of the
following; 2, an invalid route; 3, directly connected route meaning a
directly connected subnet; 4, indirect meaning that an indirect connection is used to reach the destination.
Routing Protocol Ref. 5.. This tells the route discovery protocol by which
the route was learned. This entry may be, for example, the RIP protocol
where the column value would be 8, the Cisco internal group routing
protocol where the value would be 11, the exterior gateway protocol
EGP. where the value would be 5, and so on.
Route Age Ref. 10.. Value in seconds since the route was verified or
updated.
Routing Mask Ref. 9.. The mask is an overlay showing the valid binary
digits. This is logical ANDed with the destination address in the IP
datagram before being compared to column 1 of the table.
Route Information Ref. 11.. This is a reference to MIB management
information base. definition dealing with a particular protocol. Value
depends on the type of protocol used.
Suppose a datagram or datagrams. is are. directed to a host that is not in
the routing table resident in a particular gateway. Likewise, there is a
possibility that the network address for that host is also unknown. These
problems may be resolved with the address resolution protocol ARP.
Ref. 11..
First the ARP searches a mapping table that relates IP addresses with
corresponding physical addresses. If the address is found, it returns the
correct address to the requester. If it cannot be found, the ARP broadcasts a
request containing the IP target address in question. If a device recognizes
the address, it will reply to the request where it will update its ARP cache
with that information. The ARP cache contains the mapping tables maintained by the ARP module.
There is also a reerse address resolution protocol RARP, Ref. 12.. It works
in a fashion similar to that of ARP, but in reverse order. RARP provides an
IP address to a device when the device returns its own hardware address.
This is particularly useful when certain devices are booted and only know
their own hardware address.
Routing with IP involves a term called hop. A hop is defined as a link
connecting adjacent nodes routers or gateways. in a connectivity involving
IP. A hop count indicates how many nodes routers or gateways. must be
traversed from source to destination.
399
There are two routing protocols used with IP. These are RIP routing
information protocol. and OSPF open shortest path first..
RIP is known as a distance ector protocol. As one might imagine, it uses a
database that we call a routing table. It contains two fields needed for
routing: a vector a known IP address . and distance, meaning how many
routers away is the destination. There are other fields in the table, but these
are the basic ones which provide the essential concept of the protocol. In
other words, RIP simply builds a table or database . in memory that contains
all the routes that it has learned and the distance to that network. A router
associates a cost with those interfaces. The cost is initially set to 1. As other
routers report their topology and connectivity regarding locally attached
subnets, all the information needed will be stored in each routers database
routing table.. Routing will then be based on cost, which really equates to
hop count.
OSPF Ref. 13. is a link state protocol. It maintains the state of every link
in the domain. Information is flooded to all routers in the domain. Flooding
is discussed in Chapter 10. Flooding is the process of receiving information
on one port and retransmitting to all other active ports of a router. This is
how routers receive the same information. The information is stored in a
router database that is called a link-state database. This database is identical
on every router in an autonomous system. Based on information from the
database, the Dykstra algorithm Ref. 9. is employed which seeks the
shortest-path tree based on metrics. It uses itself as the root of the tree. It is
able to build a routing table based on shortest-path routing.
11.4.4.3 Internet Control Message Protocol (ICMP). ICMP Ref. 14. is
used as an adjunct to IP when there is an error in datagram processing.
ICMP uses the basic support of IP as if it were a high-level protocol.
However, ICMP is actually an integral part of IP and is implemented in every
IP module.
ICMP messages are sent in several situationsfor example, when a
datagram cannot reach its destination, when a gateway does not have the
buffering capacity to forward a datagram, and when the gateway can direct
the host to send traffic on a shorter route.
ICMP messages typically report errors in the processing of datagrams. To
avoid the possibility of infinite regress of messages about messages, and so
on, no ICMP messages are sent about ICMP messages. Also ICMP messages
are only sent about errors in handling fragment zero of fragmented datagrams. Note: Fragment zero has the fragment offset equal to zero..
Message Formats. ICMP messages are sent using the basic IP header see
Figure 11.23.. The first octet of the data portion of the datagram is an
ICMP-type field. The data portion is the last field at the bottom of Figure
11.23. The ICMP-type field determines the format of the remaining data.
Any field labeled unused is reserved for later extensions and is fixed at zero
400
when sent, but receivers should not use these fields except to include them
in the checksum.. Unless otherwise noted under individual format descriptions, the values of the internet header fields are as follows:
Version: 4.
IHL internet header length.: length in 32-bit words.
Type of service: 0.
Total length: length of internet header and data in octets.
Identification, flags, and fragment offset: used in fragmentation, as in
basic IP described previously.
Time to live: in seconds; as this field is decremented at each machine in
which the datagram is processed, the value in this field should be at
least as great as the number of gateways which this datagram will
traverse.
Protocol: ICMP s 1.
Header checksum: The 16-bit ones complement of the ones complement sum of all 16-bit words in the header. For computing the checksum, the checksum field should be zero. The reference RFC Ref. 14.
RFC 792. states that this checksum may be replaced in the future.
Source address: The address of the gateway or host that composes the
ICMP message. Unless otherwise noted, this is any of a gateways
addresses.
Destination address: The address of the gateway or host to which the
message should be sent.
There are eight distinct ICMP messages covered in RFC 792:

1.
2.
3.
4.
5.
6.
7.
8.
Destination unreachable message

Time exceeded message
Parameter problem message
Source quench message
Redirect message
Echo or echo reply message
Timestamp or timestamp reply message
Information request or information reply message
Example: Destination Unreachable Message. The ICMP fields in this case

are shown in Figure 11.26.
IP Fields
Destination Address: The source network and address from the original
datagrams data.
401
Figure 11.26. A typical ICMP message format: destination unreachable message. Note: The
relevant RFC for ICMP is RFC 792 (Ref. 14).
ICMP Fields
Type 3
Code: 0 s net unreachable
1 s host unreachable
2 s protocol unreachable
3 s port unreachable
4 s fragmentation needed and DF set
5 s source route failed
Checksum: as above
Internet header q 64 bits of data datagram
The Internet Header Plus the First 64 Bits of the Original Datagrams Data.
These data are used by the host to match the message to the appropriate
process. If a higher-level protocol uses port numbers, they are assumed to be
in the first 64 data bits of the original datagrams data.
Description. If, according to the information in the gateways routing
tables, the network specified in the internet destination field of a datagram is
unreachable e.g., the distance to the network is infinity., the gateway may
send a destination unreachable message to the internet source host of the
datagram. In addition, in some networks the gateway may be able to
determine if the internet destination host is unreachable. Gateways in these
networks may send destination unreachable messages to the source host
when the destination host is unreachable.
If, in the destination host, the IP module cannot deliver the datagram
because the indicated protocol module or process port is not active, the
destination host may send a destination unreachable message to the source
host.
Another case is when a datagram must be fragmented to be forwarded by
a gateway, yet the Dont Fragment flag is on. In this case the gateway must
discard the datagram and may return a destination unreachable message.
It should be noted that Codes 0, 1, 4, and 5 may be received from a
gateway; Codes 2 and 3 may be received from a host.
Section 11.4.4.3 is based on RFC 792 Ref. 14..
402
Figure 11.27. Protocol layers showing the relationship of TCP with other layered protocols.
11.4.5
The Transmission Control Protocol
11.4.5.1 TCP Defined. TCP Refs. 15 and 16. was designed to provide
reliable communication between pairs of processes in logically distinct hosts
on networks and sets of interconnected networks. TCP operates successfully
in an environment where network congestion and the loss, damage, duplication, or misorder of data can occur. This robustness in spite of unreliable
communications media makes TCP well-suited to support commercial, military, and government applications. TCP appears at the transport layer of the
protocol hierarchy. Here, TCP provides connection-oriented data transfer
that is reliable, ordered, full-duplex, and flow-controlled. TCP is designed to
support a wide range of ULPs.* The ULP can channel continuous streams of
data through TCP for delivery to peer ULPs. The TCP breaks the streams
into portions which are encapsulated together with appropriate addressing
and control information to form a segmentthe unit of exchange between
TCPs. In turn, the TCP passes the segments to the network layer for
transmission through the communication system to the peer TCP.
As shown in Figure 11.27, the layer below the TCP in the protocol
hierarchy is commonly the IP layer. The IP layer provides a way for the TCP
to send and receive variable-length segments of information enclosed in
internet datagram envelopes. The internet datagram provides a means for
addressing source and destination TCPs in different networks. The IP also
deals with fragmentation and reassembly of TCP segments required to
*ULP stands for upper layer protocol. Typically, there are three ULPs: file transfer protocol
FTP., simple mail transfer protocol SMTP., telnet protocol. See Figure 11.18.
403
achieve transport and delivery through the multiple networks and interconnecting gateways. The IP also carries information on the precedence, security
classification, and compartmentation of the TCP segments, so this information can be communicated end-to-end across multiple networks.
11.4.5.2 TCP Mechanisms. TCP builds its services on top of the network
layers potentially unreliable services with mechanisms such as error detection, positive acknowledgments, sequence numbers, and flow control. These
mechanisms require certain addressing and control information to be initialized and maintained during data transfer. This collection of information is
called a TCP connection. The following paragraphs describe the purpose and
operation of the major TCP mechanisms.
PAR Mechanism. TCP uses a positive acknowledgment with retransmission
PAR. mechanism to recover from the loss of a segment by the lower layers.
The strategy with PAR is for a sending TCP to retransmit a segment at timed
intervals until a positive acknowledgment is returned. The choice of retransmission interval affects efficiency. An interval that is too long reduces data
throughput, whereas one that is too short floods the transmission media with
superfluous segments. In TCP, the timeout is expected to be dynamically
adjusted to approximate the segment round-trip time plus a factor for
internal processing; otherwise performance degradation may occur. TCP uses
a simple checksum to detect segments damaged in transit. Such segments are
discarded without being acknowledged. Hence, damaged segments are treated
identically to lost segments and are compensated for by the PAR mechanism.
TCP assigns sequence numbers to identify each octet of the data stream.
These enable a receiving TCP to detect duplicate and out-of-order segments.
Sequence numbers are also used to extend the PAR mechanism by allowing a
single acknowledgment to cover many segments worth of data. Thus, a
sending TCP can still send new data although previous data have not been
acknowledged.
Flow Control Mechanism. TCPs flow control mechanism enables a receiving
TCP to govern the amount of data dispatched by a sending TCP. The
mechanism is based on a window which defines a contiguous interval of
acceptable sequence-numbered data. As data are accepted, TCP slides the
window upward in the sequence number space. This window is carried
in every segment, enabling peer TCPs to maintain up-to-date window
information.
Multiplexing Mechanism. TCP employs a multiplexing mechanism to allow
multiple ULPs within a single host and multiple processes in a ULP to use
TCP simultaneously. This mechanism associates identifiers, called ports, to
ULP processes accessing TCP services. A ULP connection is uniquely identified with a socket, the concatenation of a port and an internet address. Each
connection is uniquely named with a socket pair. This naming scheme allows
404
a single ULP to support connections to multiple remote ULPs. ULPs which

provide popular resources are assigned permanent sockets, called well-known
sockets.
11.4.5.3 ULP Synchronization. When two ULPs wish to communicate,
they instruct their TCPs to initialize and synchronize the mechanism of
information flow on each to open the connection. However, the potentially
unreliable network layer i.e., internet protocol layer. can complicate the
process of synchronization. Delayed or duplicate segments from previous
connection attempts might be mistaken for new ones. A handshake procedure with clock-based sequence numbers is used in connection opening to
reduce the possibility of such false connections. In the simplest handshake,
the TCP pair synchronizes sequence numbers by exchanging three segments
thus the name three-way handshake.
11.4.5.4 ULP Modes. A ULP can open a connection in one of two modes:
passive or active. With a passive open, a ULP instructs its TCP to be receptie
to connections with other ULPs. With an active open, a ULP instructs its
TCP to actually initiate a three-way handshake to connect to another ULP.
Usually an active open is targeted to a passive open. This activerpassive
model supports server-oriented applications where a permanent resource,
such as a database management process, can always be accessed by remote
users. However, the three-way handshake also coordinates two simultaneous
active opens to open a connection. Over an open connection, the ULP pair
can exchange a continuous stream of data in both directions. Normally, TCP
groups the data into TCP segments for transmission at its own convenience.
However, a ULP can exercise a push service to force TCP to package and
send data passed up to that point without waiting for additional data. This
mechanism is intended to prevent possible deadlock situations where a ULP
waits for data internally buffered by TCP. For example, an interactive editor
might wait forever for a single input line from a terminal. A push will force
data through the TCPs to the awaiting process. A TCP also provides the
means for a sending ULP to indicate to a receiving ULP that urgent data
appear in the upcoming data stream. This urgent mechanism can support, for
example, interrupts or breaks. When a data exchange is complete, the
connection can be closed by either ULP to free TCP resources for other
connections. Connection closing can happen in two ways. The first, called a
graceful close, is based on the three-way handshake procedure to complete
data exchange and coordinate closure between the TCPs. The second, called
an abort, does not allow coordination and may result in the loss of unacknowledged data.
11.4.5.5 Scenario. The following scenario provides a walk-through of a
connection opening, data exchange, and connection closing as might occur
between the database management process and the user mentioned previously. The scenario focuses more on a. the three-way handshake mechanism
in connection with opening and closing and b. the positive acknowledgment
405
with retransmission mechanism supporting reliable data transfer. Although

not pictured, the network layer transfers the information between TCPs. For
the purpose of this scenario, the network layer is assumed not to damage,
lose, duplicate, or change the order of data unless explicitly noted. The
scenario is organized into three parts:
1. A simple connection opening steps 17.
2. Two-way data transfer steps 817.
3. A graceful connection close steps 1824.
Scenario Notation. The following notation is used in Figures 11.2811.30:
Start at Figure 11.28.
Figure 11.28. Two examples of a simple connection opening.
406
1. ULP B the DB manager. issues a PASSIVE OPEN to TCP B to

prepare for connection attempts from other ULPs in the system.
2. ULP A the user. issues an ACTIVE OPEN to open a connection to
ULP B.
3. TCP A sends a segment to TCP B with an OPEN control flag, called
an SYN, carrying the first sequence number shown as SEQ200. it
will use for data sent to B.
4. TCP B responds to the SYN by sending a positive acknowledgment, or
ACK, marked with the next sequence number expected from TCP A.
In the same segment, TCP B sends its own SYN with the first
sequence number for its data SEQ550..
5. TCP A responds to TCP Bs SYN with an ACK showing the next
sequence number expected from B.
6. TCP A now informs ULP A that a connection is open to ULP B.
7. Upon receiving the ACK, TCP B informs ULP B that a connection
has been opened to ULP A.
Turn to Figure 11.29.
Figure 11.29. Two examples of two-way data transfer.
407
8. ULP A passes 20 octets of data to TCP A for transfer across the open
connection to ULP B.
9. TCP A packages the data in a segment marked with current A
sequence number.
10. After validating the sequence number, TCP B accepts the data and
delivers it to ULP B.
11. TCP B acknowledges all 20 octets of data with the ACK set to the
sequence number of the next data octet expected.
12. ULP B passes 125 bytes of data to TCP B for transfer to ULP A.
13. TCP B packages the data in a segment marked with the B sequence
number.
14. TCP A accepts the segment and delivers the data to ULP A.
15. TCP A returns an ACK of the received data marked with the sequence number of the next expected data octet. However, the segment
is lost by the network and never arrives at TCP B.
16. TCP B times out waiting for the lost ACK and retransmits the
segment. TCP A receives the retransmitted segment, but discards it
Figure 11.30. Two examples of a graceful connection close.
408
because the data from the original segment have already been accepted. However, TCP A re-sends the ACK.
17. TCP B gets the second ACK.
Turn to Figure 11.30.
18. ULP A closes its half of the connection by issuing a CLOSE to TCP A.
19. TCP A sends a segment marked with a CLOSE control flag, called a
FIN, to inform TCP B that ULP A will send no more data.
20. TCP B gets the FIN and informs ULP B that ULP A is closing.
21. ULP B completes its data transfer and closes its half of the connection.
22. TCP B sends an ACK of the first FIN and its own FIN to TCP A to
show ULP Bs closing.
23. TCP A gets the FIN and the ACK, then responds with an ACK to
TCP B.
24. TCP A informs ULP A that the connection is closed. Not pictured. .
TCP B receives the ACK from TCP A and informs ULP B that the
connection is closed.
11.4.5.6 TCP Header Format. The TCP header format is shown in Figure
11.31. It should be noted that TCP works with 32-bit segments.
Figure 11.31. TCP header format. Note that one tick mark represents one bit position.
409
Source Port. The port represents the source ULP initiating the exchange.
The field is 16 bits long.
Destination Port. This is the destination ULP at the other end of the
connection. This field is also 16 bits long.
Sequence Number. Usually, this value represents the sequence number of
the first data octet of a segment. However, if an SYN is present, the
sequence number is the initial sequence number ISN. covering the SYN;
the first data octet is then numbered SYN q 1. The SYN is the syncrhonize control flag. It is the opening segment of a TCP connection. SYNs
are exchanged from either end. When a connection is to be closed, there is
a similar FIN sequence exchange.
Acknowledgment Number. If the ACK control bit* is set bit 2 of the 6-bit
control field., this field contains the value of the next sequence number
that the sender of the segment is expecting to receive. This field is 32 bits
long.
Data Offset. This field indicates the number of 32-bit words in the TCP
header. From this value the beginning of the data can be computed. The
TCP header is an integral number of 32 bits long. The field size is 4 bits.
Reserved. This is a field of 6 bits set aside for future assignment. It is set to
zero.
Control Flags. The field size is 6 bits covering six items 1 bit per item.:
1.
2.
3.
4.
5.
6.
URG: Urgent pointer field significant

ACK: Acknowledgment field significant
PSH: Push function
RST: Reset the connection
SYN: Synchronize sequence numbers
FIN: No more data from sender
Window. The number of data octets beginning with the one indicated in the
acknowledgment field which the sender of this segment is willing to
accept. The field is 2 octets in length.
Checksum. The checksum is the 16-bit ones complement of the ones complement sum of all 16-bit words in the header and text. The checksum also
covers a 96-bit pseudo-header conceptually prefixed to the TCP header.
This pseudo-header contains the source address, the destination address,
the protocol, and TCP segment length.
*ACK is a control bit acknowledge. occupying no sequence space, which indicates that the
acknowledgment field of this segment specifies the next sequence number the sender of this
segment is expecting to receive, thereby acknowledging receipt of all previous sequence
numbers.
410
Urgent Pointer. This field indicates the current value of the urgent pointer
as a positive offset from the sequence number in this segment. The urgent
pointer points to the sequence number of the octet following the urgent
data. This field is only to be interpreted in segments with the URG control
bit set. The urgent point field is two octets long.
Options. This field is variable in size; and, if present, options occupy space at
the end of the TCP header and are a multiple of 8 bits in length. All
options are included in the checksum. An option may begin on any octet
boundary. There are two cases of an option:
1. Single octet of option kind
2. An octet of option kind, an octet of option length, and the actual
option data octets
Options include end of option list, no-operation, and maximum
segment size.
Padding. The field size is variable. The padding is used to ensure that the
TCP header ends and data begin on a 32-bit boundary. The padding is
composed of zeros.
11.4.5.7 TCP Entity State Diagram. Figure 11.32 summarizes TCP operation. It shows a TCP state diagram.
Section 11.4 is based on various RFCs, MIL-STD-1778, and Refs. 9, 15, and 16.
11.4.6
Internet Protocol Version 6 (IPv6)
IPv6 is an evolutionary step improvement on IPv4. The principal problem

with IPv4 was its addressing, a 32-bit field. IPv6 has a 128-bit field, which
supports more levels of addressing.
What happened to IPv5? It is covered by RFC 1819 and is called ST2 q .
It serves as an adjunct to IP where a guaranteed QoS is required.
In IPv6 the header has been simplified and certain IPv4 fields have been
dropped. The IPv6 header is a static 40 octets long.
A new capability has been added to enable labeling of packets belonging
to particular traffic flows for which the sender requests special handling,
such as nondefault QoS or real-time service.
There have also been added authentication and privacy capabilities and
improved data integrity.
11.4.6.1 IPv6 Routing. Prefix routing is used with IPv6 where every address has an associated prefix. This is just a mask identifier which indicates
how many bits, starting from the left are used for routing and how many
411
Figure 11.32. TCP entity state diagramsummary.
identify a host. The routing prefix is a tool for building routing tables. End
stations make the prefix similar to the mask described in Chapter 10.
Existing protocols may use IPv6 addresses as well. However, they will have
to change to understand 128-bit addressing as one might imagine.
11.4.6.2 IPv6 Header. Figure 11.33 illustrates the IPv6 header. Compare
this to Figure 11.23, which shows the IPv4 header. All the fields have been
changed except the Vers version. field.
The following is a description of the IPv6 header fields:
Version: 4 bits. Internet protocol version number s 6.

Traffic class: 8 bits.
Flow label: 20 bits.
412
Figure 11.33. IPv6 header format. (From RFC-2460, Ref. 17.)
Payload length: 16 bits. Unsigned integer. It is the length of the IPv6

payload, measured in octets. Note: Any extension headers present are
considered part of the payload.
Next header: 8 bits. This means it is an 8-bit selector. It identifies the
type of header immediately following the IPv6 header. It uses the same
values as the IPv4 protocol field see RFC-1700..
Hop limit: 8 bits. Unsigned integer. Decremented by 1 by each node
that forwards the packet. The packet is discarded if the hop limit is
decremented to zero.
Source address: 128 bits. Address of the originator of the packet.
Destination address: 128 bits. Address of the intended recipient of the
packet. This may possibly not be the ultimate recipient if a routing
header is present.
Discussion on Selected Fields

Traffic Class. The 8-bit Traffic Class field is available for use by originating
nodes andror forwarding routers to identify and distinguish between different classes or priorities of IPv6 packets. At the point in time at which
specification was being written, there are a number of experiments underway
in the use of the IPv4 Type of Service andror Precedence bits to provide
various forms of Differentiated Service for IP packets, other than through
the use of explicit flow setup. The Traffic Class field in the IPv6 header is
intended to allow similar functionality to be supported in IPv4. It is hoped
by the IETFrInternet Society that those experiments will eventually lead to
413
agreement on what sorts of traffic classifications are most useful for IP

packets. Detailed definitions of the syntax and semantics of all or some of the
IPv6 Traffic Class bits, whether experimental or intended for eventual
standardization, are to be provided in separate documents.
The following general requirements apply to the Traffic Class field:
The service interface to the IPv6 service within a node must provide a
means for an upper-level protocol to supply the value of the Traffic
Class bits in packets originated by that upper-layer protocol. The default
value must be zero for all 8 bits.
Nodes that support a specific experimental or eventual standard . use of
some or all of the Traffic Class bits are permitted to change the value of
those bits in packets that they originate, forward or receive, as required
for that specific use. Nodes should ignore and leave unchanged any bits
of the Traffic Class field for which they do not support a specific use.
An upper-layer protocol must not assume that the value of the Traffic
Class bits in a received packet are the same as the value sent by the
packets source.
Flow Labels. The 20-bit Flow Label field in the IPv6 header may be used by
a source to label sequences of packets for which it requests special handling
by IPv6 routers, such as nondefault QoS or real time service. This aspect of
IPv6 is still experimental and subject to change as the requirements for flow
support in the Internet become clearer. Hosts or routers that do not support
the functions of the Flow Label field are required to set the field to zero
when originating a packet, pass the field on unchanged when forwarding a
packet, and ignore the field when receiving a packet.
IPv6 Addressing. There are three types of addresses:
Unicast. An identifier for a single interface. A unique address delivered to
a single notation.
Anycast. New for IP, an anycast address is an identifier for a set of
interfaces typically belonging to different nodes.. This is similar to
multicast, but a packet sent to an anycast address is delivered to one of
the interfaces identified by that address the nearest one, according
to the routing protocols measure of distance ..
Multicast. An identifier for a set of interfaces typically belonging to
different nodes.. A packet sent to a multicast address is delivered to all
interfaces identified by that address.
The broadcast address is not defined in IPv6. It was superseded by the
multicast address Ref. 9..
IPv6 is based on RFC 2460 Ref. 17..
414
11.5
USER DATAGRAM PROTOCOL (UDP)
UDP provides a procedure for application programs to send messages to

other programs with a minimum of protocol mechanism. The protocol is
transaction oriented and delivery and duplicate protection are not guaranteed. Applications requiring ordered reliable delivery of streams of data
should use the transmission control protocol see Section 11.4.5.. UDP
assumes that the internet protocol IPsee Section 11.4. is used as the
underlying protocol.
11.5.1
UDP Header Format and Fields
The UDP datagram header format is shown in Figure 11.34. The following is
a description of the fields in Figure 11.34.
Source Port is an optional field, when meaningful, that indicates the port
of the sending process. It may be assumed to be the port to which a reply
should be addressed in the absence of any other information. If not used, a
value of zero is inserted.
Destination Port has a meaning within the context of a particular IP
destination address.
Length is the length in octets of this user datagram including this header
and the data field. There is minimum value of length of eight octets..
Checksum is the 16-bit ones complement of the ones complement sum of
a pseudo-header of information from the IP header, the UDP header, and
the data, padded with zero octets at the end if necessary . to make a multiple
of 2 octets.
The pseudo-header conceptually prefixed to the UDP header contains the
source address, the destination address, the protocol, and the UDP length. It
is shown in Figure 11.35. This information gives protection against misrouted
datagrams. The checksum procedure is the same as used in TCP.
If the computed checksum is zero, it is transmitted as all ones the
equivalent in ones complement arithmetic .. An all-zero transmitted check-
Figure 11.34. UDP datagram header format. (From RFC 768, Ref. 18.)
THE CLNP PROTOCOL BASED ON ISO 8473
415
Figure 11.35. Format of pseudo-header. (Ref. 18)
sum value means that the transmitter generated no checksum for debugging
or for higher level protocols that dont care..
User Interface. A user interface allows the creation of new receive port. It
also allows receive operations on the receive ports that return the data octets
and an indication of source port and source address, and an operation that
allows a datagram to be sent, specifying the data, source, and destination
ports and addresses to be sent.
IP Interface. The UDP module must be able to determine the source and
destination internet addresses and the protocol field from the internet
header. One possible UDPIP interface would return the whole internet
datagram including all of the internet header in response to a receive
operation. Such an interface would also allow the UDP to pass a full internet
datagram complete with header to the IP to send. The IP would verify
certain fields for consistency and compute the internet header checksum.
Protocol Application. The major uses of this protocol is the Internet Name
Server and the Trivial File Transfer. This is protocol number 17 21 octal.
when used in the internet protocol. Other protocol numbers are listed in
RFC 762.
Section 11.5 consists of abstracts from RFC 768.
11.6
THE CLNP PROTOCOL BASED ON ISO 8473 Ref. 19.
CLNP connectionless network protocol. was designed to be used in the

context of the interworking protocol approach to the provision of the connectionless mode network service defined in ISO 8648. It is intended for use in
the subnetwork independent convergence protocol SNICP. role. CLNP
operates to construct the OSI Network Service over a defined set of underlying services. CLNP was planned as a replacement of IPv4, however, IPv6 and
other improvements resulted in the IT community remaining with IPv4 and
IPv6. We have included CLNP, sometimes referred to as TUBA TCP and
416
Figure 11.36. CLNP frame format, DT PDU. (From Figure 10, page 21, ISO-8473, Ref. 19.)
UDP over bigger addresses ., because it is constantly referenced in the

literature with some application in Europe.
11.6.1
CLNP Frame Formats
There are two different frame types used with CLNP. These are Data DT .
PDU and Error Report PDU. The formats for these frames are illustrated in
Figures 11.36 and 11.37. Each consists of a fixed part and a variable part. The
fixed part consists of those fields from the Network Layer Protocol Identifier
through the Checksum.
11.6.2
Frame Field DescriptionFIXED PART
Network Layer Protocol Identifier 1 octet.. The value of this field is set to
binary 1000 0001 to identify this network layer protocol as ISO 8473. The
value of this field is set to binary 0000 0000 to identify the inactive layer
protocol subset.
417
Figure 11.37. CLNP frame format, Error Report PDU. (From Figure 12, page 22, ISO-8473,
Ref. 19.)
Length Indicator. The length is indicated by a binary number, with a maximum value of 254 binary 1111 1110.. The length indicated is the length in
octets of the header. The value 255 1111 1111. is reserved for possible future
extensions.
Version r Protocol Identifier Extension 1 octet.. The value of this field is
binary 0000 0001, which identifies the standard Version 1 of ISO 8473.
PDU Lifetime 1 octet.. The PDU Lifetime field is encoded as a binary
number representing the remaining lifetime of the PDU, in units of 500 ms.
Flags
Segmentation Permitted (SP) 1 bit.. The Segmentation Permitted flag indicates whether segmentation is permitted. Its value is determined by the
originator of the PDU and cannot be changed by any other network entity for
the lifetime of the Initial PDU and any Derived PDUs.
418
A value of 1 indicates that segmentation is permitted. A value of 0

indicates that segmentation is not permitted. When the value of 0 is selected,
the segmentation part of the PDU header is not present, and the value of the
Segment Length field gives the total length of the PDU.
More Segments (MS) 1 bit.. The More Segments flag indicates whether or
not the data part of this PDU contains as its last octet. the last octet of the
User Data in the NSDU Network Service Data Unit.. When the More
Segments flag is set to 1, segmentation has occurred and the last octet of the
NSDU is not contained in this PDU. The More Segments flag cannot be set
to 1 if the Segmentation Permitted flag is not set to 1.
When the More Segments flag is set to 0, the last octet of the Data Part of
the PDU is the last octet of the NSDU.
Error Report ErR 1 bit.. When the Error Report flag is set to 1, the rules in
Section 6.10 of Ref. 19 are used to determine whether to generate an Error
Report PDU if it is necessary to discard this Data PDU.
When the Error Report flag is set to 0, discard of the Data PDU will not
cause the generation of an Error Report PDU.
Type Code 5 bits.. The Type Code identifies the type of PDU Data of
Error Report.. If bits 5, 4, 3, 2, 1 are set to 111000, it is a Data PDU; if they
are set to 00001, it is an Error Report PDU.
PDU Segment Length 2 octets.. The Segment Length field specifies the
entire length, in octets, of the PDU, including both header and data if
present .. When the full protocol is employed and a PDU is not segmented,
the value of this field is identical to the value of the Total Length field
located in the Segmentation Part of the header.
When the nonsegmenting protocol subset is employed, no segmentation part is present in the header. In this case, the Segment Length field
specifies the entire length of the Initial PDU, including both header and data
if present .. The Segment Length field is not changed for the lifetime of
the PDU.
PDU Checksum 2 octets.. The checksum is computed on the entire PDU
header. For a Data PDU, this includes the segmentation and option parts if
present .. For an Error Report PDU, this includes the reason for discard field
as well.
A checksum value of zero is reserved to indicate that the checksum is to
be ignored. The operation of the PDU Header Error Detection function
ensures that the value zero does not represent a valid checksum. A nonzero
value indicates that the checksum will be processed.
419
Address Part
Destination and Source Address Variable length.. The Destination and
Source Address used by this protocol are Network Service Access Point
address as defined in ISO 8348rAdd.2. The Destination and Source Addresses are of variable length. The Destination and Source Address fields are
encoded as Network Protocol Address Information using the Preferred
Binary Encoding defined in Section 8.3.1 of ISO 8348rAdd.2.
The field consists of a subfield 1 octet long which is the Destination
Address Length Indicator, then the Destination Address itself of variable
length. Then there is the Source Address Length Indicator subfield that is
1 octet long. This is followed by the Source Address itself of variable
length.
Segmentation Part. If the Segmentation Permitted Flag in the Fixed Part of
the PDU Header is set to 1, the segmentation part of the header will be
present. If the Segmentation Permitted flag is set to 0, the segmentation part
is not present the nonsegmenting protocol subset is in use..
The Segmentation part consists of three fields, each 2 octets long. These
are Data Unit Identifier, Segment Offset, and Total Length. These three
fields are described as follows.
Data Unit Identifier 2 octets.. The Data Unit Identifier identifies an Initial
PDU and hence, its Derived PDUs. so that a segmented data unit may be
correctly reassembled.
Segment Offset 2 octets.. For each Derived PDU, the Segment Offset field
specifies the relative position of the segment contained in the Data Part of
the Derived PDU with respect to the start of the Data Part of the Initial
PDU. The offset is measured in units of octets. The offset of the first
segment and hence the initial PDU. is 0, an unsegmented Initial . PDU has
a segment offset value of 0. The value of this field is a multiple of 8.
PDU Total Length 2 octets.. The Total Length Field specifies the entire
length of the Initial PDU in octets, including both the header and data. This
field is not changed for the lifetime of the Initial PDU and hence its Derived
PDUs..
Options Part. If the options part is present, it may contain one or more
parameters. The number of parameters that may be contained in the options
part is constrained by the length of the options part, which is determined by
the formula
Options part length s PDU header length
y length of fixed part q length of address part
q length of segmentation part .
and by the length of the individual optional parameters.
420
TABLE 11.6
Encoding Option Parameters
Octets
Field
n
nq1
nq2
to
nqmq1
Parameter Code
Parameter Length (e.g., m )
Parameter Value
Parameters defined in the options part may appear in any order. Duplication of options is not permitted. Receipt of a Protocol Data Unit with a
duplicated option should be treated as a protocol error. The rules governing
the treatment of protocol errors are described in Section 6.10 Error Reporting Function. of the reference document.
The encoding parameters contained within the options part are listed in
Table 11.6.
The Parameter Code field 1 octet. is encoded in binary and provides for
a maximum of 255 different parameters. No parameter code uses bits 8
and 7 with the value 00, so the actual maximum number of parameters is
lower. A parameter code of 255 binary 1111 1111. is reserved for possible
extensions.
The Parameter Length 1 octet. field indicates the length, in octets, of the
Parameter Value field. The length is indicated by a positive binary number, m,
with a theoretical maximum value of 254. The practical maximum value of m
is lower. For example, in the case of a single parameter contained within the
options part, 2 octets are required for the parameter code and the parameter
length indicators. Thus, the value of m is limited to
m s 252 y length of fixed part q length of address part
qlength of segmentation part .
Accordingly, for each successive parameter the maximum value of m decreases.
The parameter field contains the value of the parameter identified in the
parameter code field. The following parameters are permitted in the options
part.
Parameters Permitted
Padding. The padding parameter is used to lengthen the PDU header to a
convenient size as described in Section 6.12 of the reference document.
Parameter Code:
Parameter Length:
Parameter Value:
1100 1100
variable
any value is allowed
421
Security. This parameter allows a unique and unambiguous security level to

be assigned to a PDU. See Section 6.13 of the reference document..
Parameter Code:
Parameter Length:
Parameter Value:
1100 0101
variable
The high-order 2 bits of the first octet specify the
Security Format Code where:
Security Field Format

00
01
10
11
Type of Security Field

Reserved
Source Address Specific
Destination Address Specific
Globally Unique
The rest of the first octet is reserved and is set to 0. The remainder of
the Parameter Value field specifies the security level.
Source Address Specific and the Destination Address Specific fields are
coded 01 and 10, respectively, as indicated in the above table. The
Security Format Code values of 01 and 10 indicate that the remaining
octets of the parameter value field specify a security level which is
unique and unambiguous in the context of the security classification
system employed by the authority responsible for assigning the source
NSAP network service access point. Address and the destination
NSAP Address, respectively.
Globally Unique is coded 11 and is similar to the above. However, in this
case the security classification system is not specified.
Source Routing. Source routing, which is another parameter, specifies,
either completely or partially, the route to be taken from Source Network
Address to Destination Network Address.
Parameter Code:
Parameter Length:
Parameter Value:
1100 1000
variable
2 octets of control information followed by a concatenation of network-entity title entries ordered from
source to destination
The first octet of the parameter value is the type code, and has the
following significance:
0000 0000 partial source routing
0000 0001 complete source routing
all other values are reserved.
422
The second octet indicates the octet offset of the next network-entity title
entry to be processed in the list. It is relative to the start of the parameter,
such that a value of 3 indicates that the next network-entity title entry begins
immediately after this control octet. Successive octets are indicated by
correspondingly larger values of this indicator.
The third octet begins the network-entity title list. The list consists of
variable length network-entity title entries. The first octet of each entry gives
the length of the network-entity title which comprises the remainder of the
entry.
Recording of Route. The recording of route parameter identifies the intermediate systems traversed by the PDU.
Parameter Code:
Parameter Length:
Parameter Value:
1100 1011
variable
2 octets of control information followed by a concatenation of network-entity title entries ordered from
source to destination
The first octet of the parameter value is the type code and has the
following significance:
0000 0000 Partial Recording of Route progress
0000 0001 Complete Recording of Route progress
The second octet identifies the first octet not currently used for a recorded
network-entity title, and, therefore, also the end of the list. It is encoded
relative to the start of the parameter value, such that a value of 3 indicates
that no network-entity titles have yet been recorded. A value of all 1s is used
to indicate that route recording has been terminated.
The third octet begins the network-entity title list. The list consists of
variable length network-entity title entries. The first octet of each entry gives
the length of the network-entity title comprising the remainder of the entry.
Network-entity title entries are always added to the end of the list.
Other Parameters. There is a Quality of Service QoS. Maintenance parameter which conveys information about quality of service requested by the
originating Network-Service user. Another parameter is the Priority parameter that indicates the relative priority of the PDU. Intermediate systems that
support this option make use of this information in routing and ordering
PDUs for transmission.
11.6.3
423
Error Report PDU (ER)
The format of the Error Report PDU is shown in Figure 11.37. This format,
similar to the Data PDU, contains a fixed part group of fields and a variable
part. The only difference in the fixed parts is that the flag bits SP, MS, and
ErR. are always set to binary 0. The Destination Address specifies the
network-entity title of the originator of the discarded PDU. The Source
Address specifies the title of the intermediate-system or end-system
network-entity initiating the Error Report PDU.
11.6.3.1 Reason for Discard. This parameter is valid only for the Error
Report PDU.
Parameter Code:
Parameter Length:
Parameter Value:
1100 0001
2 octets
type of error encoded in binary
The parameter values are listed in Table 11.7.
TABLE 11.7
Reason for Discard Parameter Values
Parameter
Value
Class of
Error
Meaning
0000
0000
0001
0010
0011
0100
0101
0110
0111
General
Reason not specified

Protocol Procedure Error
Incorrect Checksum
PDU Discarded due to Congestion
Header Syntax Error (cannot be parsed)
Segmentation needed but not permitted
Incomplete PDU Received
Duplicate Option
1000
0000
0001
0000
0001
0010
0011
Address
Destination Address Unreachable

Destination Address Unknown
Unspecified Source Routing Error
Syntax Error in Source Routing Field
Unknown Address in Source Routing Field
Path not Acceptable
1010
0000
0001
Lifetime
Lifetime Expired while Data Unit in Transit

Lifetime Expired during Reassembly
1011
0000
0001
0010
0011
0100
PDU
Discarded
Unsupported Option not Specified

Unsupported Protocol Version
Unsupported Security Option
Unsupported Source Routing Option
Unsupported Recording of Route Option
1100
0000
Reassambly
Reassembly Interference
1001
Source
Routing
Source: Table 7, page 23, ISO 8473, Ref. 19.
424
The first octet of the parameter value contains an error type code. If the
error in the discarded Data PDU can be localized to a particular field,
the number of the first octet of that field is stored in the second octet of the
reason for discard parameter field; or if the error is a checksum error, then
the value 0 is stored in the second octet of the reason for discard field.
Error Report Data Part. This field contains the entire header of the discarded
Data PDU and may contain none, some, or all of the Data Part of the
discarded PDU.
Section 11.6 has been abstracted from ISO 8473, 1988 Ref. 19..
11.7
11.7.1
NETWORKING VIA VSATS

Rationale of VSAT Networks
There are several reasons for implementing a very-small-aperture terminal

VSAT. network. Much depends on the area of the world or even the country
one is in. In industrial nations, VSAT networks are used as private data
networks. The overwhelming application is to provide a star network connectivity as shown in Figure 11.38 where the hub is colocated with corporate
headquarters and the VSAT terminals are in branch locations. These are
generally data networks that feed data information to their headquarters
Figure 11.38. Elements and architecture of a VSAT network.
425
facility. In many cases such circuits are two-way. They are particularly
attractive for large, dispersed industries and sales organizations. These networks completely bypass local telephone companies that otherwise might
have provided such service.
In underdeveloped nations the rationale for VSAT networks is less economic and more service-oriented. In these areas the local telephone companies or administrations provide such poor service that owning ones own
network becomes very attractive. Specialized data protocols developed to
optimize usage are commonly used on hub-and-spoke configurations.
The hub-and-spoke architecture of Figure 11.38 is the most common
topology. However, another topology is gaining popularity, especially in
developing nations. This is a full mesh network where each and every VSAT
terminal has connectivity to all other terminals in the network. There is no
hub. Their antennas are larger and provide greater RF power to accommodate higher bit rates, often DS1 or E1 configurations see Chapter 7.. They
provide both voice and data connectivity. We can expect to find the more
common type of data protocols applied on these circuits.
11.7.2
Basic Description of a VSAT Network
The most common type of VSAT network has the hub-and-spoke configuration* as shown in Figure 11.38. There are three elements to the network: 1.
the hub, which is a rather large facility, 2. the VSAT terminals, which are
small, and 3. the satellite. The satellite will be in geostationary orbit. The
round-trip delay, some 500 ms, may pose a problem.
The term VSAT contains the expression very small aperture, meaning a
terminal with a small antenna. Most satellite terminal antennas are dishes.
A dish means a parabolic antenna. VSAT antennas have diameters of the
parabolic dish. ranging from 0.8 m to 2.0 m. Hub stations have antenna
diameters ranging from 5 m to 8 m. Aside from the hub, everything is small in
a VSAT network. There seems to be no limit in the number of VSAT
terminals a hub can service. One case we are familiar with has over 2000
accesses. Because the multiplier is 2000, to keep the system viable economically, the cost of a VSAT terminal must be reduced just as much as possible.
Of course, all terminals in the network must be within line-of-sight of the
satellite being used. In fact, the terminals must have elevation angles greater
than 5 degrees for C-band i.e., 6r4 GHz. operation and greater than 10, or
even 15, degrees of Ku-band i.e., 14r12 GHz. operation. The definition of
elevation angle is shown in Figure 11.39.
11.7.3
Disadvantages of a VSAT Network
There are essentially three drawbacks to the employment of VSAT networks.

The first two deal with survivability. The problem is that there are two critical
*This is a star network, if you will.
426
Figure 11.39. Definition of elevation angle for satellite communications. D, distance to the
satellite; , elevation angle.
single-point failure locations. The first is the satellite. If we lose the satellite,
the network goes down. The second is the hub. If we lose the hub for
whatever reason, the network also goes down.
One alternative is to lease standby transponder space on another nearby
satellite. This is fraught with danger. By nearby, we mean a satellite in a
position such that antennas do not have to be slewed far to acquire the
standby satellite. This reacquiring is not as easy as it seems. To keep costs
low, we have no automatic pointing and tracking capability on the VSAT
terminal antennas. Thus a well-trained technician must visit each site to go
through the reacquisition exercise. Some terminals may be kept down for
days because of this.
By regarding the hub as a single-point failure, we can build or lease a
standby hub. Either way, it implies added cost.
The third drawback is that on conventional VSAT networks, for one
VSAT to communicate with another requires routing through the hub involving a double hop i.e., from VSAT1 up to satellite and down to hub, and from
hub up to satellite and down to VSAT2.. This is not done for voice
operations, if that is what we want. It is not done because of excessive delay.
If we use the proper protocols, the delay will have little effect on data
connectivities. Seldom, however, is there a requirement for one VSAT to
direct traffic to another. Nearly always the direction of traffic is from VSAT
to hub, from hub to VSAT, or both. For instance, if it is a hotel chain
installing a system, the hub will be located with the main reservation
computer and service desk.
11.7.4
427
Important Design Variables
Probably the largest recurring cost of a VSAT network is space segment

charges. By this we mean the cost of leasing transponder space from the
satellite service provider.
Space segment charges are usually based on, or are a function of,
transponder bandwidth required for the network. To reduce these recurring
charges, we have to reduce the bandwidth as much as possible for a certain
service quality.
How can we measure service quality? Several measures can come to mind
besides BER. Time to carry out a transaction could well be one. However, for
some data services, time to carry out a transaction may have secondary
importance. Throughput, if the term is well-defined, is another measure of
service quality.
Bit error rate BER. behavior can be pretty well controlled except on
VSAT networks using the Ku-band and higher frequencies. In frequency
bands above 10 GHz, heavy rainstorms can severely degrade BER over the
period of the storm. It causes a form of fading that is very similar to
multipath fading except for some of the dispersive properties. There are
three steps that can be taken to mitigate the effects of rainfall fading:
1. Accept the condition and let ARQ carry out its function. Throughput
will degrade accordingly for the period of the rainstorm.
2. Overbuild the system as one would do to build up fade margin. This
can be done by using larger antennas, increasing transmitter power and
improving receiver noise performance. A subset of this is to make the
circuit adaptive and only increase transmit power during the storm
event.
3. Use channel coding with interleaving. This is a very effective measure if
we choose the interleaving interval to be longer statistically . than fade
duration. Such a measure is fine for data circuits, but care must be
taken on voice circuits because of the increase in delay caused by the
interleaver.
Coding refers to channel coding as opposed to source coding.. With
channel coding we systematically add redundancy to correct errors at the
receiver. We described channel coding when we discussed FEC. They are the
same. The problem with channel coding or FEC. is that it corrects random
errors and not burst errors. The fading due to rainfall gives burst errors. By
using an interleaver, which pseudorandomizes the serial bit stream of data,
we can correct for error bursts. This assumes that the duration of the block
of data is longer than the duration of a fade at least statistically .. This
block of data has nothing to do with blocks we have been working with in
the past. This is a chunk of data more properly called a span. The
interleaver stores the data for the span period and then pseudorandomly
shuffles the data while it stores the next contiguous span. The reverse process
428
is carried out at the distant end. Of course the receive interleaver must be in
synchronism with the transmit interleaver. In such a way burst errors can be
made to look like random errors. With this type of operation we can
drastically reduce burst errors and pick up several decibels of coding gain
as well.
11.7.4.1 Definitions. The term inbound means circuits and traffic from a
VSAT terminal toward the hub. The converse of this is outbound, meaning
circuits or traffic from the hub toward one or more related VSAT terminals.
11.7.4.2 Optimizing Bandwidth Usage. Here we are dealing with multiple access protocols. In Chapters 9 and 10 we dealt with multiple access
protocols such as CSMArCD, contention and polling, and token passing.
Conventional satellite systems use frequency division multiple access
FDMA. or time division multiple access TDMA.. FDMA is where a
satellite transponder frequency slot is leased. These slots are typically less
than 1 MHz up to the full bandwidth of a transponder e.g., 36, 72 MHz..
Here we have the bandwidth and have to pay for it whether we use it or not.
TDMA is where we are assigned a time slot rather than a frequency slot. We
may have to pay for it whether we use it or not. Demand assignment multiple
access DAMA. is where we are assigned a frequency slot for only the time
we will use it. When we are finished, it is returned to the pool of free
channels, ready for someone else to use.
Almost universally, outbound traffic is carried on a TDM link. A time
division multiplex TDM. link is one where frames, packets, or blocks are
transmitted in one continuous serial bit stream and are received by all
VSATs or all members of a particular VSAT family. Framerblockrpacket
message headers are examined by each receiving terminal, and only those
frames with a terminals address are delivered to the end-user. That same
serial bit stream can also carry control information to control inbound traffic.
One of the most common TDM bit rates is 56 kbps. If system traffic
warrants, 128 and 384 kbps may be used. Some systems use 1.544 Mbps on
the outbound link. Remember, as the bit rate increases, the antenna aperture
will have to increase accordingly. If not, the hub may have to increase its
radio-frequency power output assuming that the satellite transponder will act
accordingly i.e., it will also be able to increase downlink power..
For uplink traffic, a VSAT system may have a calling frequency and then a
working frequency. When a VSAT terminal has traffic, it calls the hub and
asks for a channel to pass its traffic downstream to the hub.. Often the
calling channel is a common channel for all to use. To access the channel,
some form of contention is used, such as pure ALOHA similar to
CSMArCD, Chapter 10.. Collisions are more probable here and their
resolution more difficult because of the large delays involved.
Other approaches use a common channel for all inbound traffic and do
not resort to a calling channel and working channel s.. Certainly as the user
429
VSAT. population increases, the probability of collision increases. The

tradeoff whether we approach the problem using a form of DAMA i.e.,
calling and working channels . or contention scheme requires considerable
study if we are to optimize bandwidth required.
Section 11.7 is based on material from Ref. 20.
Figure 11.40. Hypothetical reference connections for public synchronous data networks. Link
A is the data link between two adjacent data switching exchanges in a national network; link
A1 is the data link between two adjacent gateway data switching exchanges in an international connection; link B is the data link between a source DSE and a destination DSE; link B1
is the data link between a local DSE and a gateway DSE; link G1 is the data link between a
source gate DSE and a destination gateway DSE in an international connection; link C is the
data link between source DTE and destination DTE; link D is the data link between source
DTE and the source local DSE or the data link between destination DTE and destination local
DSE; link E is the data link between communicating processes. (From Figure 1r X.92, page
296, CCITT Rec. X.92, Ref. 21.)
430
11.8 HYPOTHETICAL REFERENCE CONNECTIONS FOR PUBLIC

SYNCHRONOUS DATA NETWORKS
Hypothetical reference circuits or connections HRCs. are an aid to the
network designer to apportion impairments to the various circuit segments.
They are particularly useful on international connections and on long national connections involving several or many nodes. The HRCs presented
here have been taken from CCITT Rec. X.92 Ref. 21..
Figure 11.40 shows five hypothetical reference connections that are used
to assess overall customer-to-customer performance objectives. These reference connections may be used for circuit-switched services, packet-switched
services, and leased line services in public switched data networks PSDNs..
In the figure, between points Y and Z, transmission takes place over
64-kbps digital paths, the standard digital network voice channel called DS0
in the United States and Canada and called E0 in Europe. Such paths may
include sections using modems over analog facilities. It should be assumed
that the signaling for the circuit-switched data call control follows the same
Figure 11.41. Variants to logical link D. Link D1 is the data link between data terminal
equipment in user class of service 17 and PAD equipment; link D2 is the data link between
data terminal equipment in user class of service 811 or PAD equipment and a local data
switching exchange. Note 1: A user may see two different types of logical interfaces with the
network (links D1 and D2). Note 2: Link D2 could provide an interface for a single-access
terminal as well as for a multiple-access terminal. (From Figure 2r X.92, page 270, CCITT
Rec. X.92, Ref. 21.)
REFERENCES
431
route as the data connection. Logical links in the case of packet switching are
indicated in Figure 11.40 by dashed lines.
Figure 11.41 illustrates the variants to logical link D of Figure 11.40 to
allow incorporation of packet assemblyrdisassembly PAD. facilities.
REFERENCES
1. Dixon R. Doll, Data Communications: Facilities, Networks and Systems Design,
John Wiley & Sons, New York, 1978.
2. Interface Between Data Terminal Equipment DTE . and Data Circuit-Terminating
Equipment DCE . for Terminals Operating in the Packet Mode and Connected to
Public Data Networks by Dedicated Circuit, ITU-T Rec. X.25, ITU Geneva,
October 1996.
Equipment DCE . for Synchronous Operation on Public Data Networks, CCITT
Rec. X.21, ITU Geneva, September 1992.
4. Support of Packet Mode Terminal Equipment by an ISDN, ITU-T Rec. X.31, ITU
Geneva, March 1993.
5. Uyless Black, TCPrIP and Related Protocols, McGraw-Hill, New York, 1992.
6. Packet-Switched Signaling System Between Public Networks Proiding Data Transmission Serices, CCITT Rec. X.75, Fascicle VIII.3, CCITT Plenary Assembly,
Melbourne, 1988.
7. Internet Protocol, RFC 791, DDN Network Information Center, SRI International, Menlo Park, CA, September 1981.
8. A Standard for the Transmission of IP Datagrams Oer IEEE 802 Networks, RFC
1042, DDN Network Information Center, SRI International, Menlo Park, CA,
1988.
9. Mathew Naugle, Illustrated TCPrIP, John Wiley & Sons, New York, 1999.
10. Assigned Numbers, RFC 1060, DDN Network Information Center, SRI International, Menlo Park, CA, March 1990.
11. Internet Protocol Transition Workbook, SRI International, Menlo Park, CA, March
1982.
12. Reerse Address Resolution Protocol, RFC 903, DDN Network Information Center,
SRI International, Menlo Park, CA, June 1984.
13. Open Shortest Path First OSPF ., RFC 1131, DDN Network Information Center,
SRI International, Menlo Park, CA, 1991.
14. Internet Control Message Protocol, RFC 792, DDN Network Information Center,
SRI International, Menlo Park, CA, September 1981.
15. Transmission Control Protocol, RFC 793, DDN Network Information Center, SRI
International, Menlo Park, CA, September 1981.
16. Military Standard, Transmission Control Protocol, MIL-STD-1778, U.S. Dept. of
Defense, Washington DC, August 1983.
17. Internet Protocol Version IP6, RFC-2460, from the Internet Internet Society.,
December 1998.
432
18. User Datagram Protocol UDP ., RFC-768, DDN Network Information Center,
SRI International, Menlo Park, CA, September 1981.
19. Information Processing SystemsData CommunicationsProtocol for Proiding
Connectionless-Mode Network Serice, International Standards Organization, ISO8473, 1st edition, Geneva, December 1988.
20. Roger L. Freeman, Radio System Design for Telecommunications, 2nd edition,
John Wiley & Sons, New York, 1997.
21. Hypothetical Reference Connections for Public Synchronous Data Networks, CCITT
Rec. X.92, Fascicle VIII.3, IXth Plenary Assembly, Melboune, 1988.
12
FRAME RELAY
12.1
HOW CAN NETWORKS BE SPEEDED UP?
It would seem from the meaning of speeding up networks that we were

making bits travel faster down the pipe; that by some means we had broken
the speed barrier by dramatically increasing the velocity of propagation. Of
course this is eminently not true.
If somehow the bandwidth of the medium can be increased, the bit rate
can be increased as well. Another alternative is to turn to a multilevel
transmission scheme where we can squeeze more bits per hertz of bandwidth.
Another approach is to work our way around the voice channel; turn to the
digital voice channel, for example, which will provide 64 kbps. Still another
commonly used practice is to use several digital voice channels in a grouping.
For example, six of these will produce 384 kbps, the ISDN H0 channel.
Probably the greatest pressure to speed up the network came from LAN
users who wished to extend LAN connectivity to distant destinations. Ostensibly this traffic, as described in Chapters 9 and 10, displays local data rates
from 16 to 1000 Mbps. The initial resolution of this problem was to use X.25
wide area network connectivity Chapter 11.. It provided a robust service, but
was slow and tedious. There is a better way.
What slows down X25 service? X.25 is feasible at 64 kbps, and it may be
employed even at DS1rE1 rates. The intensive processing at every node see
Table 12.1., along with continual message exchange, reduces the progress of
packets from node to originator and from node to destination. On many X.25
connectivities, multiple nodes are involved, slowing down service still further.
One clue is that X.25 was designed for circuits with poor transmission
performance degrading error rates. It worked well on circuits with bit error
performance in the range of 1 = 10y4 . Meanwhile, the underlying digital
networks of North America display BERs better than 5 = 10y10 . This begs
the question of removing the responsibility of error recovery from the service
provider. If statistically an error occurs in 1 bit in 2,000,000,000 bits, there is
a strong argument for removing error recovery actions in their entirety.
433
434
FRAME RELAY
TABLE 12.1
Functional Comparison of X.25 and Frame Relay

Function
Flag recognition r generation

Transparency
FCS checking r generation
Recognize invalid frames
Discard incorrect frames
Address translation
Fill interframe time
Manage V (S) state variable
Manage V (R) state variable
Buffer packets awaiting acknowledgment
Manager timer T-1
Acknowledge received I-frames
Check received N (S) against V (R)
Generation of rejection message
Respond to poll r final bit
Keep track of number of retransmissions
Act upon reception of rejection message
Respond to receiver not ready (RNR)
Respond to receiver ready (RR)
Multplexing of logical channels
Management of D bit
Management of M bit
Management of W bit
Management of P (S) packets sent
Management of P (R) packets received
Detection of out-of-sequence packets
Management of network layer RR
Management of network layer RNR
X.25 in ISDN
(X.31)
Frame
Relay
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
In fact, with frame relay the following salient points emerge:

1. There is no process for error recovery by the frame relay service
provider.
2. The service provider does not guarantee delivery of frames, nor are
there any sort of acknowledgments provided.
3. Frame relay only uses the first two open system interconnection OSI.
layers i.e., physical and data-link layer., thus removing layer 3 and its
intensive processing requirements.
4. Frame overhead is kept to a minimum to reduce processing time and to
increase useful throughput.
5. There is no control field and no sequence numbering.
6. Under certain circumstances frames may be discarded using the leaky
bucket concept.
HOW CAN NETWORKS BE SPEEDED UP?
435
Figure 12.1. A typical frame relay network. FRS, frame relay switch.
7. Frames are discarded without notifying originator for such reasons as

congestion and having encountered an error.
8. Frame relay operates on a statmultiplex concept.
In sum, the service that the network provides can be speeded up by
increasing data rate, eliminating error recovery procedures, and reducing
processing time. One source states that a frame relay frame takes some 20 ms
to reach the distant end statistically ., where an X.25 packet takes in excess
of 200 ms on terrestrial circuits inside CONUS over similar distances.
Another advantage of frame relay over a conventional static TDM connection is that it uses virtual connections. Data traffic is often bursty and
normally would require much larger bandwidths to support the short data
messages, and much of the time that bandwidth would remain idle. Virtual
connections of frame relay only use the required bandwidth for the period of
the burst or usage. This is one reason frame relay is used so widely to
interconnect LANs over a WAN. Figure 12.1 shows a typical frame relay
network.
12.1.1
The Genesis of Frame Relay
Frame relay derives from the Integrated Services Digital Networks. ISDN.
link access procedure for the D-channel LAPD., or D-channel layer 2.
LAPD is discussed in Section 13.7. Its importance has taken on such a
436
FRAME RELAY
magnitude that the ITU-T Organization has formulated Rec. I.122, Framework for Proiding Additional Packet-Mode Bearer Serices Ref. 1., and Rec.
I.233, ISDN Frame Mode Bearer Serices Ref. 2.. Even the term LAPD,
although modified in certain cases, continues to be used for frame relay
applications.
Frame relay has become an ANSI initiative. There is also the Frame Relay
Forum, consisting of manufacturers and users of frame relay equipment,
which many feel is leading this imaginative initiative. So when we discuss
frame relay, we must consider what specifications a certain system is designed around:
ANSI, based on ANSI specifications and their publication dates

Frame Relay Forum with publication dates
ITU-T organization and its most current recommendations
There are also equivalent ANSI specifications directly derived from ITU-T
recommendations such as ANSI T1.617-1991 Ref. 3.. The term core aspects
of ISDN LAPD, or DL-CORE, refers to a reduced subset of LAPD found in
Annex A of ITU-T Rec. Q.922 Ref. 4.. The basic body of Q.922 presents
CCITTrITU-T specification for frame relay. This derivative is called LAPF
rather than LAPD. For all intents and purposes, the material found in ANSI
T1.618 Ref. 5. is identical to that found in Annex A of Q.922.
To properly describe frame relay from our perspective, we briefly give an
overview of the ANSI T1.618-1991 and T1.606-1990 Ref. 6.. This is followed
by some fairly well identified variants.
12.2
INTRODUCTION TO FRAME RELAY
Frame relay may be considered a cost-effective outgrowth of ISDN meeting

high-data-rate e.g., 2 Mbps. and low-delay data communications requirements. Frame relay encapsulates data files. These may be considered
packets, although they are called frames. Thus frame relay is often compared to CCITT Rec. X.25 packet service. Frame relay was designed for
current transmission capabilities of the network with its relatively wider
bandwidths* and excellent error performance e.g., BER better than 1 =
10y7 ..
The incisive reader will note the use of the term bandwidth. It is used
synonymously with bit rate. If we were to admit at first approximation 1 bit
per hertz of bandwidth, such use is acceptable. We are mapping frame relay
bits into bearer channel bits probably on a one-for-one basis. The bearer
channel may be a DS0rE0 64-kbps channel, a 56-kbps channel of a DS1
configuration, or multiple DS0rE0 channels in increments of 64 kbps up to
*I would rather use the term greater bit rate capacity.
437
1.544r2.048 Mbps. We may also map the frame relay bits into a SONET or
SDH configuration Chapter 15.. The final bearer channel may require more
or less bandwidth than that indicated by the bit rate. This is particularly true
for such bearer channels riding on radio systems, and to a lesser extent on a
fiber-optic medium or other transmission media. The reader should be aware
of certain carelessness of language used in industry publications.
Frame relay works well in the data rate range from 56 kbps up to
1.544r2.048 Mbps. It is being considered for the 45-Mbps DS3 rate for still
additional speed.
ITU-Ts use of the ISDN D channel for frame relay favors X.25-like
switched virtual circuits SVCs.. However, ANSI recognized that the principal application of frame relay was interconnection of LANs, and not to
replace X.25. Because of the high data rate of LANs megabit range.,
dedicated connections are favored. ANSI thus focused on permanent virtual
connections PVCs.. With PVCs, circuits are set up by the network,* not by
the endpoints. This notably simplified the signaling protocol. Also, ANSI
frame relay does not support voice or video.
As mentioned above, the ANSI frame relay derives from ISDN LAPD
core functions. The core functions of the LAPD protocol that are used in
frame relay as defined here. are as follows:
Frame delimiting, alignment, and transparency provided by the use of

HDLC flags and zero-bit insertionrextraction
Frame multiplexingrdemultiplexing using the address field
Inspection of the frame to ensure that it consists of an integer number
of octets prior to zero-bit insertion or following zero-bit extraction
Inspection of the frame to ensure it is not too long or too short
Detection of but not recovery from. transmission errors
Congestion control functions
In other words, ANSI has selected features from the LAPD structurerprotocol, rejected others, and added some new features. For instance, the control
field was removed, but certain control functions have been incorporated as
single bits in the address field. These are the CrR bit commandrresponse.,
DE bit discard eligibility., FECN bit, and BECN bit forward explicit
congestion notification and backward explicit congestion notification..
12.2.1
The Frame Structure
User traffic passed to a Frame Access Device FRAD. is segmented into

frames with a maximum length information field or with a default length of
*That is, set up by the switching nodes in the network.
LAPD is a derivative of HDLC see Chapter 3..
438
FRAME RELAY
Figure 12.2. Frame relay ANSI frame format with 2-octet addressing. Note: The default
address field length is 2 octets. By bilateral agreement, it may be extended to either 3 or 4
octets by using the address field extension bit ea. From ANSI T1.618-1991, Fig. 1, page 3,
(Ref. 5.)
262 octets and a recommended length ANSI. of at least 1600 octets when
the application is LAN interconnectivity. The minimum information field is
1 octet.
Figure 12.2 shows the overall frame relay frame structure. As mentioned
above, it uses HDLC opening and closing flags 01111110.. The closing flag
may also serve as the opening flag of the next contiguous frame; however,
receivers must be able to accommodate reception of one or more consecutive
flags on a bearer channel.
Address Field. The address field consists of at least 2 octets, but may be
extended to 3 or 4 octets as shown in Figure 12.3. There is no control field as
there is in HDLC, LAPD, or LAPB.
In its most reduced version, there are just 10 bits allocated to the address
field in 2 octets of bit positions. The remainder of the bits serve as control
functions. . The 10 bits can support up to 1024 logical connections i.e.,
2 10 s 1024..
It should be noted that the number of addressable logical connections is
multiplied because they can be reused at each nodal switch. interface. That
439
Figure 12.3. Address field formats. Source: ANSI T1.619-1991, Fig. 6, page 6, (Ref.5).
is, an address in the form of a data-link connection identifier DLCI. has

meaning only on one trunk between adjacent nodes. The switch node. that
receives a frame is free to change the DLCI before sending the frame onward
over the next link. Thus, the limit of 1024 DLCIs applies to the link, not the
network.
Information Field. Information field follows the address field and precedes
the frame check sequence. The maximum size of the information field is an
implementation parameter, and the default maximum is 262 octets. ANSI
chose this default maximum to be compatible with LAPD on the ISDN D
channel, which has a 2-octet control field and a 260-octet maximum information field. All other maximum values are negotiated between users and
networks and between networks. The minimum information field size is 1
octet. The field must contain an integer number of octets; partial octets are
not allowed. A maximum of 1600 octets is encouraged for applications such
as LAN interconnects to minimize the need for segmentation and reassembly
by user equipment.
A number of frame relay service providers have standardized their networks at considerably longer information fields, between 4000 and 5000
octets.
440
FRAME RELAY
Transparency. As with HDLC, X.25, and LAPD, the transmitting data-link

layer must examine the frame content between opening and closing flags and
inserts a 0 bit after all sequences of five contiguous 1s including the last 5
bits of the FCS. to ensure that a flag or an abort sequence is not simulated
within the frame. At the other side of the link, a receiving data-link layer
must examine the frame contents between the opening and closing flags and
must discard any 0 bit that directly follows five contiguous 1s.
Frame Check Sequence. Frame check sequence FCS. is based on the
generator polynomial X 16 q X 12 q X 5 q 1. The CRC processing includes
the content of the frame existing between, but not including, the final bit of
the opening flag and the first bit of the FCS, excluding the bits inserted for
transparency. The FCS, of course, is a 16-bit sequence. If there are no
transmission errors detected ., the FCS at the receiver will have the sequence
00011101 00001111.
Order of Bit Transmission. The order of bit transmission is that the octets
are transmitted in ascending numerical order and that inside an octet bit, 1 is
the first bit transmitted.
Field Mapping Convention. When a field is contained within a single octet,
the lowest bit number of the field represents the lowest-order value. When a
field spans more than one octet, the order of bit values progressively
decreases as the octet number increases within each octet. The lowest bit
number associated with the field represents the lowest-order value.
An exception to the preceding field mapping convention is the data-link
layer FCS field, which spans 2 octets. In this particular case, bit 1 of
the first octet is the high-order bit and bit 8 of the second octet is the loworder bit.
Invalid Frames. An invalid frame is a frame that:
is not properly bounded by two flags e.g., a frame abort.;

has fewer than 3 octets between the address field and the closing flag;
does not consist of an integral number of octets prior to zero bit
insertion or following zero bit extraction;
contains an FCS error;
contains a single-octet address field; and
contains a data-link connection identifier DLCI. that is not supported
by the receiver.
Invalid frames are discarded without notification to the sender, with no

further action.
441
Frame Abort. Frame abort consists of seven or more contiguous 1 bits; upon
receipt of an ABORT, the data-link layer ignores the frame currently being
received.
12.2.2
Address Field Discussion
Figure 12.3 shows the ANSI-defined address field formats. Included in the
field are the address field extension bits; a reserved bit to support a
commandrresponse CrR. indication bit; forward and backward explicit
congestion indicator FECN and BECN. bits; discard eligibility indicator
DE.; a data-link connection identification DLCI. field; and, finally, a bit to
indicate whether the final octet of a 3- or 4-octet address field is the
low-order part of the DLCI or DL-CORE control information. The minimum
and default length of the address field is 2 octets. However, the address field
length may be extended to 3 or 4 octets. To support a larger DLCI address
range, the 3- or 4-octet address fields may be supported at the usernetwork
interface or networknetwork interface based on bilateral agreement.
12.2.2.1 Address Field Variables
Address Field Extension Bit (EA). The address field range is extended by
reserving the first transmitted bit of the address field octets to indicate the
final octet of the address field. If there is a 0 in this bit position, it indicates
that another octet of the address field follows this one. If there is a 1 in the
first bit position, it indicates that this octet is the final octet of the address
field. As an example, for a 2-octet address field, bit 1 of the first octet is set
to 0 and bit one of the second octet is set to 1.
It should be understood that a 2-octet address field is specified by ANSI.
It is a users option whether a 3- or 4-octet field is desired.
CommandrResponse Bit (CrR). The CrR bit is not used by the DL-CORE
protocol, and the bit is conveyed transparently.
Forward Explicit Congestion Notification (FECN) Bit. This bit may be set by
a congested network to notify the user that congestion avoidance procedures
should be initiated, where applicable, for traffic in the direction of the frame
carrying the FECN indication. This bit is set to 1 to indicate to the receiving
end-system that the frames it receives have encountered congested resources.
The bit may be used to adjust the rate of destination-controlled transmitters.
While setting this bit by the network or user is optional, no network shall
ever clear this bit i.e., set to 0.. Networks that do not provide FECN shall
pass this bit unchanged.
442
FRAME RELAY
Backward Explicit Congestion Notification (BECN) Bit. This bit may be set
by a congested network to notify the user that congestion avoidance procedures should be initiated, where applicable, for traffic in the opposite
direction of the frame carrying the BECN indicator. This bit is set to 1 to
indicate to the receiving end-system that the frames it transmits may encounter congested resources. The bit may be used to adjust the rate of
source-controlled transmitters.
While setting this bit by the network or user is optional according to the
ANSI specification, no network shall ever clear i.e., set to 0. this bit.
Networks that do not provide BECN shall pass this bit unchanged.
Discard Eligibility Indicator (DE) Bit. This bit, if used, is set to 1 to indicate a
request that a frame should be discarded in preference to other frames in a
congestion situation. Setting this bit by the network or user is optional. No
network shall ever clear i.e., set to 0. this bit. Networks that do not provide
DE capability shall pass this bit unchanged. Networks are not constrained to
only discard frames with DE equal to 1 in the presence of congestion.
Data-Link Connection Identifier (DLCI). This is used to identify the logical
connection, multiplexed within the physical channel, with which a frame is
associated. All frames carried within a particular physical channel and having
the same DLCI value are associated with the same logical connection.
The DLCI is an unstructured field. For 2-octet addresses, bit 5 of the
second octet is the least significant bit. For 3- and 4-octet addresses, bit 3 of
the last octet is the least significant bit. In all cases, bit 8 of the first octet is
the most significant bit.
The structure of the DLCI field may be established by the network at the
usernetwork interface subject to bilateral agreements.
In order to allow for compatibility of call control and layer management
between BrH and D channels, the following ranges of DLCIs are reserved
and preassigned see Table 12.2.. The DLCIs have local significance only.
DLCI on the D Channel. The six most significant bits bits 8 to 3 of the first
octet. correspond to the service access point identifier SAPI. field in ANSI
standard . T1.602 Ref. 7..
The DLCI subfield bits 8 to 3 of the first octet. values that apply on a D
channel are reserved for specific functions to ensure compatibility with
operation of the D channel that may also use ANSI T1.602 protocols.
A 2-octet address format for the DLCI is assumed when used on the D
channel. Note: For frame relay in the D channel, only DLCI values in the
range 512991 SAPI s 3161. will be assigned.
Table 12.3 gives DLCI values for D channel.
DLCI or DL-CORE Control Indicator (DrC). The DrC indicates whether the
remaining six usable bits of that octet are to be interpreted as the lower
DLCI bits or as DL-CORE control bits. This bit is set to 0 to indicate that
TABLE 12.2
443
DLCI Values, Use of DLCIs
DLCI Range
Function
10 bits DLCIs (Note 1)
0 (Note 2)
115
16511
512991
9921007
10081022
1023 (Note 2)
In channel signaling, if required

Reserved
Network option: on non-D channels, available for support of user
information
Logical link identification for support of user information (Note 6)
Layer 2 management of frame mode bearer service
Reserved
In channel layer 2 management, if required
0 (Note 2)
11023
102432,767
32,76863,487
63,48864,511
64,51265,534
65,535 (Note 2)

Reserved
information
Reserved
0 (Note 2)
12047
204865,535
65,536126,975
126,976129,023
129,024131,070
131,071 (Note 2)

Reserved
information
Reserved
0 (Note 2)
1131,071
131,0724,194,303

Reserved
information
Note 1. These DLCIs apply when a 2-octet address field is used or when a 3-octet address field is
used with DrC s 1.
Note 2. Only available within non-D channel.
Note 3. These DLCIs apply for non-D channels when a 3-octet address field is used with DrC s 0.
Note 6. The use of semipermanent frame mode connections may reduce the number of DLCIs
available from this range.
Source: Table 1rQ.922, page 7, CCITT Rec. Q.922, ITU Geneva, 1992 (Ref. 4 ).
444
FRAME RELAY
TABLE 12.3 DLCI Values for D Channel

(2-Octet Address Format)
DLCI Values
Function
512991
Assigned using frame relay

connection procedures
Source: Ref. 5.
the octet contains DLCI information. When this bit is set to 1, it indicates
that the octet contains DL-CORE control information. The DrC is limited
to use in the last octet of the 3- or 4-octet-type address field. The use of this
indication for DL-CORE control is reserved as there have not been any
additional control functions defined that need to be carried in the address
field. Thus this indicator has been added to provide possible future expansion of the protocol.
12.3
DL-CORE PARAMETERS (AS DEFINED BY ANSI)
DLCI VALUE parameter conveys the DLCI agreed to be used between core
entities in support of DL-CORE connection. Its syntax and usage are
described above.
DL-CORE connection endpoint identifier CEI . uniquely defines the DLCORE connection.
Physical connection endpoint identifier ph-CEI . uniquely identifies a physical connection to be used in support of a DL-CORE connection.
12.3.1
Procedures
For permanent frame relay bearer connections, information related to the

operation of the DL-CORE protocol in support of DL-CORE connection is
maintained by DL-CORE management. For demand frame relay bearer
connections, layer 3 establishes and releases DL-CORE connections on
behalf of the DL-CORE sublayer. Therefore, information related to the
operation of the DL-CORE protocol is maintained by coordination of layer 3
management and DL-CORE sublayer management through the operation of
the local system environment.
Connection Establishment. When it is necessary to notify the DL-CORE
sublayer entity either because of establishment of a demand frame relay call,
because of notification of reestablishment of a permanent frame relay bearer
connection, or because of system initialization . that a DL-CORE connection
is to be established, the DL-CORE layer management entity signals an
MC-ASSIGN request primitive to the DL-CORE sublayer entity.
The DL-CORE sublayer entity establishes the necessary mapping between
supporting ph-connection, the core-CEI, and the DLCI. In addition, if it has
TRAFFIC AND BILLING ON FRAME RELAY
445
not already done so, it begins to transmit flags on the physical connection
except on the D channel.
Connection Release. When it is necessary to notify the DL-CORE sublayer
entity either because of release of a demand frame relay call or because of
notification of failure of a permanent frame relay bearer connection. that a
DL-CORE connection is to be released, the DL-CORE layer management
entity signals the MC-REMOVE request primitive to the DL-CORE sublayer
entity.
12.4
TRAFFIC AND BILLING ON FRAME RELAY
Figure 12.4 shows a typical traffic profile on a conventional public telephone

network, whereas Figure 12.5 shows a typical profile of bursty traffic over a
frame relay network. Such bursty traffic is typical for a LAN. The primary
employment of frame relay is to interconnect LANs at a distance.
Turning to Figure 12.5, the traffic is bursty. With conventional leased data
circuits, we have to pay for the bandwidth* whether it is used or not. On the
other hand, with frame relay, we only pay for the time used. Billing can be
handled in one of three ways:
1. CIR Committed Information Rate. is a data rate subscribed to by a
user. This rate may be exceeded for short bursts during the peak period
as shown in Figure 12.5.
2. We can pay a flat rate.
3. We can pay per packet i.e., frame..
Figure 12.4. Typical traffic profile of a public switched telephone network. (From Ref. 8.
Courtesy of Hewlett-Packard.)
*I prefer the use of the expression bit rate capacity.
446
FRAME RELAY
Figure 12.5. Typical bursty traffic of frame relay. Note the traffic levels indicated. (From Ref. 8.
Courtesy of Hewlett-Packard.)
Note that on the right-hand side of Figure 12.5 there is the guaranteed
transmission bit rate equivalent to the CIR. Depending on the traffic load
and congestion, during short periods a user may exceed the CIR. However,
there is a point where the network cannot sustain further increases in traffic
without severe congestion resulting. Traffic above such levels is arbitrarily
discarded by the network without informing the originator.
12.5
PVCs AND SVCs
Most frame relay service providers offer permanent virtual circuits PVCs. to
transport the frame relay traffic. Many frame relay standards describe PVCs
as dedicated circuits whereas they often really mean proisioned circuits.
Provisioned has a slightly different connotation than dedicated. Provisioned means that a particular frame relay port accesses a preselected route,
the end points being known. Once the circuit becomes live wi.e., the
originating FRAD frame relay access device. has frames to transmitx, the
contracted bit rate capacity is allocated on that route. With many service
providers, there is an alternative backup route assigned in addition to the
primary route. PVCs require no circuit setup signaling procedures, only
verification that the route is alive and healthy.
An SVC is a switched virtual circuit. In this case the circuit is not
provisioned but set up by signaling procedures when there are frames to be
delivered to a particular location. wRefer to ITU-T Rec. X.76, Amendment 1
08r97. Ref. 16.x.
12.6
TWO TYPES OF INTERFACES: UNI AND NNI
A usernetwork interface UNI. is that interface between a user FRAD and

an FR switch. The interface between FR switches is called an NNI or
CONGESTION CONTROL: A DISCUSSION
447
Figure 12.6. Illustration of interfaces: UNI and NNI. (From Figure 2rI.372, page 2, ITU-T Rec.
I.372, Ref. 15).
networknetwork interface.* The UNI is described in ANSI T.618-1991 for

OSI layer 2 T.617., in ITU-T Rec. Q.933 for layer 3 SVC, in ITU-T Rec.
Q.922 for layer 2, in X.36 for layer 2, and in X.36 Amd 1 for layers 2 and 3
for PVCs. The NNI is covered in ITU-T Rec. X.76 SVC. and X.76 Amd 1.
See Refs. 27, 10, and 14.
Specifically for PVCs, layer 3 circuit maintenance messages are employed,
which are described in ANSI T.617A Annex D and ITU-T Q.933 Annex A
Ref. 11..
Figure 12.6 distinguishes UNIs and NNIs in frame relay networks. In the
figure, an NNI or UNI can be either a PVC or SVC.
It is emphasized that all frame relay revenue bearing traffic is transported
through layer 2 frames. A typical layer 2 frame is illustrated in Figure 12.3.
12.7
Congestion in the user plane occurs when traffic arriving at a resource

exceeds the networks capacity. It can also occur for other reasons such as
equipment failure. Network congestion affects the throughput, delay, and
frame loss experienced by the end user.
End-users should reduce their offered load in the face of network congestion. Reduction of offered load by an end-user may well result in an increase
in the effective throughput available to the end-user during congestion.
Congestion avoidance procedures, including optional explicit congestion
notification, are used at the onset of congestion to minimize its negative
effects on the network and its users.
Explicit notification is a procedure used for congestion avoidance and is
part of the data transfer phase. Users should react to explicit congestion
*An NNI is sometimes called a networknode interface.
448
FRAME RELAY
notification i.e., optional but highly desirable .. Users that are not able to act
on explicit congestion notification should have the capability to receive and
ignore explicit notification generated by the networks.
Congestion recovery and the associated implicit congestion indication due
to frame discard are used to prevent network collapse in the face of severe
congestion. Implicit congestion detection involves certain events available to
the protocols operating above the core function to detect frame loss e.g.,
receipt of a REJECT frame, timer recovery.. Upon detection of congestion,
the user reduces the offered load to the network. Use of such reduction by
users is optional.
12.7.1
Network Response to Congestion
Explicit congestion signals are sent in both forward toward frame destination. and backward toward frame source. directions. Forward explicit congestion notification is provided by using the FECN bit in the address field.
Backward explicit congestion notification is provided by one of two methods.
When timely reverse traffic is available, the BECN bit in an appropriate
address field may be used. Otherwise, a single consolidated link layer
management message may be generated by the network. The consolidated
link layer management CLLM. message travels on the U-plane physical
path. The generation and transport of CLLM by the network are optional.
All networks shall transport the FECN and BECN bits without resetting.
12.7.2
User Response to Congestion
Reaction by the end-user to the receipt of explicit congestion notification is

rate-based. Annex A to ANSI T1.618-1991 Ref. 5. gives an example of user
reaction to FECN and BECN.
End-User Equipment Employing Destination-Controlled Transmitters. Enduser reaction to implicit congestion detection or explicit congestion notification FECN indications ., when supported, is based on the values of FECN
indications that are received over a period of time. The method is consistent
with commonly used destination-controlled protocol suites, such as OSI class
4 transport protocol operated over the OSI connectionless service.
End-User Equipment Employing Source-Controlled Transmitters. End-user
reaction to implicit congestion notification BECN indication ., when supported, shall be immediate when a BECN indication or a CLLM is received.
This method is consistent with implementation as a function of data-link-layer
elements of procedure commonly used in source-controlled protocols such as
CCITT Rec. Q.922 elements of procedure.
The following figure illustrates frame relay congestion control:
12.7.3
449
Consolidated Link Layer Management (CLLM) Messages
The CLLM uses XID frames for the transport of functional information. We
may remember in the discussion of HDLC Section 3.5. that an XID frame
was an exchange identification frame. It was used for reporting station identification. In frame relay, CLLM uses XID frames for network management as
an alternative for congestion control. CLLM messages originate at network
nodes to the frame relay interface usually housed in a router near or
incorporated with the user equipment.
As we mentioned, BECNrFECN bits in frames must pass congested nodes
in the forward or backward direction. Suppose for a given user no frames
pass in either direction; thus that user has no knowledge of network congestion because at that moment the user is not transmitting or receiving frames.
Frame relay standards do not permit a network to generate frames with the
DLCI of the congested circuit. CLLM covers this contingency. It has a
DLCI s 1023 reserved.
The use of CLLM is optional. If it is used, it may or may not operate in
conjunction with BECNrFECN. The CLLM frame format dedicates 1 octet
for the cause of congestion such as excessive traffic, equipment or facility
failure, preemption, or maintenance action.
The CLLM message is based on the ISO 8885 definition of XID frames
for transport of function information. The frame format of a CLLM message
is illustrated in Figure 12.7. Each parameter is described in the CLLM using
sequence-type-length value. The functional fields for CLLM are described
450
FRAME RELAY
Figure 12.7. Frame format of a CLLM message, with 2-octet address field. IA5, International
Alphabet No. 5 (Ref.12) (see Section 5.3). (From Figure C.1rX.36, page 33, ITU-T Rec. X.36,
Ref. 10.)
below with reference to the figure. The CLLM message may be transmitted
whenever a congestion control procedure is being performed as a result of
network congestion, line or equipment failure, or the performance of maintenance functions. All fields are binary coded unless otherwise specified.
Field Descriptions, CLLM Message Format, Figure 12.7
Fields 1 and 2. Addressing, self-explanatory.
Field 3, XID Control Field. For a CLLM frame, coded 10101111 in binary.
Field 4, Format Identifier. Coded 130 decimal and 10000010 in binary.
Field 5, Group Identifier Field. Coded decimal 15 and 00001111 in binary.
Fields 6 and 7, Group Length Octets. This 16-bit field describes the length,
in octets, in the remainder of the group field. The maximum value of
the group field is 256.
Field 8, Group Field Value Field. Consists of two or more parameter fields.
The parameter set identification with a parameter value of 0. identifies
a set of private parameters within the group field in accordance with
TABLE 12.4
451
Codes for Cause in CLLM Message
Bits
87654321
00000010
00000011
00000110
00000111
00001010
00001011
00010000
00010001
Cause
Network congestion due to excessive trafficshort term
Network congestion due to excessive trafficlong term
Facility or equipment failureshort term
Facility or equipment failurelong term
Maintenance actionshort term
Maintenance actionlong term
Unknownshort term
Unknownlong term
All other values are reserved
Source: Table C.1rX.36, page 35, ITU-T Rec. X.36, ITU Geneva 4r95 (Ref. 10 ).
ISO 8885rDAD3 as an identifier to be determined. Other parameters

appear in the following order: case identifier and DLCI identifier.
Parameter Set Identification Field is always present; otherwise the
CLLM message is ignored. There are three subfields under this heading. Octet 8 contains the parameter identification field for the first
parameter and is set to 0 in accordance with ISO 8885rDAD3. Parameter 0 identifies the set of private parameters within this group.
Field 9, Parameter Set Identification Length Field. Contains the length
parameter 0 and is set to binary value 100 a decimal value of 4..
Fields 1013, Parameter Field Value. Identify the usage of the XID frame
private parameter group for ITU-T Rec.. I.122 private parameters.
Octet 10 contains IA5 International Alphabet No. 5. value I decimal
105.. Octet 11 contains IA5 values 1 decimal 49.; and octets 12 and 13
each contains the IA5 value of 2 decimal 50..
Fields 1416, Cause Identifier Parameter Field. See Table 12.4.. These
fields are explained in the following text.
Field 14, Parameter Identifier Field. Is set to decimal 2 binary 00000010.
Field 15, Parameter Length Field. It gives the length of the cause identifier and is set to binary 1 i.e., 00000001.
Field 16, Cause Value. This octet gives the cause as determined by the
congested network node whose layer management module originated
the message. The value for the cause is to indicate the network status of
the layer management entity issuing the message e.g., congestion,
failure, or maintenance operation.. The binary values for cause are
shown in Table 12.4.
Field 17 and Onwards, DLCI Identifier Parameter Field. This field is used to
determine the DLCI corresponding to the cause listed in the CLLM
described above. If the DLCI identifier is missing, then the frame is
452
FRAME RELAY
ignored. Octet 17 contains the parameter identifier field. When this

field is set to decimal 3 binary 00000011., the octets that follow this
parameter contain the DLCIs. of the frame relay connections. that
are congested.
Field 18, Parameter Length. This contains the length of the DLCIs. being
reported, in octets. For example, if n. DLCIs are being reported and
they are of length 2 octets each, this will be 2 times n. in octet size.
Field 19 Onward to FCS Field. This is the Parameter Value field. Here is a
listing of the DLCI values. which identify the logical links. that have
encountered a congested state. The DLCI field in a frame relay frame.
is 10 bits long and is contained in bits 8 to 3 of the first octet pair and
bits 8 to 5 of the next octet pair. Bit 8 of the first octet is the most
significant bit and bit 5 of the second octet is the least significant bit.
Bits 2 to 1 in the first octet and 4 to 1 in the second octet are reserved.
Network Congestion, Use of the CLLM. When the network encounters

resource congestion as a result of excessive DTE traffic levels and the level of
traffic continues at these high levels, the network may be forced to discard
traffic or bring the system to a halt for recovery. By transmitting a CLLM
message to the DTE indicating the underlying cause of congestion, the
network informs the DTE of the possibility of an action being taken. It
should be noted that CLLM messages generated at times of congestion are
intended only for providing notification in the opposite direction to the
traffic congestion direction. CLLM messages may signal a congestion situation to the transmitting DTE when there is no traffic in the reverse direction.
12.7.4
Action of a Congested Node
When a node is congested, it has several alternatives it may use to mitigate or

eliminate the problem. It may set the FECN and BECN bits to 1 in the
address field andror use the CLLM message. Of course, the purpose of
explicit congestion notification is to either a. inform the edge node at the
network ingress of congestion so that the edge node can take the appropriate
action to reduce the congestion, b. notify the source that the negotiated
throughput has been exceeded, or c. do both.
One of the strengths of the CLLM is that it contains a list of DLCIs that
correspond to the congested frame relay bearer connections. These DLCIs
indicate not only the sources currently active causing the congestion but also
those sources that are not active. The reason for the latter is to prevent those
sources that are not active from becoming active and thus causing still further
congestion. It may be necessary to send more than one CLLM message if all
the affected DLCIs cannot fit into a single frame.
FLOW CONTROL AND POSSIBLE APPLICATIONS OF FECN AND BECN BITS
453
12.8 FLOW CONTROL AND POSSIBLE APPLICATIONS

OF FECN AND BECN BITS
ANSI in T1.618-1991 Annex A Ref. 5. presents some sample procedures and
suggestions for the application of FECN and BECN information and user
reaction to such congestion notifications.
12.8.1
FECN Usage
User Behavior on Receipt FECN Bit. ANSI recommends Ref. 5. that users
compare the number of frames in which the FECN bit is set with the number
of frames in which the FECN bit is cleared over a measurement interval %.
If the number of set FECN bits a set bit in this context is a bit set to 1
indicating some form of congestion. is equal to or exceeds the number of
clear FECN bits bits set to 0. received during this period, then the user
should reduce its throughput to 78 0.875. of its previous value. If the number
of set FECN bits is less than the number of clear FECN bits, then the user
may increase the information rate by 161 of its throughput.
The suggested measurement interval % is approximately equal to four
times the end-to-end transit delay. However, other mechanisms not dependent on timers may be used by the terminal if the effect is similar.
A slow-start mechanism is recommended, so as to cause convergence
toward equilibrium on the connection. ANSI suggests that the initial data
rate should be set to the throughput value agreed at the time of call
establishment, or less, in order to avoid an impulse load on the network at
the time the user begins transmitting. If the connection has been idle for a
long period of time in the order of tens of seconds., the offered rate or less
should be returned to the throughput.
Use of Windows as an Approximation to Rate-Based Control. For some
implementations, it may be convenient to use a window-based mechanism as
an approximation to rate-based control. Such implementations may or may
not be able to measure their offered rate, or to relate to the throughput
negotiated with the network during connection establishment. The actual
offered rate is limited by end-to-end transit delay, access rate, window size,
and frame size.
In the event that a windowed protocol is in use, then the user, in
responding to the FECN bit, compares the number of frames received with
the FECN bit set with the number of frames with the FECN bit clear. The
measurement interval for this shall be twice the interval during which the
number of frames equal to the current working window size is transmitted
and acknowledged i.e., two window turns.. If the number of set FECN bits is
equal to or exceeds the number of clear FECN bits received during this
454
FRAME RELAY
period, then the user will reduce its working window size variable to 78 0.875.
of its previous value. However, it need never reduce its working window
below the size of one frame. If the number of FECN bits is less than the
number of clear FECN bits, then the user increases its window size variable
by the size of one frame, not to exceed the maximum window size for that
virtual circuit. After the adjustment is made, the set and clear FECN bit
counters are reset to zero and the comparison begins again.
The working window is initialized at a small value, such as one frame, in
order to avoid impulse load upon the network at the time a user begins
transmitting. If the connection has been idle for a long period of time, in the
order of tens of seconds, it may be appropriate to again reduce the window
size to its initial value. There may be a maximum window size or information
rate that a connection can accommodate, as limited by the end-system
policies; the working window should not be adjusted beyond such a value. It
should also be observed that this algorithm is relatively insensitive to the loss
of acknowledgments carried in the user data when they are used to carry
window adjustment information back to the source.
12.8.2
BECN Usage
BECN Congestion Control on a Step Count S. S, as defined in the ANSI

specification Annex A, is a function of twice the end-to-end transit time, with
the information rate in each direction multiplied by the ratio of the forward
rate to the reverse data rate.
If a frame with a BECN bit is received and the users offered rate is
greater than the throughput, the user reduces its offered rate to the throughput rate agreed upon for the frame relay connection.
If S consecutive frames are received with the BECN bit set, the user
reduces its rate to the next step rate below the current offered rate.
Further rate reduction should not occur until S consecutive frames are
received with the BECN bit set. The step rates are:
0.675 times throughput
Note: Networks should be engineered such that reduction below 0.5 times
throughput never becomes necessary, and they may take alternative action
e.g., rerouting. to prevent such occurrence.
When the user has reduced its rate due to receipt of BECN, it may
increase its rate by a factor of 0.125 times throughput after any Sr2
consecutive frames are received with the BECN bit clear.
POLICING A FRAME RELAY NETWORK
455
Network Procedures for Setting the BECN Bit. The network should, if
possible, set the BECN bit before it becomes necessary to discard frames.
The network should continue to set the BECN bit whenever it is in such
condition, and it may elect to provide some hysteresis to prevent oscillation.
If the congestion condition deteriorates, the network should discard frames
sent in excess of throughput at the access node, as well as discard those
marked with the DE bit, in preference to other frames. At this stage, the
network is in a moderately congested state, and BECN bits should continue
to be set in frames that are not discarded.
If the congestion condition further deteriorates to the point that frames
that are neither in excess of throughput nor are marked with the DE bit are
being discarded, a severe congestion condition exists. The network should
continue to use the BECN to encourage users to reduce their transmission
rates andror take further action such as clearing or rerouting calls to restore
control.
12.9
12.9.1

Introduction
Frame relay switches may carry out a policing function on accessing users.
The result of such action is discarded frames. Similar policing actions are
employed on ATM circuits Chapter 16..
12.9.2
Definitions
Access Rate (AR). The data rate, expressed in bits per second, of the user
access channel D, B, or H channel .. The rate at which users can offer
data to the network is bounded by the access rate.
Excess Burst Size ( B e ). The maximum amount of uncommitted data in bits.
that the network will attempt to deliver over the measurement interval
T .. These data may or may not be contiguous i.e., may appear in one
frame or in several frames, possibly with interframe idle flags.. Be is
negotiated at call establishment for demand establishment of communication. or at service subscription time for permanent establishment of
communication.. Excess burst data may be marked for discard eligibility
with the DE bit. by the network.
Measurement Interval ( T ). The time interval over which rates and burst
sizes are measured. In general, the duration of T is proportional to the
burstiness of the traffic. Except as noted below, T is computed as T s
BcrCIR or T s BerAR.
Committed Information Rate (CIR). The rate, expressed in bits per second,
at which the network agrees to transfer information under normal conditions. This rate is measured over the measurement interval T. CIR is
456
FRAME RELAY
negotiated at call establishment or service subscription time. Data marked

discard eligible DE. is not accounted for in CIR.
Committed Burst Size ( B c ). The maximum amount of data in bits. that a
network agrees to transfer under normal conditions over a measurement
interval T. These data may or may not be contiguous i.e., they may
appear in one frame or in several frames, possibly with interframe idle
flags.. Bc is negotiated at call establishment for demand establishment of
communication. or service subscription time for permanent establishment
of communication..
Fairness. An attempt by the network to maintain the negotiated quality of
service for all users operating under normal conditions i.e., within their
CIR and Bc , without discard due to congestion.. For example, the network
may discard frames offered in excess of CIR, and it may reject any call
attempts that would cause the network resources to be overcommitted.
Offered Load. The bits offered to the network by an end user, to be
delivered to the selected destination. The information rate and burst
length offered to the network could exceed the negotiated class of service
parameters. The offered load consists of the user data portion of frames
and therefore excludes flags, FCS, and inserted zeros.
12.9.3
Relationship Among Parameters
Figure 12.8 illustrates the relationship of access rate, excess burst, committed
burst, committed information rate, discard eligibility indicator, and measurement interval parameters. The CIR, Bc , and Be parameters are negotiated at
call establishment time for demand establishment of communication or
established by subscription for permanent establishment of communication.
Access rate is established by subscription for permanent access connections
or during demand access connection establishment. Each end-user and the
network participate in the negotiation of these parameters to agreed-upon
values.
These negotiated values are then used to determine the measurement
interval parameter, T, and when the discard eligibility indicator if used. is
set. These parameters are also used to determine the maximum allowable
end-user input levels. The relationship among parameters can be used at any
instant of time, T0 , to measure the offered load over the interval T0 , Tc q T ..
Similarly, the offered load over any interval t y T, t . may be measured at
any instant of time t, as long as the measurement function retains memory of
user activity over the previous interval T. One way of doing this is by use of a
leaky bucket algorithm described below.
The measurement interval is determined as shown in Table 12.5. The
network and the end-users may control the operation of the discard eligibility
indicator DE. and the rate enforcer functions by adjusting at call setup the
CIR, Bc and Be parameters in relation to the access rate. If both the CIR
457
Figure 12.8. Illustration of relationships among parameters. Note 1: Number of frames and
size of frames are for illustrative purposes only. Note 2: Frame may be discarded at ingress
node. This is a region of rate enforcement. Note 3: Frame marked as discard eligible. (From
Figure 7, page 9, ANSI T1.606a, Ref. 9.)
TABLE 12.5
Congestion Parameter State
CIR
Bc
Be
)0
)0
s0
)0
)0
s0
)0
s0
)0
Measurement
Interval ( T )
T s Bc rCIR
T s Bc rCIR
T is a networkdependent value
Source: Reference 9, paragraph 10.2.
and Bc parameters are not equal to zero, then T s BcrCIR. In addition

there are two special conditions:
1. When CIR s access rate, Bc s 0, and Be s 0, both access rates must
be equal i.e., ingress s egress..
2. When CIR s 0 Bc must s 0. and Be ) 0, then T is a network-dependent value.
458
FRAME RELAY
Notes
1. The ingress and egress access rates do not have to be equal; however,
when the ingress access rate is substantially higher than the egress
access rate, continuous input of Be frames at the ingress interface may
lead to persistent congestion of the network buffers at the egress
interface, and a substantial amount of the input Be data will be
discarded.
2. When a frame, entering the ingress node, consumes the remaining
capacity of Be or Bc , reducing it to zero, the action taken on that frame
is network-dependent.
Figure 12.8 is a static illustration of the relationship among time, cumulative
bits of user data, and rate. In this example, the user sends four frames during
the measurement interval T0 q T .. The slope of the line marked CIR is
BcrT. Bits are received at the access rate by the ingress node. of the access
channel. Because the sum of the number of bits contained in frames 1 and 2
is not greater than Bc , the network does not mark these frames with the
discard eligibility indicator DE.. The sum of the number of bits in frames 1,
2, and 3 is greater than Bc , but not greater than Bc q Be ; therefore frame 3
is marked discard eligible. Because the sum of the number of bits received by
the network in frames 1, 2, 3, and 4 exceeds Bc q Be , frame 4 is discarded at
the ingress node. This figure does not address the case in which the end-user
sets the DE bit.
This section is based on Ref. 9, paragraph 10.2.
12.10
QUALITY OF SERVICE PARAMETERS
The quality that frame-relaying service provides is characterized by the values

of the following parameters. ANSI adds in Ref. 6 that the specific list of
negotiable parameters is for further study.
Throughput
Transit delay
Information integrity
Residual error rate
Delivered errored. frames
Delivered duplicated frames
Delivered out-of-sequence frames
Lost frames
Misdelivered frames
Switched virtual call established delay
NETWORK RESPONSIBILITIES
459
Switched virtual call clearing delay

Switched virtual call establishment failure
Premature disconnect
Switched virtual call clearing failure
12.11
NETWORK RESPONSIBILITIES
The frame are routed through the network on the basis of an attached label
i.e., the DLCI value in the frame.. This label is a logical identifier with local
significance. In the virtual call case, the value of the logical identifier and
other associated parameters such as layer 1 channel delay may be requested
and negotiated during call setup. Depending on the value of the parameters,
the network may accept or reject the call. In the case of the permanent
virtual circuit, the logical identifier and other associated parameters
are defined by means of administrative procedures e.g., at the time of subscription..
The usernetwork interface structure allows for the establishment of
multiple virtual calls or permanent virtual circuits, or both, to many destinations over a single access channel.
Specifically, for each connection, the bearer service:
1. Provides bidirectional transfer of frames.
2. Preserves their order as given at one usernetwork interface if and
when they are delivered at the other end. Note: No sequence numbers
are kept by the network. Networks are implemented in such a way that
frame order is preserved. .
3. Detects transmission, format, and operational errors such as frames
with an unknown label.
4. Transports the user data contents of a frame transparently. Only
frames address and FCS fields may be modified by network nodes.
5. Does not acknowledge frames.
At the usernetwork interface, the FRAD, as a minimum, has the following responsibilities:
1. Frame delimiting, alignment, and transparency provided by the use of
HDLC flags and zero-bit insertion.
2. Virtual circuit multiplexingrdemultiplexing using the address field of
the frame.
3. Inspection of the frame to ensure that it consists of an integer number
of octets prior to zero-bit insertion or following zero-bit extraction.
4. Inspection of the frame to ensure that it is not too short or too long.
5. Detection of transmission, format, and operational errors.
460
FRAME RELAY
A frame received by a frame handler may be discarded if the frame:

1. Does not consist of an integer number of octets prior to zero-bit
insertion or following zero-bit extraction.
2. Is too long or too short.
3. Has an FCS that is in error.
The network will discard a frame if it:
1. Has a DLCI value that is unknown.
2. Cannot be routed further due to internal network conditions.
A frame can be discarded for other reasons such as exceeding negotiated
throughout.
12.12
FRAME RELAY SIGNALING FUNCTIONS
Frame relay SVC circuits require setup procedures. PVC circuits require
periodic health and welfare checks. These functions are carried out by OSI
layer 3 signaling actions. In the case of PVC, the signaling functions are
reduced to a set of management processes. No setup or release procedures
are necessary. On the other hand, SVC operation requires circuit setup and
release processes and utilizes a somewhat extended group of management
action messages.
12.12.1
PVC Management Procedures
These procedures are based on periodic transmission of a STATUS ENQUIRY message by the DTE and a STATUS message by the DCE. These
messages are principally used to make a provisioned circuit active; if it is not,
then to determine the general health of the circuit. This includes link
integrity verification of the DTErDCE interface, notification to the DTE of
the status of a PVC, notification of the addition of a PVC, and the detection
by the DTE of the deletion of a PVC. These messages are transferred across
the bearer channel using layer 2 unnumbered information UI. frames.
12.12.1.1 Message Definition. Both messages are transferred on DLCI
s 0. The FECN, BECN, and DE bits are not used and set to 0. The 3 octets
following the address field have fixed values. Figure 12.9 illustrates the PVC
management frame format.
461
Figure 12.9. PVC management frame format with 2-octet addressing. (From Figure 7rX.36,
These 3 octets are coded as follows:
The first octet is the control field of a UI frame with P bit set to 0.
The second octet is the protocol discriminator information element of
the message.
The third octet is the dummy call reference information element of the
message.
The other information elements are reviewed below.

STATUS ENQUIRY Message. This message is sent to request the status of
PVCs or to verify link integrity. Specific information elements for this
message are described in Table 12.6, and they are in the order indicated in
the table.
TABLE 12.6
Message-Specific Information Elements in STATUS ENQUIRY Message
Message type: STATUS ENQUIRY

Significance: Local
Information Element
Message type
Report type
Link integrity verification
a
Length in octets.
Source: Table 2rX.36, page 15, Ref. 10.
Direction: Both
Direction
Type
Length a
Both
Both
Both
Mandatory
Mandatory
Mandatory
1
3
4
462
FRAME RELAY
TABLE 12.7
Message-Specific Information Elements in a STATUS Message
Message type: STATUS

Significance: Local
Information Element
Message type
Report type
Link integrity verification
PVC status
Direction: Both
Direction
Type
Length a
Both
Both
Both
Both
Mandatory
Mandatory
Optional r Mandatory (Note)
Optional r Mandatory (Note)
1
3
4
57
Note: Optional or mandatory depending on the type of report.

a
Length in octets.
Source: Table 3rX.36, page 15, Ref. 10.
STATUS Message. This message is sent in response to a STATUS ENQUIRY message to indicate the status of PVCs or for a link integrity
verification. Optionally, it may be sent at any time to indicate the status of a
single PVC. Message-specific information elements for this message are given
in Table 12.7, and they are given in the order indicated in the table.
Moreover, the PVC status information element may occur several times.
Message Type Coding (Octet 7)
STATUS ENQUIRY message:
STATUS message:
0111 0101
0111 1101
Report Type Coding (Octets 8, 9, and 10)

Octet 8: 0101 0001
Octet 9: Length of report type contents
Octet 10: Type of report
0000 0000 Full status status of all PVCs on the DTEDCE interface .
0000 0001 Link integrity verification only
0000 0010 Single PVC asynchronous status
All other values are reserved.
The purpose of the Report Type information element is to indicate the type
of enquiry requested when included in a STATUS ENQUIRY message or
the contents of the STATUS message. The length of this information
element is 3 octets.
Link Integrity Verification (Octets 11, 12, 13, and 14)
Octet 11: 0101 0011
Octet 12: Length of link integrity verification contents s 2
463
Octet 13: Send sequence number

Octet 14: Receive sequence number
The purpose of the Link Integrity verification information element is to
exchange sequence numbers between the DCE and DTE on a periodic basis.
The length of the contents of the link integrity information element is binary
encoded in octet 2. The send sequence number in octet 3 indicates the
current send sequence number of the originator of the message, and the
receive sequence number in octet 4 indicates the send sequence number
received in the last received message.
PVC Status. The purpose of the PVC status information element is to
indicate the status of existing PVCs at the interface. The information
element can be repeated, as necessary, in a message to indicate the status of
all PVCs on the DTEDCE interface. The length of this information
element depends on the length of the DLCIs being used on the DTEDCE
interface. The length of this information element is 5 octets when a default
address format 2 octets. is used. The format of the PVC status information
element is defined in Figure 12.10, where the default address format is used.
Bit 6 of octet 3 is the most significant bit in the DLCI.
Bit 2 of the last octet of each PVC status information element is Active
bit, which is coded 1 to indicate the PVC is active and coded 0 to indicate
the PVC is inactive. An active indication means that the PVC is available to
be used for data transfer. An inactive indication means that the PVC is
configured but is not available for data transfer.
Bit 3 of the last octet of each PVC information element is the Delete
bit, which is coded 1 to indicate that the PVC is deleted and coded 0 to
indicate that the PVC is configured.
Bit 4 of the last octet for each PVC status information element is New
bit, which is coded 1 to indicate the PVC is newly configured and coded 0 to
indicate that the PVC is already configured.
The PVC status information elements are arranged in the messages in
ascending order of DLCIs; the PVC with the lowest DLCI is first, the second
lowest is second, and so on. The maximum number of PVCs that can be
indicated in a message is limited by the maximum frame size.
The Delete bit is only applicable for timely notification using the
optional PVC asynchronous status report. When this bit is set to 1, the New
and Active bits have no significance and are set to 0 upon transmission and
not interpreted upon reception.
12.12.1.2 Description of Operation: Health and Welfare, PVC Link. The
operation is based on periodic polling, as described below, to verify the
integrity of the link and to report the status of PVCs.
464
FRAME RELAY
465
Periodic Polling
General. The DTE initiates the polling as described below.
1. The DTE sends STATUS ENQUIRY message to the DCE and starts
polling timer T391. When T391 expires, the DTE repeats the above
action.
The STATUS ENQUIRY message typically requests a link integrity
verification exchange only report type 0000 0001.. However, every N391
polling cycle, the DTE requests full status of all PVCs report type 0000
0000..
2. The DCE responds to each STATUS ENQUIRY message with a
STATUS message and starts or restarts the polling verification timer
T392 used by the network to detect errors. The STATUS message sent
in response to a STATUS ENQUIRY contains the link integrity verification and Report Type information elements. If the content of the
Report Type information element specifies full status, then the STATUS message must contain one PVC status information element for
each PVC configured on the DTEDCE interface.
3. The DTE shall interpret the STATUS message depending upon the
type of report contained in this STATUS message. The DCE may
respond to any poll with a full status message in case of a PVC status
change or to report the addition or deletion of PVC on the DTEDCE
interface. If it is a full status message, the DTE should update the
status of each configured PVC.
Link Integrity Verification. The purpose of the link integrity verification
information element is to allow the DTE and the DCE to determine the
status of the signaling link DLCI s 0.. This is necessary since these procedures use the unnumbered information UI. frame. Figure 12.11 shows the
normal link integrity verification procedure.
The DTE and DCE maintain the following internal counters:
The send sequence counter maintains the value of the send sequence
number field of the last link integrity verification information element
sent.
The receive sequence counter maintains the value of the last received
send sequence number field in the link integrity verification information
element and maintains the value to be placed in the next transmitted
received sequence number field.
466
FRAME RELAY
Figure 12.11. Link integrity verification. (From Figure 11rX.36, page 20, ITU-T Rec. X.36,
Ref. 10.)
The following procedure is used:

1. Before any messages are exchanged, the DCE and DTE set the send
sequence counter and receive sequence counter to zero.
2. Each time the DTE sends a STATUS ENQUIRY message, it increments the send sequence counter and places its value in the send
sequence number field. It also places the current value of the receive
sequence counter into the receive sequence number field of the link
integrity verification information element. The DTE increments the
send sequence counter using modulo 256. The value zero is skipped.
3. When the DCE receives a STATUS ENQUIRY from the DTE, the
DCE checks the receive sequence number received from the DTE
against its send sequence counter.
The received send sequence number is stored in the receive sequence
counter. The DCE then increments its send sequence counter and places its
current value in the send sequence number field and the value of the receive
sequence number counter the last received send sequence number. into the
receive sequence number field of the outgoing link integrity verification
information element. The DCE then transmits the completed STATUS
message back to the DTE. The DCE increments the send sequence counter
using modulo 256. The value zero is skipped.
4. When the DTE receives a STATUS from the DCE in response to a
STATUS ENQUIRY, the DTE checks the receive sequence number
received from the DCE against its send sequence counter. The received
send sequence number is stored in the receive sequence counter.
467
Note: The value zero in the receive sequence number indicates that the
field contents are undefined; this value is normally used after initialization.
The value zero is not to be used in the send sequence number field so that
the receive sequence number will never contain the value zero to differentiate the undefined condition from the normal modulo round-off.
Signaling the Actiity Status of PVCs. In response to a STATUS ENQUIRY
message sent by the DTE containing a Report Type information element sent
to full status, the DCE reports in a STATUS message to the DTE the
activity status of each PVC configured on the DTEDCE interface with PVC
status information elements one per PVC..
The Report Type information element in this STATUS message is set to
full status. Also in response to a STATUS ENQUIRY message set by the
DTE containing a Report Type information element set to link integrity
verification only, the DCE may respond with a STATUS message containing
a Report Type information element set to full status in case of a PVC
status change. Each PVC status information element contains an Active bit
indicating the activity status of that PVC.
The action that the DTE takes based on the value of the Active bit is
independent of the action based on the New bit. The DTE could get a PVC
status information element with the New bit set to 1 and the Active bit set
to 0.
If the DTE receives a PVC status information element with the Active bit
set to 0, the DTE shall stop transmitting frames on the PVC until it receives
a PVC status information element for that PVC with the Active bit set to 1.
When the Active bit is set to 1, the Delete bit must be set to 0 on
transmission. The Delete bit is not interpreted in the full status reporting
STATUS message. When the Delete bit set to 1 in the optional asynchronous
status message, the active bit has no significance. Other action taken by the
DTE is implementation-dependent.
Since there is a delay between the time the network makes a PVC active
and the time the DCE transmits a PVC status information element notifying
the DTE, there is a possibility of the DTE receiving frames on a PVC
marked as inactive. The action the DTE takes on receipt of frames on an
inactive PVC is implementation-dependent.
Since there is a delay between the time the network detects that a PVC
has become inactive and the time the DCE transmits a PVC status information element notifying the DTE, there is a possibility of the DCE receiving
frames on an inactive PVC. The action the DCE takes on receipt of frames
for an inactive PVC is network-dependent and may include the dropping of
frames on the inactive PVC.
The DCE will signal that a PVC is active if the following criteria are met:
The PVC is configured and available for data transfer in the network
from the local DCE to the remote DCE.
468
FRAME RELAY
There is no service affecting condition at both local and remote

DTEDCE interfaces.
In the case bidirectional procedures are used at the remote DTEDCE
interface, the remote DTE indicates that the PVC is present and active.
Note that in the case bidirectional procedures are used at the local DTEDCE
interface, this indication is independent of the indication received from the
local DTE.
This section is based on Ref. 10.
12.12.2
Signaling Required for SVCs
Switched virtual circuits SVCs. at the UNI require call setup and release
procedures. Certain link maintenance messages, similar to those discussed in
Section 12.12.1, are also required for SVC operation. Table 12.8 lists call
control messages employed in the various connection control procedures.
Note that the entries in Table 12.8 are presented in alphabetical order,
not in chronological operational. order. The readers attention is called to
the fact that STATUS ENQUIRY and STATUS messages that are described
here differ from those given in Section 12.12.1. They are not interchangeable.
The most important message in the Call Establishment group is the
SETUP message. This message is sent by the calling user to the network and
by the network to the called user to initiate the actual call establishment
procedure. Its 23 information elements in sequential order are shown in
Table 12.9.
The first three information elements of all call control messages listed in
Table 12.9 have some commonality. The first octet is the protocol discriminator. This simply tells the receiving device that this is a call control message. It
distinguishes it from other message types. It is identical for all the call control
messages listed in Table 12.8 and is coded 00001000. It should be noted that
these call control messages are valid for UNI and are not necessarily valid for
the NNI.
TABLE 12.8
Messages Required for Call Connection Control of SVCs

Message Type
Message Type
Call Establishment Messages
Call Clearing Messages
ALERTING
CALL PROCEEDING
CONNECT
CONNECT ACKNOWLEDGE
PROGRESS
SETUP
DISCONNECT
RELEASE
RELEASE COMPLETE
Miscellaneous Messages
STATUS
STATUS ENQUIRY
TABLE 12.9
469
SETUP Message Content
Message type: SETUP

Significance: Global
Information Element
Protocol discriminator
Call reference
Message type
Bearer capability
Channel identification
Data-link connection identifier
Progress indicator
Network-specific facilities
Display
End-to-end transit delay
Packet layer binary parameters
Link layer core parameters
Link layer protocol parameters
X.213 priority
Calling party number
Calling party subaddress
Called party number
Called party subaddress
Transit network selection
Repeat indicator
Low layer compatibility
High layer compatibility
Useruser
Direction: Both
Direction
Type
Length
Both
Both
Both
Both
Both
Both
Both
Both
nu
Both
Both
Both
Both
Both
Both
Both
Both
Both
un
Both
Both
Both
Both
M
M
M
M
O (Note 1)
O (Note 2)
O (Note 3)
O (Note 4)
O (Note 5)
O (Note 7)
O (Note 8)
O (Note 9)
O (Note 10)
O (Note 11)
O (Note 12)
O (Note 13)
O (Note 14)
O (Note 15)
O (Note 16)
O (Note 17)
O (Note 18)
O (Note 19)
O (Note 20)
1
2*
1
45
2*
26
24
2*
(Note 6)
211
23
227
29
28
2*
223
2*
223
2*
1
216
24
(Note 21)
470
FRAME RELAY
The second part of every message is the call reference. It has a minimum
length of 2 octets. Its purpose is to identify the call or facility
registrationrcancellation at the local UNI to which the particular message
applies. The call reference does not have end-to-end significance across the
network.
Bits 8, 7, 6, and 5 of the first octet of the call reference are coded 0000;
bits 4, 3, 2, and 1 provide the length of the call reference values. The second
octet, bit 8, is a flag; the remainder of the octet and octets 3 onwards, if
required, carry the actual call reference value.
The third information element is message type. Its intent is to identify the
purpose of the message being sent. The coding of the message is shown in
Table 12.10. Note that bit 8 the first bit of an octet as shown. is coded binary
0 and is reserved for future use as a possible extension bit.
TABLE 12.10
471
Coding for Message Types
Bits
8765 4321
0000 0000
Escape to nationally specified message types; see Note.
000- ---0 0001

0 0010
0 0111
0 1111
0 0011
0 0101
Call establishment messages:

ALERTING
CALL PROCEEDING
CONNECT
CONNECT ACKNOWLEDGE
PROGRESS
SETUP
010- ---0 0101

0 1101
1 1010
Call clearing messages:

DISCONNECT
RELEASE
RELEASE COMPLETE
011- ---0 0000

1 1101
1 0101
Miscellaneous messages:
SEGMENT
STATUS
STATUS ENQUIRY
Note: When used, the message type is defined in the following octet (s ), according to
national specification. The extension mechanism (bit 8 of the message type) is
independent of the escape mechanism for the message.
Source: Table 4-1rQ.933, page 15, Ref. 11.
Each information element has a unique identifier coding as illustrated in

Table 12.11. This 1-octet element tells what the information element is going
to do or what information it carries.
Due to space limitation, we cannot explore each call control message and
information element. It will be useful, however, to explore several examples.
The SETUP message was presented in Table 12.9. Another example is shown
in Figure 12.12, which gives the link layer core parameters information
element. The specific coding of the core parameters is given in Table 12.12,
which consists of three parts. Figure 12.13 shows the format of the called
party number information element. Table 12.13 gives the called party information element coding.
The information element Link Layer Core Parameter, Figure 12.12
and the three subsequent tables deal with configuring an SVC. Among the
configuration parameters are Outgoing and Incoming Maximum Frame Mode
Information field size octets 3ad., Throughput octets 4 and 5., and
Committed and Excess Burst Size octets 69.. Some or all of these parameters can be included in the SLA service level agreement. issued at contract
signing or can be negotiated on-the-fly during SVC setup.
472
FRAME RELAY
TABLE 12.11
Information Element Unique Identifier Coding

Max Length
(Octets)
(Note 1)
8765 4321
1:::: ---000 ---001 ---101 ---0::: ::::
000 0000
000 0100
000 1000
001 0100
001 1000
001 1001
001 1110
010 0000
010 1000
100 0010
100 0100
100 1000
100 1001
100 1100
100 1101
101 0000
101 0001
101 0011
101 0111
110 1100
110 1101
111 0000
111 0001
111 1000
111 1100
111 1101
111 1110
111 1111
Single octet information elements:

reserved
Shift (Note 2)
Repeat indicator
Variable length information elements:
Segmented message
Bearer capability
Cause (Note 2)
Call state
Data-link connection identifier (Note 7)
Progress indicator (Note 2)
Network specific facilities (Note 2)
Display
End-to-end transit delay
Packet layer binary parameters (Note 7)
Link layer core parameters (Note 7)
Link layer protocol parameters (Note 7)
Connected number
Connected subaddress
X.213 priority (Note 7)
Report type (Note 7)
Link integrity verification (Note 7)
PVC status (Notes 2 and 7)
Called party number
Transit network selection (Note 2)
Low layer compatibility (Note 6)
Useruser
Escape for extension (Note 3)
All other values are reserved (Note 5)
1
1
5
32
3
(Note 4)
6
4
(Note 4)
82
11
3
27
9
(Note 4)
23
8
3
4
5
(Note 4)
23
(Note 4)
23
(Note 4)
14
4
131
473
Figure 12.12. Format of the Link Layer Core Parameters information element. (From Figure
4-4rQ.933, page 21, ITU-T Rec. Q.933, Ref. 11.)
Note 1. All the parameters are optional and position independent. If certain parameters are
not included, the network default value will be used. The term outgoing is defined in the
direction from calling user to called user. The term incomingis defined in the direction from
called user to calling user.
Note 2. When octet 4 is present, octet 4a and octet 4b shall also be present. Additionally,
octet group 4c and 4d may be included.
Table notes continued on page 474.
474
FRAME RELAY
Called party number, Figure 12.13, can be based on either ITU-T Rec.
E.164 numbering or ITU-T Rec. X.121 numbering. Table 12.13 describes
octet 3 Figure 12.13., bits 4, 3, 2, and 1. If they are coded 0001, the E.164
numbering plan applies; if coded 0011, the X.121 numbering plan applies.
Octets. 4 onwards gives the applicable number digits. IA5 refers to CCITT
International Alphabet No. 5. Refer to CCITT Rec. T.50 Ref. 13..
12.13
COMPATIBILITY ISSUES
Standardization agencies have not exactly standardized frame relay. Early on

we identified three such agencies:
1. ITU
2. ANSI
3. The Frame Relay Forum
The Frame Relay Forum generally follows ITU-T practice. ANSI Frame
Relay specifications parallel the ITU in many respects. We cite only a few of
the differences. One is STATUS ENQUIRY and STATUS. Several information elements differ; even the number of information elements. Q.933 SVC
STATUS ENQUIRY and STATUS for SVC differs from X.36 for PVC.
Then there are proprietary frame relay systems. One such is AT & Ts
closed-loop system. All we need is one octet differing from one protocol
system to another, and incompatibility results. Customer beware!
Notes to Figure 12.12 Continued.

Note 4. Included only in the SETUP message.
Note 7. Throughputand Measurement interval (T)are defined as Committed Information
Rate (CIR) and Committed rate measurement interval (Tc ) in Rec. I.370, respectively. See
Section 12.9.2.
Note 8. Octet group 8 is used to indicate the committed burst size magnitude when the value
cannot be coded in octet group 6. When the incoming committed burst size field is not
included (in octet group 6), the incoming magnitude field has no significance.
Note 9. Octet group 9 is used to indicate the excess burst size magnitude when the value
cannot be coded in octet group 7. When the incoming excess burst size field is not included
(in octet group 6), the incoming magnitude field has no significance.
TABLE 12.12
475
Link Layer Core Parameters Information Element, Specific Coding
Maximum frame mode information field (FMIF) (octets 3, 3a, 3b, 3c, and 3d)
The frame mode information field size is the number of user data octets after the address field
and before the FCS field in a frame mode frame. The count is done either before zero-bit
insertion or following zero-bit extraction. If the frame mode information field is symmetrical,
octets 3a and 3b indicate the size in both directions, and octets 3c and 3d are absent. The
maximum size of the frame mode information field is a system parameter and is identified as
N203.
Outgoing maximum FMIF size (octets 3a and 3b)
The outgoing maximum FMIF size is used to indicate the maximum number of end user data
octets in a frame for the calling user to the called user direction. The size is in octets and is
encoded in binary.
Incoming maximum FMIF size (octets 3c and 3d)
The incoming maximum FMIF size is used to indicate the maximum number of end user data
octets in a frame for the called user to the calling user direction. The size is in octets and is
encoded in binary.
Note 1. The maximum frame mode information field size allowed for the D-channel is 262
octets. For B- and H-channels a larger FMIF size is allowed (e.g., up to 4096 octets). The
default value of the frame mode information field size is defined in Recommendation Q.922.
The users may negotiate a maximum FMIF size that is lower than the maximum FMIF size that
the network can deliver. The network is not required to enforce the negotiated FMIF value.
Throughput (octets 4, 4a, 4b, 4c, and 4d)
The purpose of the throughput field is to negotiate the throughput for the call. Throughput is
the average number of frame mode information field bits transferred per second across a
user-network interface in one direction, measured over an interval of duration T.
This field, when present in the SETUP message, indicates requested throughput, which is the
lesser of the throughput requested by the calling user and throughput available from the
network(s), but is not less than the minimum acceptable throughput. When present in the
CONNECT message, it indicates the agreed throughput, which is the throughput acceptable
to the calling user, the called user, and the network(s).
If the throughput is asymmetrical (i.e., the values in the incoming and outgoing directions are
different), octets 4a and 4b indicate throughput in the outgoing direction (from the calling
user) and octets 4c and 4d indicate throughput in the incoming direction (to the calling user).
If the throughput is symmetrical, octets 4a and 4b indicate throughput in both directions, and
octets 4c and 4d are absent.
Throughput is expressed as an order of magnitude (in powers of 10) and an integer multiplier.
For example, a rate of 192 kbps is expressed as 192 = 10 3 .
Magnitude (octets 4a and 4c)
This field indicates the magnitude of the throughput. This is expressed as a power of 10.
Bits
7 6 5
0 0 0 10 0
0 0 1 10 1
0 1 1 10 2
0 1 1 10 3
1 0 0 10 4
1 0 1 10 5
1 1 0 10 6
Table continued on page 476
476
FRAME RELAY
TABLE 12.12
( Contined )
Note 2. To ensure that various implementations will encode particular rates in a consistent
fashion, the coding of the magnitude and multiplier shall be such that the multiplier is as small
as possible; i.e. the multiplier shall not be evenly divisible by 10. For example, a rate of 192
kbps shall be expressed as 192 = 10 3 , not as 1920 = 10 2 .
Multipler (octets 4a, 4b, 4c, and 4d)
This field indicates, in binary, the value by which the magnitude shall be multiplied to obtain
the throughput.
Minimum acceptable throughout
The purpose of the minimum acceptable throughput field is to negotiate the throughput for
the call. Minimum acceptable throughput is the lowest throughput value that the calling user is
willing to accept for the call. If the network or the called user is unable to support this
throughout, the call shall be cleared.
This field, which is present only in the SETUP message, is carried unchanged through the
network(s). Its value may not be greater than the requested throughput.
If the minimum acceptable throughput is asymmetrical (i.e., the values in the incoming and
outgoing directions are different), octets 5a and 5b indicate minimum acceptable throughput
in the outgoing direction (from the calling user) and octets 5c and 5d indicate minimum
acceptable throughput in the incoming direction (to the calling user). If the minimum
acceptable throughput is symmetrical, octets 5a and 5b indicate throughput in both directions, and octets 5c and 5d are absent.
Minimum acceptable throughput is expressed as an order of magnitude (in powers of 10) and
an integer multiplier. For example, a rate of 192 kbps is expressed as 192 = 10 3 .
Magnitude (octets 5a and 5c)
Same as octets 4a and 4c coding.
Multiplier (octets 5a, 5b, 5c, and 5d)
This field indicates, in binary, the value by which the magnitude shall be multiplied to obtain
the minimum acceptable throughput.
Committed burst size
This field indicates the maximum amount of data (in bits) that the network agrees to transfer,
under normal conditions, over a measurement interval ( T ). This data may or may not be
interrupted (i.e., may appear in one frame or in several frames, possibly with interframe idle
flags). T is calculated using the following combinations:
Throughput
Committed burst
size ( Bc )
(Note 3)
Excess burst
size ( Be )
Measurement interval
(T )
)0
)0
)0
T s Bc rThroughput
)0
)0
s0
T s Bc rThroughput
s0
s0
)0
Default (Note 4)
T s Be rAccess rate
Note 3. The coding of this field is in octets. Therefore, the committed burst size is 8 = the
contents of this field. If the committed burst size is symmetrical, octets 6a and 6b indicate the
size in both directions, and octets 6c and 6d are absent.
TABLE 12.12
477
( Continued )
Note 4. However, the ingress and egress Access Rate (AR) do not have to be equal.
However, when the ingress AR is substantially higher than the egress AR, continuous input of
Be frames at the ingress interface may lead to persistent congestion of the network buffers at
the egress interface, and a substantial amount of the input Be data may be discarded.
Networks may define the interval T to be less than Be rAR in which case the default is not
used.
Outgoing committed burst size (octets 6a and 6b)
The outgoing committed burst size (in octets) is binary coded.
Incoming committed burst size (octets 6c and 6d)
The incoming committed burst size (in octets) is binary coded.
Excess burst size
This field indicates the maximum amount of uncommitted data (in bits) that the network will
attempt to deliver over measurement interval ( T ). This data may appear in one frame or in
several frames. If the data appears in several frames, these frames can be separated by
interframe idle flags. Excess burst may be marked Discard Eligible (DE) by the network.
Note 5. The coding of this field is in octets. Therefore, the excess burst size is 8 = the
contents of the field.
If the excess burst size is symmetrical, octets 7a and 7b indicate the size in both directions,
and octets 7c and 7d are absent.
Outgoing excess burst size (octets 7a and 7b)
The outgoing excess burst size (in octets) is binary coded.
Incoming excess burst size (octets 7c and 7d)
The incoming excess burst size (in octets) is binary coded.
Committed burst size magnitude (octets 8 and 8a)
The purpose of the Committed burst size magnitude is to indicate the magnitude of the
Committed burst size. It is expressed as a power of 10. It is multiplied by the Committed burst
size value (octet group 6) to give the actual value of the Committed burst size. When the
incoming Committed burst size field is not included (in octet group 6), the incoming
magnitude has no significance.
The outgoing and incoming Bc magnitudes are coded as follows:
Bits
3r6
2r5
1r4
10 0
10 1
10 2
10 3
10 4
10 5
10 6
478
FRAME RELAY
TABLE 12.12
( Continued )
Excess burst size magnitude (octets 9 and 9a)

The purpose of the Excess burst size magnitude is to indicate magnitude of the Excess burst
size. It is expressed as a power of 10. It is multiplied by the Excess burst size value (octet
group 7) to give the actual field of the Excess burst size. When the incoming Excess burst
size value is not included (in octet group 7), the incoming magnitude has no significance.
The outgoing and incoming Be magnitude are coded as a power of 10 as follows:
Bits
3r6
2r5
1r4
10 0
10 1
10 2
10 3
10 4
10 5
10 6
All others values are reserved.

Source: Table 4-5rQ.933, ITU-T Rec. Q.933, Ref. 11.
Figure 12.13. Called party number information element. (From Figure 4-14rQ.931, page 75,
ITU-T Rec. Q.931, ITU, Ref. 13.)
TABLE 12.13
479
Called Party Information Element Coding
Type of number (octet 3) (Note 1)

Bits
7 6 5
0 0 0
Unknown (Note 2)
0 0 1
International number (Note 3)
0 1 0
National number (Note 3)
0 1 1
Network specific number (Note 4)
1 0 0
Subscriber number (Note 3)
1 1 0
Abbreviated number (Note 5)
1 1 1
Reserved for extension

Note 1. For the definition of international, national, and subscriber number, see ITU-T
Recommendation 1.330.
Note 2. The type of number unknown is used when the user or the network has no
knowledge of the type of number, (e.g., international number, national number, etc.). In this
case the number digits field is organized according to the network dialing plan; for example,
prefix or escape digits might be present.
Note 3. Prefix or escape digits shall not be included.
Note 4. The type of number network specific number is used to indicate
administrationrservice number specific to the serving network (e.g., used to access an
operator).
Note 5. The support of this code is network-dependent. The number provided in this
information element presents a shorthand representation of the complete number in the
specified numbering plan as supported by the network.
Numbering plan identification (octet 3)
Numbering plan (applies for type of number s 000, 001, 010, and 100)
Bits
4 3 2 1
0 0 0 0
Unknown (Note 6)
0 0 0 1
ISDN r telephony numbering plan (ITU-T Rec. E.164)
0 0 1 1
Data numbering plan (ITU-T Rec. X.121)
0 1 0 0
Telex numbering plan (ITU-T Rec. F.69)
1 0 0 0
National standard numbering plan
1 0 0 1
Private numbering plan
1 1 1 1
Reserved for extension

Table continued on page 480
480
FRAME RELAY
TABLE 12.13
( Continued )
Note 6. The numbering plan unknown is used when the user or network has no knowledge
of the numbering plan. In this case the number digits field is organized according to the
network dialing plan; for example, prefix or escape digits might be present.
Number digits (octets 4, etc.)
This field is coded with IA5 characters, according to the formats specified in the appropriate
numberingrdialing plan.
Note that the E.164 and X.121 numbering plans may be found in the 3rd ed., Reference Manual for
Telecommunication Engineering, 3rd edition, John Wiley & Sons, June 2001.
Source: Table 4-9, page 76, ITU-T Rec. Q.931, ITU Geneva, May 1998, Ref. 13.
REFERENCES
1. Framework for Frame Mode Bearer Serices, ITU-T Rec. I.122, ITU, Geneva,
March 1993.
2. Frame Mode Bearer Serices, CCITT Rec. I.233, ITU Geneva, 1992.
3. ISDN Signaling Specification for Frame Relay Bearer Serice for Digital Subscriber
Signaling System Number 1 DSS1., ANSI T1.617-1991, ANSI, New York, 1991.
4. ISDN Data Link Layer Serice for Frame Mode Bearer Serices, CCITT Rec. Q.922,
ITU Geneva, 1992.
5. Integrated Serices Digital Network ISDN . Core Aspects of Frame Protocol for
Use with Frame Relay Bearer Serice, ANSI T1.618-1991, ANSI, New York, 1991.
6. ISDNArchitectural Framework and Serice Description for Frame Relaying Bearer
Serice, ANST T1.606-1990, ANSI, New York, 1990.
7. Telecommunications: Integrated Serices Digital Network ISDN . Data Link Layer
Signaling Specification for Application at the UserNetwork Interface, ANSI T1.6021989, ANSI, New York, 1989.
8. Frame Relay & SMDS, a seminar, Hewlett-Packard Co., Burlington, MA, October
1993.
9. Integrated Serices Digital Network ISDN . Architectural Framework and Serice
Description for Frame-Relaying Bearer Serice Congestion Management and Frame
Size ., ANSI T1.606a, ANSI, New York, 1992.
Equipment DCE . for Public Data Networks Proiding Frame Relay Data Transmission Serice by Dedicated Circuit, ITU-T Rec. X.36, ITU Geneva, October 1995.
11. Integrated Serices Digital Network ISDN . Digital Subscriber Signaling System No.
1 DSS 1. Signaling Specifications for Frame Mode Switched and Permanent
Virtual Connection Control and Status Monitoring, ITU-T Rec. Q.933, ITU Geneva,
October 1995.
REFERENCES
481
12. International Reference Alphabet IRA. Formerly International Alphabet No. 5 or

IA5 . Information Technology7-bit Coded Character Set for Information Interchange, CCITT Rec. T.50, ITU Geneva, September 1992.
13. ISDN UserNetwork Interface Layer 3 Specification for Basic Call Control, ITU-T
Rec. Q.931, ITU Geneva, May 1998.
14. Network-to-Network Interface Between Public Data Networks Proiding the Frame
Relay Data Transmission Serice, ITU-T Rec. X.76, ITU Geneva, April 1995.
15. Frame Relay Bearer Serice Network-to-Network Interface Requirements, ITU-T Rec.
I.372, ITU Geneva, March 1973.
16. Switched Virtual Circuits, ITU-T Rec. X.76, Amendment 1, ITU Geneva, August
1997.
13
INTEGRATED SERVICES
DIGITAL NETWORKS (ISDNs)
13.1
INTRODUCTION
Integrated services digital network ISDN. might be defined as an extension

of the digital telephone network to the subscriber where both voice and data
are available in an integrated way. In theory, ISDN supports a wide variety of
telecommunication services such as:
Voice
TelexrTeletext
Telemetry
Data: circuit- and packet-switched
Conference TV, slow scan TV
Facsimile, including CCITT Group 4
ISDN is based on the public switched digital network, namely, the familiar
PCM network discussed in Chapter 7. In other words, the principal vehicle
for signal transport is the 64-kbps digital channel. When ISDN is implemented at a user premise, it assumes that access is available for 64-kbps
channels at the local serving exchange and that CCITT signaling system No. 7
is in place and operational for the intervening network end-to-end.
ISDN integrates services. It does so by providing a 64-kbps bearer channel
for all the services listed. In North America the digital channel may be a
56-kbps channel.
13.2
13.2.1
ISDN STRUCTURES
ISDN User Channels
At the ISDN usernetwork interface, the two most common structures for
user access transmission links are the B channel at 64-kbps transmission rate
483
484
INTEGRATED SERVICES DIGITAL NETWORKS (ISDNs)
and the D channel at 16 kbps. The B channel is the basic user channel and
serves any one of the traffic types listed in Section 13.1.
The D channel is a 16-kbps channel primarily used for signaling. Secondarily, if capacity permits, it may be used for lower-speed data connectivity to
the network.
Note: The D channel, when used for primary rate service, is a 64-kbps
channel. In this case, it is used exclusively for signaling.
There are also H channels for higher transmission rates. The H0 channel
is 384 kbps; there is an H1 group of transmission rates where H11 is 1536
kbps and H12 is 1920 kbps.
The H channels do not carry signaling. Their purpose is to serve the
higher data rate user for such things as conference television, digitized
program channels for broadcasters ., fast facsimile, and higher data rate
service for computer data.
13.2.2
Basic and Primary Rate Interfaces
The basic interface structure is composed of two B channels and a D

channel, commonly expressed as 2B q D. The D channel at this interface is
16 kbps. Often the literature uses the abbreviation BRI for basic rate
interface.
Appendix I to CCITT Rec. I.412 ITU-T I.412 in Ref. 1. states that the
basic access may also be B q D or just D.
The primary rate B-channel interface structures are composed of n B
channels and one D channel, where the D channel in this case is 64 kbps.
There are two primary data rates:
1.544 Mbps s 23B q D from the North American T1 configuration .
2.048 Mbps s 30B q D from the European E1 configuration .
For the usernetwork access arrangement containing multiple interfaces,
it is possible for the D channel in one structure not only to serve the signaling
requirements of its own structure but also to serve another primary rate
structure without an activated D channel. When a D channel is not activated,
the designated time slot may or may not be used to provide an additional B
channel, depending on the situation, such as 24B with 1.544 Mbps.
There are a number of H-channel interface structures where we would
assume a primary rate configuration. H0, for example, is useful for video
conferencing with its 384-kbps data rate i.e., 6 = 64 kbps.. Three H0
channels can be accommodated within the North American DS1 T1. in time
slots 16, 712, and 1318. If a signaling channel is not required, then a
fourth H0 channel can be handled in time slots 1824.
At the E1 2.048-Mbps rate, five H0 channels can be accommodated and
the preferred assignment is to use channels 16, 712, 1319 excluding
USER ACCESS AND INTERFACE
485
channel 16., 2025, and 2631. Any time slot not used for H0 channels can
be used for B channels. Of course, channel 16 in the E-1 configuration is the
signaling channel.
There are two categories of H1 channels: H11 and H12. H11 is the
1.536-Mbps rate, and H12 is the 1920-kbps rate. This may be carried on the
DS1 1.544-Mbps configuration, but a signaling channel, if required, would
have to be carried on a separate DS1 configuration. For E-1, H1 would be
carried in time slots 125 excluding 16. of the 2.048-Mbps primary rate
interface.
The industry often uses the acronym PRI to designate the primary rate
interface.
13.3
13.3.1

General
The objective of the ISDN designers was to provide a telecommunication

service that would be ubiquitous and universal. Whether this ambitious goal
is being met can be argued, especially in light of the evolving Asynchronous
Transfer Mode ATMChapter 16. and the several varieties of Digital
Subscriber Line DSL..* This latter can provide downstream transmission
rates as high as 6 Mbps on a subscriber pair. Other last mile broadband
communication techniques also compete such as LMDS and cable television
CATV.. See Chapter 17.
ISDN was designed around the copper transmission medium i.e., the
twisted pair.. Within 20 years much of the copper plant will be replaced by
fiber optics Chapter 15.. Some futurists argue 10 years and others 30 years.
Nevertheless, enterprise networks now demand transmission rates in excess
of what ISDN has been designed for. In the business community, this
demand, among other requirements, is for local area network LAN. connectivity over wide area network WAN. distances. Fiber can certainly do a
better job than wire pair. Corporate intranets require megabit service rather
than kilobit service offered by ISDN on copper pairs. ISDN is still very
attractive for small businesses and residential service.
Figure 13.1 shows generic ISDN user connectivity to the network. We can
select either the basic or primary rate service e.g., 2B q D, 23B q D, or
30B q D. to connect to the ISDN network. The objectives of any digital
interface design, and specifically those of ISDN access and interface, are:
1. Electrical and mechanical specification
2. Channel structure and access capabilities
*ISDN basic rate may also be considered a DSL digital subscriber line. technique.
LMDS stands for Local Multipoint Distribution System. This is a radio system that operates in
the 28- to 30-GHz band.
486
Figure 13.1. ISDN generic users.
3.
4.
5.
6.
Usernetwork protocols
Maintenance and operation
Performance
Services
ISDN specifications may be found in the ITU-T I Recommendations and,

for North American application, in relevant Telcordia and ANSI documents.
In either case, expect these to cover the six items listed previously.
Figure 13.2 shows the conventional ISDN reference model. It delineates
interface points for the user. In the figure, NT1 network termination 1.
provides the physical layer interface; it is essentially equivalent to open
system interconnection OSI. layer 1. The functions of the physical layer
include:
Transmission facility termination

Layer 1 maintenance functions and performing monitoring
Timing
Power transfer
Layer 1 multiplexing
Interface termination, including multidrop termination employing layer
1 contention resolution
Figure 13.2. ISDN reference model.
487
Network termination 2 NT2. can be broadly associated with OSI layers 1,

2, and 3. Among the examples of equipment that provide NT2 functions are
user controllers, servers, LANs, and private automatic branch exchanges
PABXs.. Among the NT2 functions are:
Layers 1, 2, and 3 protocol processing

Multiplexing layers 2 and 3.
Switching
Concentration
Interface termination and other layer 1 functions
Maintenance functions
A distinction must be drawn here between North American and European

practice. In Europe, telecommunication administrations have historically
been, in general, national monopolies that are government controlled. In
North America i.e., United States and Canada. they are private enterprises,
often very competitive. Thus in Europe, in many cases, the NT1 is the
property of the telecommunication administration and the NT2 is the responsibility of the customer. Of course, in North America they both belong to the
user, and there is a U interface between NT1 and ISDN switch. On the
other hand, CCITT defines NT1 as part of the network even though it is on
the users premises.
It should be noted that there is an overall trend outside of North America
to privatize telecommunications such as has happened in the United Kingdom and is scheduled to take place in other countries such as Germany,
Mexico, and Venezuela.
TE1 in Figure 13.2 is the terminal equipment and has an interface that
complies with ISDN terminalnetwork interface specifications at the S
interface. We call this equipment ISDN-compatible. TE1 covers functions
broadly belonging to OSI layer 1 and higher OSI layers. Among the equipment are digital telephones, computer workstations, and other devices in the
user end-equipment category that are ISDN compatible.
TE2 in Figure 13.2 refers to equipment that does not meet the ISDN
terminalnetwork interface at point S. TE2 adapts the equipment to meet
the ISDN terminalnetwork interface. This process is assisted by the terminal adapter TA..
Reference points T, S, and R are used to identify the interface available at
those points. T and S are identical electrically, mechanically, and from the
point of view of protocol. Point R relates to the TA interface, or, in essence,
it is the interface of that nonstandard i.e., non-ISDN. device. The U
interface provides for transmission facility compatibility and is referenced in
North American ISDN specifications.
We will return to usernetwork interfaces once the stage is set for ISDN
protocols looking into the network from the user.
488
13.4
ISDN PROTOCOLS AND PROTOCOL ISSUES
When fully implemented, ISDN will provide both circuit and packet switching. For the circuit-switching case, now fairly broadly available in North
America, the B channel is fully transparent to the network, permitting the
user to utilize any protocol or bit sequence so long as there is end-to-end
agreement on the protocol that will be used. Of course the protocol itself
should be transparent to any bit sequence.
It is the D channel that carries the circuit-switching control function for its
related B channels. Whether it is the 16-kbps D channel associated with BRI
or the 64-kbps D channel associated with PRI, it is that channel which
transports the signaling information from the users ISDN terminal from NT
to the first serving telephone exchange of the telephone company or administration. Here the D-channel signaling information is converted over to
CCITT No. 7 signaling data employing ISUP ISDN user part. of signaling
system No. 7. Thus it is the D channels responsibility for call establishment
setup., supervision, termination takedown., and all other functions dealing
with network access.
The B channel in the case of circuit switching is serviced by NT1 or NT2
using OSI layer 1 functions only. The D channel carries out OSI layers 1, 2,
and 3 functions such that the B-channel protocol established by a family of
ISDN end users will generally make layer 3 null in the B channel where the
networking function is carried out by the associated D channel.
With packet switching, two possibilities emerge. The first basically relies
on the B channel to carry out OSI layers 1, 2, and 3 functions at separate
packet-switching facilities PSFs.. The D channel is used to set up the
connection to the local switching exchange at each end of the connection.
This type of packet-switched offering provides 64-kbps service. The second
method utilizes the D channel exclusively for lower data rate switched service
Figure 13.3. Simplified concept of ISDN switching.
ISDN PROTOCOLS AND PROTOCOL ISSUES
489
Figure 13.4. Detailed diagram of the ISDN circuit-switching concept.
where the local interface can act as a CCITT ITU-T. X.25 data communication equipment DCE. device.
Figure 13.3 is a simplified conceptual diagram of ISDN circuit switching. It
shows the B channel riding on the public digital network and also shows the
D channel, which is used for signaling. Of course, the D channel is a separate
channel. It is converted to a CCITT No. 7 signaling structure and in this form
may traverse several signal transfer points STPs. and may be quasi-associated or disassociated from its companion B channel s.. Figure 13.4 is a
detailed diagram of the same ISDN circuit-switching concept. The reader
should note the following in the figure: a. Only users at each end have a
peer-to-peer relationship available for all seven OSI layers of the B channel.
As the call is routed through the system, there is only layer 1 physical layer.
interaction at each switching node along the call route. b. The D channel
requires the first three OSI layers for call setup to the local switching center
at each end of the circuit. c. The D-channel signaling data are turned over
to CCITT signaling system No. 7 SS No. 7. at the near- and far-end local
switching centers. d. SS No. 7 also utilizes the first three OSI layers for
circuit establishment, which requires the transfer of control information. In
SS No. 7 terminology this is called the message transfer part. There is a fourth
layer called the user part in SS No. 7. There are three user parts: telephone
user part, data user part, and ISDN user part ISUP. depending on whether
the associated B channel is in telephone, data, or ISDN service for the user.
490
Figure 13.5. ISDN packet service for lower data rates using the D channel. TA, terminal
adaptor; NT, network termination 2 andror 1; ET, exchange termination; TE1, terminal
equipment 1; PH, packet handling function. Note 1: This figure is only an example of many
possible configurations and is included as an aid to the text describing the various interface
functions. Note 2: In some implementations, the PH functions logically belonging to the ISDN
may reside physically in a node of the PSPDN. The service provided is still the ISDN virtual
circuit service. Note 3: See Recommendation X.325. Note 4: See Recommendation X.320.
Note 5: This connection is either on demand or semipermanent, but has no relevance with the
user network procedures. Only internal procedures between the ET and the PH are required.
(From Figure 2-3 r X.31, page 6, ITU-T Rec. X.31, Ref. 2.)
Figure 13.5 shows packet service for lower data rates where the D channel
is involved.
13.5
ISDN NETWORKS
In this context, ISDN networking is seen as a group of access attributes

connecting an ISDN user at either end to the local serving exchange i.e., the
local switch.. This is shown in Figure 13.6, the basic architectural model
of ISDN.
Figure 13.7 shows the ISDN reference configuration of public ISDN
connection type, Figure 13.8 illustrates the access connection element model,
Figure 13.9 shows the national tandemrtransit connection element model,
and Figure 13.10 shows the private ISDN access connection element.
The national transit network CCITT terminology. is the public switched
digital telephone network. That network, whether DS1DS4-based or
E1E5-based,* provides the two necessary attributes for ISDN compatibility:
1. 64-kbps channelization
2. Separate channel signaling based on CCITT Signaling System No. 7
*See Chapter 7 for descriptions of DS1DS4 and E1E5 digital network configurations.
ISDN NETWORKS
491
Figure 13.6. The basic architectural model of ISDN. TE, terminal equipment; BLLF, basic
low-layer functions; ALLF, additional low-layer functions; BHLF, basic high-layer functions;
AHLF, additional high-layer functions. Note a: In certain national situations, ALLF may also be
implemented outside the ISDN, in special nodes or in certain categories of terminals. Note b:
The ISDN local functional capabilities correspond to functions provided by a local exchange
and possibly including other equipment, such as electronic cross-connect equipment,
muldexes, and so on. Note c: These functions may either be implemented within ISDN or be
provided by separate networks. Possible applications for basic high-layer functions and for
additional high-layer functions are contained in Recommendation I.210. Note d: For signaling
between international ISDNs, CCITT signaling system No. 7 shall be used. (From Figure
1rI.324, page 3, CCITT Rec. I.324, Ref. 3.)
Figure 13.7. Reference configuration of public ISDN connection type. IRP, internal reference
point; CRF, connection-related functions; CE, connection element. (From Figure 3rI.324,
page 8, CCITT Rec. I.324, Ref. 3.)
492
Figure 13.8. Access connection element model. NT1, network termination 1; S, 64-kbps
circuit switch; C, signaling handling and exchange control functions; Mpx, (remote) multiplexer; RSU, remote switching unit andror concentrator; CRF, connection-related function.
(From Figure 4rI.324, page 9, CCITT Rec. I.324, Ref. 3.)
Figure 13.9. National tandem r transit connection element model. S, 64-kbps circuit switch;
C, signaling handling and exchange control functions; CRF, connection-related function.
(From Figure 5rI.324, page 10, CCITT Rec. I.324, Ref. 3.)
Figure 13.10. Private ISDN access connection element. S, 64-kbps circuit switch; CH,
signaling handling and private ISDN exchange control function; CRF, connection-related
function. (From Figure 8rI.324, page 11, CCITT Rec. I.324, Ref. 3.)
ISDN PROTOCOL STRUCTURES
493
Connections from the user at the local connecting exchange interface

include:
Basic service BRI. 2B q D s 192 kbps CCITT. specified .

s 160 kbps North American ISDN 1.
Both rates include overhead bits.
Primary service 23B q Dr30B q D s 1.544r2.048 Mbps
13.6
13.6.1

ISDN and OSI
Figure 13.11 shows the ISDN relationship with OSI. OSI was discussed in
Chapter 2.. As is seen in the figure, ISDN concerns itself with only the first
three OSI layers. OSI layers 4 to 7 are peer-to-peer connections and the
end-users responsibility. Remember that the B channel is concerned with
OSI layer 1 only. We showed the one exception; that is, when the B channel
is used for packet service, it will have those first three OSI layers to interface.
The D channel with its signaling and control functions is the exception to
the previous statement. The D channel interfaces with CCITT Signaling
System No. 7 at the first serving exchange. D channels handle three types of
information: signaling s., interactive data p., and telemetry t..
The layering of the D channel has followed the intent of the OSI
reference model. The handling of the p and t data can be adapted to the OSI
Figure 13.11. A generic communication context showing ISDNs relationship with the
seven-layer OSI model. Note that the end-system protocol blocks may reside in the subscribers TE or network exchanges or other ISDN equipment.
494
Figure 13.12. Correspondence among the ISDN D channel, CCITT SS No. 7, and the OSI
model. MTP, message transfer part; UP, user part. (From Ref. 4. Copyright IEEE, New York,
1985.)
model; the s data, by its very nature, cannot.* Figure 13.12 shows the
correspondence between D-channel signaling protocols, SS No. 7 levels, and
the OSI seven-layer model.
13.6.2
Layer 1 Interface, Basic Rate (CCITT)
The SrT interface of the reference model, Figure 13.2 or layer 1 physical
interface ., requires a balanced metallic transmission medium i.e., copper
pair. for each direction of transmission four-wire. capable of supporting 192
kbps. Again, this is the NT interface of the ISDN reference model.
Layer 1 provides the following services to layer 2 for ISDN operation:
1. The transmission capability by means of an appropriately encoded bit
stream of the B and D channels and any timing and synchronization
functions that may be required.
2. The signaling capability and the necessary procedure to enable customer terminals andror network terminating equipment to be deactivated when required and reactivated when required.
3. The signaling capability and necessary procedures to allow terminals to
gain access to the common resource of the D channel in an orderly
fashion while meeting the performance requirements of the D-channel
signaling system.
*CCITT No. 7 signaling system, like any other signaling system, has the primary quality of
service measure as post-dial delay. This is principally the delay in call setup. To reduce the
delay time as much as possible, it is incumbent upon system engineers to reduce processing time
as much as possible. Thus SS No. 7 truncates OSI to 4 layers, because each additional layer
implies more processing time, thus more post-dial delay.
TABLE 13.1
495
Primitives Associated with Layer 1

Specific Name
Generic
L1 l L2
PH-DATA
PH-ACTIVATE
PH-DEACTIVATE
M l L1
MPH-ERROR
MPH-ACTIVATE
MPH-DEACTIVATE
MPH-INFORMATION
Parameter
Priority
Indicator
Message
Unit
X
X
X
X
(Note 2)
X
X
X
Request
Indication
X
(Note 1)
X
Message Unit
Content
Layer 2 peer-to-peer
message
Type of error or
recovery from a
previously reported
error
Connectedr
disconnected
Note 1: PH-DATA request implies underlying negotiation between layer 1 and layer 2 for the acceptance
of the data.
Note 2: Priority indication applies only to the request type.
Source: Table 1rI.430, ITU-T Rec. I.430, 11r95, page 2 (Ref. 6 ).
4. The signaling capability and procedures and necessary functions at

layer 1 to enable maintenance functions to be performed.
5. An indication to higher layers of the status of layer 1.
13.6.2.1 Primitives Between Layer 1 and Other Entities. A primitie is
data relating to the development or use of software that is used in developing
measures or quantitative descriptions of software. Primitives are directly
measurable or countable, or may be given a constant value or condition for a
specific measure IEEE definition, Ref. 5..
In this discussion, primitives represent in an abstract way the logical
exchange of information and control between layer 1 and other entities or
interfaces.
The primitives to be passed across the boundary between layers 1 and 2 or
to the management entity are defined and summarized in Table 13.1. The
parameter values associated with these primitives are also summarized in the
table. For further information, the reader may consult ITU-T Rec. I.211
which describes the syntax and use of these primitives.
13.6.2.2 Interface Functions. The S and T functions for the BRI consist
of three time division multiplexed bit streams: two 64-kbps B channels and
one 16-kbps D channel for an aggregate bit rate of 192 kbps. Of this 192
kbps, the 2B q D configuration accounts for only 144 kbps. The remaining 48
kbps are overhead bits whose function is briefly described in the following
text.
496
Figure 13.13. Frame structure at reference points S and T. This is the frame structure of the
multiplexed bit stream for a 2B q D configuration, ITU-T interface, level 1. (From Figure
3rI.430, page 7, ITU-T Rec. I.430, Ref. 6.)
The functions covered at the interface include bit timing at 192 kbps to
enable the TE and NT to recover information from the aggregate bit stream.
This timing provides 8-kHz octet timing for the NT and TE to recover the
time division multiplexed channels i.e., 2B q D multiplexed .. Other functions include D-channel access control, power feeding, deactivation, and
activation.
Interchange circuits are required of which there is one in either direction
of transmission i.e., from the local switching point to and from the NT..
They are used to transfer digital signals across the interface. All of the
functions described above, except for power feeding, are carried out by
means of a digitally multiplexed signal. The format of this signal is shown in
Figure 13.13 and described in the next section.
It should be noted that NT1 in Europe is considered to be owned by the
network, meaning the local telephone company or administration. In the
United States, NT1 is owned by the customer. Transmission between
the network and the customer is 2-wire. 2-wire to 4-wire conversion takes
place in NT1. Echo suppressors at each end of the 2-wire transmission line
allow two-way data to pass without harmful interference one to the other.
13.6.2.3 Frame Structure. In both directions, the bits are grouped into
frames of 48 bits each. The frame structure is identical for all configurations,
whether point-to-point or point-to-multipoint. However, the frame structures
497
TABLE 13.2A Notes on Bit Positions and Groups for a BRI

Frame, Direction TE to NT
Bit Position
Group
1 and 2
3 to 11
12 and 13
14 and 15
16 to 24
25 and 26
27 to 35
36 and 37
38 to 46
47 and 48
Framing signal with balance bit

B1 channel (first octet) with balance bit
D-channel bit with balance bit
FA auxiliary framing bit for Q bit with balance bit
B2 channel (first octet) with balance bit
B1 channel (second octet) with balance bit
B2 channel (second octet) with balance bit
Source: Table 2rI.430, page 7 (Ref. 6 ).
TABLE 13.2B Notes on Bit Positions and Groups for a BRI

Frame, Direction NT to TE
Bit Position
1 and 2
3 to 10
11
12
13
14
15
16 to 23
24
25
26
27 to 34
35
36
37
38 to 45
46
47
48
Group
Framing signal with balance bit
B1 channel (first octet)
E-, D-echo channel bit
D-channel bit
Bit A used for activation
FA auxiliary framing bit
N bit (coded as defined in 6.3 of I.430)
B2 channel (first octet)
D-channel bit
M, multiframing bit
B1 channel (second octet)
D-channel bit
S
B2 channel (second octet)
D-channel bit
Frame balance bit
Note: The use of the S bit is optional, and when not used it is set to binary
ZERO.
are different for each direction of transmission. These structures are shown
in Figure 13.13, where the upper half of the figure shows the bit format
direction NT to TE; the lower half, TE to NT. Table 13.2A gives explanatory
notes for the direction TE to NT, and Table 13.2B, the direction NT to TE.
Allow that NT refers to the network side of the circuit and TE refers to the
terminal side. Note that each group is DC balanced by the last bit, the L bit.
498
13.6.2.4 Line Code. For both directions of transmission, pseudoternary

coding is used with 100% pulse width, as shown in the following figure:
Coding is performed such that a binary 1 is represented by no line signal 0

volts., whereas a binary 0 is represented by a positive or negative pulse. The
first binary signaling following the framing balance bit is the same polarity as
the balance bit. Subsequent binary 0s alternate in polarity. A balance bit is a
0 if the number of 0s following the previous balance bit is odd. A balance bit
is a binary 1 if the number of 0s following the previous balance bit is even.
The balance bit tends to limit the buildup of a DC component on the line.
13.6.2.5 Timing Considerations. The NT derives its timing from the
network clock. A TE synchronizes its bit, octet, and frame timing from the
NT, which has derived its timing from the ISDN bit stream being received
from the network. The NT uses this derived timing to synchronize its
transmitter clock.
13.6.2.6 BRI Differences in the United States. Telcordia and ANSI
prepared ISDN BRI standards fairly well modified from the CCITT I
Recommendation counterparts. Telephone company administrations in the
United States are at variance with those in most other countries. Furthermore, it was Telcordias intention to produce equipment that was cost-effective and that would easily interface with existing the North American telephone plant.
One point, of course, is where the telephone company responsibilities end
and customer responsibilities begin. This is called the U interface. The
tendency toward a 2-wire interface, rather than 4-wire interface, is another
point. Line waveform is yet another. Rather than a pseudoternary line
waveform, the United States uses 2B1Q. The line bit rate is 160 kbps rather
than that recommended by CCITT, namely, 192 kbps. More detailed explanation of North American practice follows. The 2B q D frames differ significantly from those recommended by CCITT.
DSL Bandwidth Allocation and Frame Structure. The digital subscriber line
DSL. modulation rate is 80 kilobauds, equivalent to 160 kbps. The effective
160-kbps signal is divided up into 12 kbps for synchronization words, 144
kbps for 2B q D of customer data, and 4 kbps of DSL overhead.
499
The synchronization technique used is based on transmission of nine

quaternary . symbols every 1.5 ms, followed by 216 bits of 2B q D data and 6
bits of overhead. The synchronization word provides a robust method of
conveying line timing and establishes a 1.5-ms DSL basic frame for multiplexing subrate signals. Every eighth synchronization word is inverted i.e.,
the 1s become 0s and the 0s become 1s. to provide a boundary for a 12-ms
superframe composed of eight basic frames. This 12-ms interval defines an
appropriate block of customer data for performance monitoring and permits
a more efficient suballocation of the overhead bits among various operational
functions.
The 48 overhead bits sometimes referred to as maintenance bits or M bits.
available per superframe are allocated into 24 bits for a 2-kbps eoc, 12 bits
for a crc check covering the 1728 bits of customer data in a superframe, plus
8 particular overhead bits, and one febe bit for communicating block errors
detected by the crc check to the far end. In determining relative positions of
the various overhead bit functions, febe and crc bits were placed away from
the start of the superframe to allow time for the crc calculation associated
with the preceding superframe to be completed.
The eoc is the embedded operations channel and permits network operations systems to access essential operations functionality in the NT1. There
are 12 cyclic redundancy check crc . bits within each superframe; they cover
the 1728 bits of 2B q D customer data plus the 8 M4 bits, in the previous
superframe. This provides a means of block error check. Two crc bits are
assigned to each of the basic frames 3 through 8. This is illustrated in Figures
13.14 and 13.15.
The acronym febe stands for far-end block error bit which is assigned
to each superframe. The febe bit indicates whether a block error was detected in the 2B q D customer data in a preceding superframe in the
opposite direction of transmission. An outgoing febe bit with a value of 1
indicates that the crc of the preceding incoming superframe matched the
value computed from the transmitted data; a value of 0 indicates a mismatch,
which, in turn, indicates one or more bit errors.
2B q D Customer Data Bit Pattern. There are 216 2B q D bits placed in
each 1.5-ms basic frame, for a customer data rate of 144 kbps. The bit
pattern before conversion to quaternary form and after reconversion to
binary form. for the 2B q D data is
B 1 B 1 B 1 B 1 B 1 B 1 B 1 B 1 B 2 B 2 B 2 B 2 B 2 B 2 B 2 B 2 DD
where B 1 and B 2 are bits from the B 1 and B 2 channels, and D is a bit from
the D channel. This 18-bit pattern is repeated 12 times per DSL basic frame.
The 2B1Q Waveform. It is convenient to express the 2B1Q waveform as
q3, q1, y1, y3 because this indicates symmetry about zero, equal spacing
500
Figure 13.14. LT to NT1 2B1Q superframe technique and overhead bit assignments.
NT1-to-LT superframe delay is offset by LT-to-NT1 superframe by 60 " 2 quats (about 0.75
ms). All bits other than the Sync word are scrambled. 1, reserved bit for future standard
(set s 1); ( )m and ( )s, magnitude bit and sign bit for given quat; act, activation bit; crc,
cyclic redundancy check: covers 2B q D and M4; dea, deactivation bit; eoc, embedded
operations channel; a, address bit; dm, datarmessage indicator; i, information (datarmessage); febe, far-end block error bit. (From Figure 3.1a, page 3-9, Ref. 7.)
between states, and convenient integer magnitudes. The block synchronization word SW. contains nine quaternary elements repeated every 1.5 ms:
q3,q 3,y 3,y 3,y 3,q 3,y 3,q 3,q 3
Line Configurations. One basic access is provided by a single, 2-wire DSL
interface. Operation on this 2-wire line is full-duplex. To avoid interface
between transmitted and received signals on the same line, an echo canceler
is used. Echo cancellation involves adaptively forming a replica of the echo
signal arriving at the receiver from its local transmitter and subtracting it
from the signal at the input of the receiver.
13.6.3
Layer 1 Interface, Primary Rate
This interface is applicable for the 1.544- or 2.048-Mbps data rates.
501
Figure 13.15. NT1 to LT 2B1Q superframe technique and overhead bit assignment. NT1-to-LT
superframe delay is offset by LT-to-NT1 superframe 60 " 2 quats (about 0.75 ms). All bits
other than the Sync word are scrambled. 1, reserved bit for future standard (set s 1). ( )m
and ( )s, magnitude bit and sign bit for given quat; act, activation bit; crc, cyclic
redundancy check: covers 2B q D and M4; cso, cold start only bit; dea, deactivation bit; febe,
far-end block error bit; eoc, embedded operations channel; a, address bit; dm, data r message
indicator; i, information (data r message); ntm, NT1 in Test Mode bit; nib, network indicator
bit from LULT and LUNT to LT; ps 1 and ps 2 , power status bits (s 1 from NT1 to LT). (From
Figure 3-1b, page 3-10, Ref. 7.)
13.6.3.1 Interface at 1.544 Mbps

Bit Rate and Synchronization
NETWORK CONNECTION CHARACTERISTICS. The network delivers except as
noted below. a signal synchronized from a clock having a minimum accuracy
of 1 = 10y11 stratum 1; see Ref. 8 for stratum definition.. When synchronization by a stratum 1 clock has been interrupted, the signal delivered by the
network to the interface will have a minimum accuracy of 4.6 = 10y6
stratum 3..
While in normal operation, the TE1rTArNT2 transmits a 1.544-Mbps
signal having an accuracy equal to that of the received signal by locking the
frequency of its transmitter signal to the long-term average of the incoming 1.544-Mbps signal or by providing equal signal frequency accuracy
from another source. ITU-T Rec. I.431 Ref. 9. advises against this latter
alternative.
502
RECEIVER BIT STREAM SYNCHRONIZED TO A NETWORK CLOCK
Receier Requirements. Receivers of signals across interface I a operate

with an average transmission rate in the range of 1.544 Mbps " 4.6
ppm. However, operation with a received signal transmission rate in the
range of 1.544 Mbps " 32 ppm is required in any maintenance state
controlled by signalsrmessages passed over the m bits and by an alarm
indication signal AIS.. In normal operation the bit stream is synchronized to stratum 1.
Transmitter Requirements. The average transmission rate of signals transmitted across interface I a by the associated equipment is the same as
the average transmission rate of the received bit stream.
Note: The I a and I b interfaces are located at the inputroutput port of TE
or NT.
TE1rTA OPERATING BEHIND AN NT2 THAT IS NOT SYNCHRONIZED TO A NETWORK
CLOCK
Receier Requirements. Receivers of signals across interface I a operate

with a transmission rate in the range of 1.544 " 32 ppm.
Transmitter Requirements. The transmitted signal across interface I a is
synchronized to the received bit stream.
SPECIFICATION OF OUTPUT PORTS.
The signal specification for output ports is
summarized in Table 13.3.
TABLE 13.3
Digital Interface at 1.544 Mbps
Bit rate
Pair(s) in each direction of transmission
Code
Test load impedance
Nominal pulse shape
Signal level
(Notes 2 and 3)
Power at 772 kHz

Power at 1544 kHz
1544 kbps
One symmetrical pair
B8ZS (Note 1)
100 ohm resistive
See pulse mask (Note 2) a
q12 dBm to q19 dBm
At least 25 dB below the
power at 772 kHz
Note 1: B8ZS is modified AMI code in which eight consecutive binary ZEROs are replaced with
000 q y 0 y q if the preceding pulse was positive (q ) and with 000 y q 0 q y if the preceding
pulse was negative (y ).
Note 2: The pulse mask and power level requirements apply at the end of a pair having a loss at
772 kHz of 0 to 1.5 dB.
Note 3: The signal level is the power level measured in a 3-kHz bandwidth at the output port for an
all binary ONEs pattern transmitted.
a
Given in Ref. 17, Chapter 7, Figure 7.24.

503
Figure 13.16. Frame structure of 1.544-Mbps interface.
TABLE 13.4
Multiframe Structure
F Bits
Assignments
Multiframe
Frame Number
Multiframe
Bit Number
FAS
Control r Maintenance
CRC Bits
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
194
387
580
773
966
1159
1352
1545
1738
1931
2124
2317
2510
2703
2896
3089
3282
3475
3668
3861
4054
4247
4440
e1
e2
e3
e4
e5
e6
FAS, frame alignment signal

Frame Structure. The frame structure is shown in Figure 13.16. Each time
slot consists of consecutive bits, numbered 1 through 8. Each frame is 193
bits long and consists of an F bit framing bit. followed by 24 consecutive
time slots. The frame repetition rate is 8000 frames per second.
Table 13.4 shows the multiframe structure called extended superframe in
the United States . which is 24 frames long. It takes advantage of the more
advanced search algorithms for frame alignment. There are 8000 F bits
504
frame alignment bits. transmitted per second i.e., 8000 frames a second, 1 F
bit per frame.. In 24 frames, with these new strategies, only 6 bits are
required for frame alignment as shown in Table 13.4. The remaining 18 bits
are used as follows; There are 12 m bits used for control and maintenance,
and the 6 e bits are used for CRC6 error checking.
Time Slot Assignment. Time slot 24 is assigned to the D channel, when this
channel is present.
A channel occupies an integer number of time slots and the same time slot
positions in every frame. A B channel may be assigned any time slot in the
frame, an H0 channel may be assigned any six slots in a frame in numerical
order not necessarily consecutive., and an H11 channel may be assigned
slots 1 to 24. The assignments may vary on a call-by-call basis.
CODES FOR IDLE CHANNELS, IDLE SLOTS, AND INTERFRAME TIME FILL. A pattern
including at least three binary 1s in an octet is transmitted in every time slot
that is not assigned a channel e.g., time slots awaiting channel assignment on
a per-call basis, residual slots on an interface that is not fully provisioned,
etc.. and in every time slot of a channel that is not allocated to a call in both
directions. Interframe layer 2. time fill consists of contiguous HDLC flags*
transmitted on the D channel when its layer 2 has no frames to send.
13.6.3.2 Interface at 2.048 Mbps

Frame Structure. There are 8 bits per time slot and 32 time slots per frame,
numbered 0 through 31. The number of bits per frame is 256 i.e., 32 = 8.,
and the frame repetition rate is 8000 frames per second. Time slot 0 provides
frame alignment, and time slot 16 is assigned to the D channel when that
channel is present. A channel occupies an integer number of time slots and
the same time slot position in every frame. A B channel may be assigned any
time slot in the frame; and an H0 channel may be assigned any six time slots,
in numerical order, not necessarily consecutive. The assignment of time slots
may vary on a call-by-call basis. An H12 channel is assigned time slots 115
and 1731 in a frame. Time slots 131 provide bit-sequence-independent
transmission.
Timing Considerations. The NT derives its timing from the network clock.
The TE synchronizes its timing bit, octet, framing. from the signal received
from the NT and synchronizes accordingly the transmitted signal. In an
unsynchronized conditionthat is, when the access that normally provides
network timing is unavailablethe frequency deviation of the free-running
clock shall not exceed "50 ppm. A TE shall be able to detect and to
interpret the input signal within a frequency range of "50 ppm.
*For information about HDLC flags, see Section 3.4.
OVERVIEW OF THE LAYER 2 INTERFACE: LAPD
505
Any TE which provides more than one interface is declared to be a

multiple access TE and is capable of taking the synchronizing clock frequency
from its internal clock generator from one or more than one access or all
access links. and synchronizing the transmitted signal at each interface
accordingly.
Codes for Idle Channels and Idle Time Slots, Interframe Fill. A pattern
including at least three binary 1s in an octet is transmitted in every time slot
that is not assigned to a channel e.g., time slots awaiting channel assignment
on a per-call basis, residual slots on an interface is not fully provisioned, etc..,
as well as in every time slot of a channel that is not allocated to a call in both
directions.
Interframe layer 2. time fill consists of contiguous HDLC flags 01111110.
which are transmitted on the D channel when its layer 2 has no frames to
send.
Frame alignment and CRC procedures can be found in ITU-T Rec. G.
706, paragraph 4 Ref. 10..
13.7 OVERVIEW OF THE LAYER 2 INTERFACE: LINK ACCESS

PROCEDURE FOR THE D CHANNEL (LAPD)
The link access procedure for the D channel LAPD.* is used to convey
information between layer 3 entities across the ISDN usernetwork interface
using the D channel.
A serice access point SAP. is a point at which the data-link layer provides
services to its next higher OSI layer or layer 3. Associated with each datalink layer OSI layer 2. is one or more data-link connection endpoints see
Figure 13.17.. A data-link connection endpoint is identified by a data-link
connection endpoint identifier as seen from layer 3 and by a data-link connection identifier DLCI. as seen from the data-link layer.
Cooperation between data-link-layer entities is governed by a specific
protocol to the applicable layer. In order for information to be exchanged
between two or more layer 3 entities, an association must be established
between layer 3 entities in the data-link layer using a data-link-layer protocol.
This association is provided by the data-link layer between two or more
SAPs, as shown in Figure 13.18. Data-link message units are conveyed
between data-link-layer entities by means of a physical connection. Layer 3
uses serice primities to request service from the data-link layer. A similar
interaction takes place between layer 2 and layer 1.
*Note the similarity between LAPD described here and HDLC covered in Chapter 3. In fact,
LAPD is a subset of HDLC.
506
Figure 13.17. Entities, service access points (SAPs), and endpoints. (From Figure 2rQ.920,
page 2, ITU-T Rec. Q.920, Ref. 11.)
Figure 13.18. Data-link connections between two or more SAPs. (From Figure 3rQ.920,
page 3, ITU-T Rec. Q.920, Ref. 11.)
Between the data-link layer and its adjacent layers there are four types of
service primitives:
1.
2.
3.
4.
Request
Indication
Response
Confirm
These functions are shown diagrammatically in Figure 13.19.

The REQUEST primitive is used where a higher layer is requesting
service from the next lower layer. The INDICATION primitive is used by a
layer providing service to notify the next higher layer of activities related to
the REQUEST primitive. The RESPONSE primitive is used by a layer to
acknowledge receipt from a lower layer of the INDICATION primitive. The
507
Figure 13.19. Functions of service primitives. Note: The same principle applies for
data-link-layerphysical-layer interactions. (From Figure 4rQ.920, page 4, ITU-T Rec. Q.920,
Ref. 11.)
CONFIRM primitive is used by the layer providing the requested service to

confirm that the requested activity has been completed.
Figure 13.20 shows the data-link-layer reference model. All data-link-layer
messages are transmitted in frames delimited by flags, where a flag is a
unique sequence bit pattern. The frame structure as defined in ITU-T Rec.
I.441 is briefly described in this section.
The LAPD includes functions for:
1. The provision of one or more data-link connections on a D channel.
Discrimination between the data-link connections is by means of a
DLCI contained in each frame.
2. Frame delimiting, alignment, and transparency, allowing recognition of
a sequence of bits transmitted over a D channel as a frame.
3. Sequence, control, which maintains the sequential order of frames
across a data-link connection.
4. Detection of transmission, format, and operational errors on a data
link.
5. Recovery from detected transmission, format, and operational errors.
Notification to the management entity of unrecovered errors.
6. Flow control.
There is unacknowledged and acknowledged operation. With unacknowledged operation, information is transmitted in unnumbered information UI.
frames. At the data-link layer the UI frames are unacknowledged. Transmission and format errors may be detected, but no recovery mechanism is
defined. Flow control mechanisms are also not defined. With acknowledged
508
Figure 13.20. Data-link-layer reference model. (From Figure 5rQ.920, page 5, ITU-T Rec.
Q.920, Ref. 11.)
operation, layer 3 information is transmitted in frames that are acknowledged

at the data-link layer. Error recovery procedures based on retransmission of
unacknowledged frames are specified. For errors that cannot be corrected by
the data-link layer, a report to the management entity is made. Flow control
procedures are also defined.
Unacknowledged operation is applicable for point-to-point and broadcast
information transfer. However, acknowledged operation is applicable only for
point-to-point information transfer.
Two forms of acknowledged information are defined:
1. Single-frame operation
2. Multiframe operation
For single-frame operation, layer 3 information is sent in sequenced information 0 SI0. and sequenced information 1 SI1. frames. No new frame is sent
until an acknowledgment has been received for a previously sent frame. This
means that only one unacknowledged frame may be outstanding at a time.
With multiple-frame operation, layer 3 information is sent in numbered
information I. frames. A number of I frames may be outstanding at the same
509
time. Multiple-frame operation is initiated by a multiple-frame establishment

procedure using set asynchronous balanced moderset asynchronous balanced mode extended SABMrSABME. command.
13.7.1
Layer 2 Frame Structure for Peer-to-Peer Communication
There are two frame formats used on layer 2 frames:

1. Format A, for frames where there is no information field
2. Format B, for frames containing an information field
These two frame formats are shown in Figure 13.21.
The following discussion briefly describes the frame content sequences
and fields. for the LAPD layer 2. frame.
Flag Sequence. All frames start and end with a flag sequence consisting of
one 0 bit followed by six contiguous 1 bits and one 0 bit. These flags are
called the opening and closing flags. If sequential frames are transmitted, the
closing flag of one frame is the opening flag of the next frame.
Figure 13.21. Frame formats for layer 2 frames. (From Figure 1rQ.921, page 20, ITU-T Rec.
Q.921, Ref. 12.)
510
Address Field. As shown in Figure 13.21, the address field consists of 2

octets and identifies the intended receiver of a command frame and the
transmitter of a response frame. A single-octet address field is used for
LAPB operation.
Control Field. The control field consists of 1 or 2 octets. It identifies the type
of frame, either command or response. It contains sequence numbers where
applicable. Three types of control field formats are specified:
1. Numbered information transfer I format.
2. Supervisory functions S format.
3. Unnumbered information transfers and control functions U format.
Information Field. The information field of a frame, when present, follows
the control field and precedes the frame check sequence FCS.. The information field contains an integer number of octets:
For SAPs supporting signaling, the default value is 128 octets.

For SAPs supporting packet information, the default value is 260 octets.
Frame Check Sequence (FCS) Field. The FCS field is a 16-bit sequence and
is the 1s complement of the modulo 2 sum of:
1. The remainder of X raised to the k power:
X 15 q X 14 X 1 q 1
divided by the generating polynomial
X 16 q X 12 q X 5 q 1
where k is the number of bits in the frame existing between but not
including the final bit of the opening flag and the first bit of the FCS,
excluding bits inserted for transparency, and
2. The remainder by modulo 2 division by the generating polynomial given
above of the product of X 16 by the content of the frame defined above.
Transparency, mentioned above, ensures that a flag or abort sequence is
not imitated within a frame. On the transmit side the data-link layer
examines the frame content between the opening and closing flag sequences
and inserts a 0 bit after all sequences with five contiguous 1 bits including
the last 5 bits of the FCS.. On the receive side the data-link layer examines
the frame contents between the opening and closing flag sequences and
discards any 0 bit that directly follows five contiguous 1 bits.
511
Figure 13.22. LAPD address field format. EA, address field extension bit; CrR, commandrresponse field bit; SAPI, service access point identifier; TEI, terminal endpoint identifier. (From
Figure 5rQ.921, page 23, ITU-T Rec. Q.920, Ref. 12.)
Address Field Format. The address field is shown in Figure 13.22. It contains address field extension bits EA., commandrresponse indication bit
CrR., a data-link-layer service access point identifier SAPI. subfield, and a
terminal endpoint identifier TEI. subfield.
ADDRESS FIELD EXTENSION BIT EA.. The address field range is extended by
reserving the first transmitted bit of the address field to indicate the final
octet of the address field. The presence of a 1 in the first bit position of an
address field octet signals that it is the final octet of the address field. The
double-octet address field for LAPD operation has bit 1 of the first octet set
to a 0 and bit 1 of the second octet set to a 1.
COMMANDrRESPONSE FIELD BIT CrR.. The CrR bit identifies a frame as
either a command or a response. The user side sends commands with the
CrR bit set to 0 and responses with the CrR bit set to 1. The network side
does the opposite; that is, commands are sent with the CrR bit set to 1, and
responses are sent with the CrR bit set to 0.
In keeping with HDLC rules,* commands use the address of the peer
data-link entity while responses use the address of their own data-link-layer
entity. In accordance with these rules, both peer entities on a point-to-point
data-link connection use the same DLCI composed of a SAPITEI where
SAPI and TEI conform to the definitions contained as follows.
Service Access Point Identifier (SAPI). The SAPI identifies a point at which
data-link services are provided by a data-link-layer entity to a layer 3 or
management entity. Consequently, the SAPI specifies a data-link-layer entity
that should process a data-link-layer frame. The SAPI allows 64 SAPs to be
specified, where bit 3 of the address field octet containing the SAPI is the
least significant binary digit and bit 8 is the most significant. The SAPI values
are allocated as shown in Table 13.5.
*LAPD, as we know, is a derivative of HDLC. See Chapter 3.
512
TABLE 13.5
SAPI
Value
0
1
16
63
All Others
Allocation of SAPI Values

Related Layer 3 or Management Entity
Call control procedures
Reserved for packet mode communications using Q.931 call control
procedures
Packet communication conforming to X.25 Level 3 procedures
Layer 2 management procedures
Reserved for future standardization
Terminal Endpoint Identifier (TEI). The TEI for a point-to-point link connection may be associated with a single terminal equipment TE.. A TE may
contain one or more TEIs used for point-to-point data transfer. The TEI for
a broadcast data-link connection is associated with all user-side data-link-layer
entities containing the same SAPI. The TEI subfield allows 128 values where
bit 2 of the address field octet containing the TEI is the least significant bit
and bit 8 is the most significant bit. The following conventions apply for
assignment of these values:
1. TEI for broadcast data-link connection. The TEI subfield bit pattern
111 1111 s 127. is defined as the group TEI. The group TEI is
assigned to the broadcast data-link connection associated with the
addressed SAP.
2. TEI for point-to-point data-link connection. The remaining TEI values
are used for point-to-point data-link connections associated with the
addressed SAP. The range of TEI values are as follows:
Values 063:
Nonautomatic TEI assignment user equipment
Values 64126: Automatic TEI assignment user equipment
Control Field Formats. Control field formats are shown in Table 13.6. The
control field identifies the type of frame, either a command or a response.
The control field contains sequence numbers, where applicable. Three types
of control field formats are specified: numbered information transfer I
format., supervisory functions S format., and unnumbered information
transfers and control functions U format..
Further discussion of control field formats may be found in Section 3.4.
LAPD Primitives. The following comments clarify the semantics and usage of
primitives. Primitives consist of commands and their respective responses
associated with the services requested of a lower layer. The general syntax of
a primitive is:
XX-generic name-type: parameters
513
TABLE 13.6
Control Field Formats
Control Field Bits (Modulo 128)
N(S)
I format
N(R)
S format
N(R)
U format
N(S) Transmitter send sequence number

N(R) Transmitter receive sequence number
S
Supervisory function bit
M
PrF
X
1
Octet 4
Octet 4
PrF
PrF
5
Octet 4
Modifier function bit

Poll bit when issued as a command,
final bit when issued as a response
Reserved and set to 0
Source: Table 4rQ.921, page 26 (Ref.12 ).
TABLE 13.7
Primitives Associated with the Data-Link Layer

Type
Generic
Name
Parameters
Priority Message
Request IndicationResponse Confirm Indicator
Unit
L3 l L2
DL-ESTABLISH
DL-RELEASE
DL-DATA
X
X
X
X
X
X
X
X
DL-UNIT DATA
M l L2
MDL-ASSIGN
MDL-REMOVE
MDL-ERROR
X
X
X
X
X
MDL-UNIT DATA
MDL-XID
L2 l L1
PH-DATA
X
X
X
X
X
PH-ACTIVATE
PH-DEACTIVATE
M l L1
MPH-ACTIVATE
MPH-DEACTIVATE
MPH-INFORMATION
L3 l L2:
L2 l L1:
M l L2:
M l L1:
Layer 3rdata-link-layer boundary

Data-link-layerrphysical-layer boundary
Management entityrdata-link-layer boundary
Management entityrphysical-layer boundary
Source: Table 6rQ.921, page 32 (Ref. 12 ).
Message
Unit
Contents
Layer 3
peer-to-peer message
Layer 3
TEI value, CES
TEI value, CES
Reason for error
message
Management function
Connection
management
information
Data-link-layer
Connectedr
disconnected
514
where XX designates the layer providing the service. XX is DL for the

data-link layer, PH for the physical layer, or MDL for the management entity
to the data-link-layer interface. Table 13.7 gives the primitives associated
with the data-link layer.
13.8
OVERVIEW OF LAYER 3
The layer 3 protocol of course deals with the D channel and its signaling
capabilities. It provides the means to establish, maintain, and terminate
network connections across an ISDN between communicating application
entities. A more detailed description of the layer 3 protocol may be found in
ITU-T Rec. Q.931, Ref. 14..
Layer 3 utilizes functions and services provided by its data-link layer, as
described in Section 13.7 under LAPD functions. These necessary layer 2
support functions are listed and briefly described as:
Establishment of the data-link connection

Error-protected transmission of data
Notification of unrecoverable data-link errors
Release of data-link connections
Notification of data-link-layer failures
Recovery from certain error conditions
Indication of data-link-layer status
Layer 3 performs two basic categories of functions and services in the

establishment of network connections. The first category directly controls the
connection establishment. The second category includes those functions
relating to the transport of messages in addition to the functions provided by
the data-link layer. Among these additional functions are the provision of
rerouting of signaling messages on an alternative D channel where provided.
in the event of D-channel failure. Other possible functions include multiplexing and message segmenting and blocking. The D-channel layer 3 protocol is
designed to carry out establishment and control of circuit-switched and
packet-switched connections. Also, services involving the use of connections
of different types, according to user specifications, may be effected through
multimedia call control procedures.
Functions performed by layer 3 include:
1. The processing of primitives for communicating with the data-link
layer.
2. Generation and interpretation of layer 3 messages for peer-level communications.
OVERVIEW OF LAYER 3
515
3. Administration of timers and logical entities e.g., call references . used

in call control procedures.
4. Administration of access resources, including B channels and packetlayer logical channels e.g., ITU-T X.25..
5. Checking to ensure that services provided are consistent with user
requirements, such as compatibility, address, and service indicators.
The following functions may also be performed by layer 3:
1. Routing and Relaying. Network connections exist either between users
and ISDN exchanges or between users. Network connections may
involve intermediate systems which provide relays to other interconnecting subnetworks and which facilitate interworking with other
networks. Routing functions determine an appropriate route between
layer 3 addresses.
2. Network Connection. This function includes mechanisms for providing
network connections making use of data-link connections provided by
the data-link layer.
3. Coneying User Information. This function may be carried out with or
without the establishment of a circuit-switched connection.
4. Network Connection Multiplexing. Layer 3 provides multiplexing of call
control information for multiple calls onto a single data-link connection.
5. Segmenting and Reassembly SAR .. Layer 3 may segment and reassemble layer 3 messages to facilitate their transfer across a usernetwork
interface.
6. Error Detection. Error detection functions are used to detect procedural errors in the layer 3 protocol. Error detection in layer 3 uses,
among other information, error notification from the data-link layer.
7. Error Recoery. This includes mechanisms for recovering from detected errors.
8. Sequencing. This includes mechanisms for providing sequenced delivery of layer 3 information over a given network connection when
requested. Under normal conditions layer 3 ensures the delivery of
information in the sequence in which it is submitted by the user.
9. Congestion Control and User Data Flow Control. Layer 3 may indicate
rejection or unsuccessful indication for connection establishment requests to control congestion within a network. Typical is the congestion control message to indicate the establishment or termination
of flow control on the transmission of USER INFORMATION
messages.
10. Restart. This function is used to return channels and interfaces to an
idle condition to recover from certain abnormal conditions.
516
13.8.1
Layer 3 Specification
The layer 3 specification is contained in ITU-T Recs. Q.930r931 Refs. 13

and 14.. It includes both circuit-switched and packet-switched operation.
There are 23 message types for circuit-mode connection control. These are
shown in Table 13.8, and the content elements of each are given in over 50
tables in the specification. One typical table is Table 13.9. Additional tables
are provided for packet switching.
The following are several explanatory notes for the tables found in ITU-T
Rec. Q.931 and for Table 13.9. The letters M and O mean mandatory
and optional, respectively. Letters n and u refer to network and user,
respectively, and give the direction of traffic such as n u network to user.
and u n user to network.. An asterisk ). in the table means undefined
maximum length.
TABLE 13.8
Messages for Circuit-Mode Connection Control
Call establishment messages:

ALERTING
CALL PROCEEDING
CONNECT
CONNECT ACKNOWLEDGE
PROGRESS
SETUP
SETUP ACKNOWLEDGE
Call information phase messages:
RESUME
RESUME ACKNOWLEDGE
RESUME REJECT
SUSPEND
SUSPEND ACKNOWLEDGE
SUSPEND REJECT
USER INFORMATION
Call clearing messages:
DISCONNECT
RELEASE
RELEASE COMPLETE
Miscellaneous messages:
CONGESTION CONTROL
FACILITY
INFORMATION
NOTIFY
STATUS
STATUS ENQUIRY
Source: Table 3-1rQ.931, page 16, Ref. 14.
OVERVIEW OF LAYER 3
TABLE 13.9
SETUP Message Content
Information Element
Call reference
Message type
Sending complete
Repeat indicator
Bearer capability
Facility
Progress indicator
Network specific facilities
Display
Keypad facility
Signal
Switchhook
Feature activation
Feature indication
Called party number
Transit network selection
Low layer compatibility
Useruser
Direction
Type
Length
Both
Both
Both
Both
Both
Both
Both
Both
Both
Both
nu
un
nu
un
un
nu
Both
Both
Both
Both
un
Both
Both
Both
M
M
M
O (Note 1)
O (Note 2)
M (Note 3)
O (Note 4)
O (Note 5)
O (Note 6)
O (Note 7)
O (Note 8)
O (Notes 10, 12)
O (Note 11)
O (Note 12)
O (Note 12)
O (Note 12)
O (Note 13)
O (Note 14)
O (Note 15)
O (Note 16)
O (Note 17)
O (Note 18)
O (Note 19)
O (Note 20)
1
2*
1
1
1
413
2*
2*
24
2*
Note 9
234
23
23
24
25
2*
223
2*
223
2*
216
24
Note 21
517
518
13.8.1.1 General Message Format and Information Elements Coding.

Within this protocol, every message consists of the following parts:
1.
2.
3.
4.
Call reference
Message type
Other information elements, as required
Information elements 1, 2, and 3 are common to all messages and are always
present, while information element 4 is specific to each message type. This
organization is illustrated in the example shown in Figure 13.23.
The term default implies that the value defined should be used in the
absence of any assignment, or in the negotiation of alternative values. When
a field, such as the call reference value, extends over more than 1 octet, the
order of bit values progressively decreases as the octet number increases. The
least significant bit of the field is represented by the lowest-numbered bit of
the highest-numbered octet field.
Protocol Discriminator. The purpose of the protocol discriminator is to
distinguish messages for usernetwork call control from other messages to
be defined . within ITU-T Rec. Q.931. It also distinguishes messages of
ITU-T Rec. Q.931 from those OSI network layer protocol units which are
coded to other ITU-T Recommendations and other standards.
Call Reference. The purpose of the call reference is to identify the call or
facility registrationrcancellation request at the local usernetwork interface
to which the particular message applies. The call reference does not have
end-to-end significance across ISDNs.
OVERVIEW OF LAYER 3
519
Figure 13.23. General message organization example. (From Figure 4-1rQ.931, page 67,
ITU-T Rec. Q.931, ITU-T Rec. Q.931, Ref. 14.)
As a minimum, all networks and users must be able to support a. a call

reference value of 1 octet for a basic usernetwork interface and b. a call
reference value of 2 octets for a primary rate interface. The call reference
information element includes the call reference value and the call reference
flag.
Call reference values are assigned by the originating side of the interface
for a call. These values are unique to the originating side within a particular
D-channel layer 2 logical link connection. The call reference value is assigned
at the beginning of a call and remains fixed for the lifetime of the call except
in the case of call suspension .. After a call ends or after a successful
suspension, the associated call reference value may be reassigned to a later
call. Two identical call reference values on the same D-channel layer 2
logical link connection may be used when each value pertains to a call
originated at opposite ends of the link.
The call reference flag can take the values of 0 or 1. The call reference
flag is used to identify which end of the layer 2 logical link originated a call
reference. The origination side always sets the call reference flag to 0. The
destination always sets the call reference flag to 1.
Hence the call reference flag identifies who allocated the call reference
value for this call, and the only purpose of the call reference flag is to resolve
simultaneous attempts to allocate the same call reference value.
Message Type. The purpose of the message type is to identify the function
of the message being sent. For instance, 00000101 indicates a call setup
message.
Other Information Elements. Forty other information elements are listed,
such as sending complete, congestion level, call identity, datertime,
and calling party number.
520
13.9
13.9.1
ISDN PACKET MODE REVIEW

Introduction
Two main services for packet-switched data transmission are defined for
packet-mode terminals connected to the ISDN:
Case A: Access to a PSPDN PSPDN services. PSPDN s packet-switched
public data network.
Case B: Use of an ISDN virtual circuit service
13.9.2
Case A: Configuration When Accessing PSPDN Services
This configuration is shown in Figure 13.24 and refers to the service of Case
A, thus implying a transparent handling of packet calls through an ISDN.
Only access via the B channels is possible. In this context, the only support
that an ISDN gives to packet calls is a physical 64-kbps circuit-mode semipermanent or demand transparent network connection type between the appropriate PSPDN port and the X.25 DTE q TA or TE1 at the customer
premises.
In the case of semipermanent access, the X.25 DTE q TA or TE1 is
connected to the corresponding ISDN port at the PSPDN AU.. The TA,
when present, performs only the necessary physical channel rate adaption
Figure 13.24. Case A: Configuration when accessing PSPDN services. AU, ISDN access unit
ports; TA, terminal adaptor; NT, network termination 2 and r or 1; ET, exchange termination;
TE1, terminal equipment 1. (From Figure 2-1rX.31, page 3, ITU-T Rec. X.31, ITU-T Rec. X.31,
Ref. 2.)
521
between the user at the R reference point and the 64-kbps B-channel rate.
D-channel layer 3 messages are not used in this case.
In the case of demand access to the PSPDN, which is shown in the upper
portion of Figure 13.24, the X.25 DTE q TA or TE1 is connected to an
ISDN port at the PSPDN AU.. The AU is also able to set up 64-kbps
physical channels through the ISDN.
In this type of connection, an originating call will be set up over the B
channel toward the PSPDN port using the ISDN signaling procedure prior to
starting X.25 layer 2 and layer 3 functions. This is done by using either
hot-line e.g., direct call. or complete selection methods. Moreover, the TA,
when present, performs user rate adaption to 64 kbps. Depending on the
data rate adaption technique employed, a complementary function may be
needed at the AU of the PSPDN.
In the complete selection case, two separate numbers are used for outgoing access to the PSPDN:
The ISDN number of the access port of the PSPDN, given in the
D-channel layer 3 SETUP message Q.931.
The address of the called DTE indicated in the X.25 call request packet
The corresponding service requested in the D-channel layer 3 SETUP

message is ISDN circuit-mode bearer services.
For calls originated by the PSPDN, the same considerations as above
apply. In fact, with reference to Figure 13.24, the ISDN port of the PSPDN
includes both rate adaption if required. and path setting-up functions. When
needed, DTE identification may be provided to the PSPDN by using the call
establishment signaling protocols in D-channel layer 3 Q.931.. Furthermore,
DCE identification may be provided to the DTE, when needed, by using the
same protocols.
For the demand access case, X.25 layer 2 and layer 3 operation in the B
channel as well as service definitions are found in ITU-T Ref. X.32 Ref. 15..
Some PSPDNs may operate the additional DTE identification procedures
defined in Rec. X.32 Ref. 15. to supplement the ISDN-provided information
in Case A.
13.9.3
Case B: Configuration for the ISDN Virtual Circuit Service
This configuration refers to the case where a packet handling PH. function
is provided within the ISDN. The configuration in Figure 13.25 relates to the
case of X.25 link and packet procedures conveyed through the B channel. In
this case, the packet call is routed, within an ISDN, to some PH function
where the complete processing of the X.25 call can be carried out.
The PH function may be accessed in various ways depending on the
related ISDN implementation alternatives. In any case, a B-channel connection set up torfrom a PH port supports a. the necessary processing for
522
Figure 13.25. Case B: Configuration for the ISDN virtual circuit service (access via B
channel). TA, terminal adaptor; NT, network termination 2 andror 1; ET, exchange termination; TE1, terminal equipment 1; PH, packet handling function. Note: 1: This figure is only an
example of many possible configurations and is included as an aid to the text describing the
various interface functions. Note 2: In some implementations the PH functions logically
belonging to the ISDN may reside physically in a node of the PSPDN. The service provided is
still the ISDN virtual circuit service. Note 3: See Recommendation X.325. Note 4: See
Recommendation X.320. (From Figure 2-2rX.31, page 5, ITU-T Rec. X.31, Ref. 2.)
B-channel packet calls, b. standard X.25 functions for layer 2 and layer 3,
and c. possible path setting-up functions for layer 1 and possible rate
adaption.
The configuration shown in Figure 13.5 in Section 13.4. refers to the case
of X.25 packet layer procedures conveyed through the D channel. In this case
a number of DTEs can operate simultaneously through a D channel by using
connection identifier discrimination at ISDN layer 2. The accessed port of
PH is still able to support X.25 packet layer procedures.
It should be pointed out that the procedures for accessing a PSDTS
public switched data transmission service. through an ISDN usernetwork
interface over a B or D channel are independent of where the service
provider chooses to locate PH functionsthat is, in either a. a remote
exchange or packet-switching module in an ISDN or, b. the local exchange.
However, the procedures for packet access through the B channel or the D
channel are different.
In both cases of B- and D-channel accesses, in the service of Case B, the
address of the called DTE is contained in the X.25 call request packet. The
establishment of the physical connection from the TArTE1 to the packet
handling functions is done on the basis of the requested bearer service ISDN
virtual circuit service.; therefore, the user does not provide any addressing
information in the layer 3 procedures D channel, Q.931..
13.9.4
523
Service Aspects
13.9.4.1 Access to PSPDN ServicesCase A

Service Characteristics. The ISDN offers a 64-kbps circuit-switched or
semipermanent transparent network connection type between the TArTE1
and the PSPDN port AU.. In the switched access case the AU must be
selected by the called address in the D-channel signaling protocol when the
TArTE1 sets up the circuit-switched connection to the AU. In the nonswitched access case, layer 3 D channel, Q.931. call control messages are
not used.
Because the packet-switched service provider is a PSPDN, some DTEs are
PSPDN terminals; they are handled by the PSPDN. Other DTEs may access
the PSPDN without subscribing to the PSPDN permanently. In the first case,
the same services as PSPDN services are maintained, including facilities,
quality of service QoS. characteristics, and DTE-DCE interfaces. In the case
where a DTE is not subscribing to the PSPDN, it will be provided with a
limited set of PSPDN facilities see ITU-T Rec. X.32..
Every DTE will be associated with one or more ISDN E.164. numbers. In
addition, a DTE may be associated with one or more X.121 numbers
assigned by the PSPDNs. associated by the DTE. The method for X.25
packets to convey numbers from the ISDN numbering plan and the relationship with ITU-T Rec. X.121 are described in ITU-T Rec. E.166.
Basic Rules. Packet data communications, when using a switched B channel,
will be established by separating the establishment phase of the B channel
and the control phase of the X.25 virtual circuits using the X.25 protocol
link layer and packet layer.. In general, ISDN has no knowledge of the
customer terminal equipment or configuration. The incoming B-channel
connection establishment will have to employ the D-channel signaling procedure in ITU-T Rec. Q.931.
13.9.4.2 Access to the ISDN Virtual Circuit ServiceCase B
Service Characteristics. The virtual circuit service provided within the ISDN
is aligned with what is described in ITU-T Series Recommendations e.g., in
terms of facilities, quality of service, etc...
The service and facilities provided, as well as the quality of service
characteristics, are those of the ISDN. Existing features of the X-Series
Recommendations may be enhanced, and additional features may also be
developed taking into account the new ISDN customer capabilities. A number from the ISDN numbering plan will be associated with one or more
TArTE1. The numbering plan is set out in ITU-T Rec. E.164 Ref. 16..
User Access Capabilities. In this case both B and D channels can be used
for accessing the ISDN virtual circuit service.
524
Access Through the B ChannelBasic Rules. Packet data communications,

when using a switched B channel, is established by separating the establishment phase of the B channel and the control phase of the virtual circuits
using the X.25 protocol link layer and packet layer..
In general, an ISDN has no knowledge of the customer terminal equipment or configuration. In the demand access case, the incoming B-channel
connection establishment uses the signaling procedures of clause 6 of Q.931
Ref. 14. D-channel layer 3..
Access Through the D ChannelBasic Rules. The following must be adhered to for TE access to the PSDTS, as defined in the X-Series Recommendations particularly X.25..
A single SAPI s 16 LAPD link, as viewed by both the network and the
user, must support multiplexing of logical channels at X.25 layer 3. Additionally, because the user may have a multipoint access, and because a single TA
or TE1 is allowed to operate with more than one TE1 terminal endpoint
identifier ., the network must support the presence of multiple SAPI s 16
LAPD logical links simultaneously operating at ISDN layer 2. This results in
the requirement that the network be able to support simultaneous layer 2
and X.25 layer 3 multiplexing for D-channel packet mode connections.
All X.25 packets, including call request and incoming call packets, must be
transported to and from the TE in numbered information frames I-frames.
in an SAPI s 16 LAPD link.
An incoming call packet is transmitted to a TE only after the public
networks check at least the following:
Compatibility of user facilities contained in the incoming call packet

with the called subscriber profile when present
Availability of the X.25 logical channel, either two-way or incoming, on
which the incoming call packet is sent.
Section 13.9 is based on ITU-T Rec. X.31, Ref. 2.
REFERENCES
1. ISDN UserNetwork InterfacesInterface Structures and Access Capabilities, CCITT
Rec. I.412, Fascicle III.8, IXth Plenary Assembly, Melbourne, 1988.
2. Support of Packet Mode Terminal Equipment by an ISDN, ITU-T Rec. X.31, ITU
Geneva, November 1995.
3. ISDN Network Architecture, CCITT Rec. I.324, ITU Geneva, 1991.
4. William Stallings, ed., Tutorial: Integrated Serices Digital Network ISDN ., IEEE
Computer Society Press, Washington DC, 1985.
5. IEEE Standard Dictionary of Electrical and Electronic Terms, 6th edition, IEEE
Press, New York, 1996.
REFERENCES
525
6. Basic UserNetwork Interface Layer 1 Specification, ITU-T Rec. I.430, ITU Geneva,
November 1995.
7. ISDN Basic Access Transport System Requirements, Technical Reference TR-TSY000397, Issue 1, October 1988, Morristown, NJ.
8. Roger L. Freeman, Telecommunication Transmission Handbook, 4th edition, John
9. Primary Rate UserNetwork InterfaceLayer 1 Specification, ITU-T Rec. I.431,
ITU Geneva, March 1993.
10. Frame Alignment and Cyclic Redundancy Check Procedures Relating to Basic Frame
Structures Defined in Rec. G.704, CCITT Rec. G.706, ITU Geneva, April 1991.
11. Digital Subscriber Signaling System No. 1 DSS1. : ISDN UserNetwork Interface
Data Link LayerGeneral Aspects, ITU-T Rec. Q.920, ITU Geneva, March 1993.
12. ISDN UserNetwork InterfaceData Link Layer Specification, ITU-T Rec. Q.921,
ITU Geneva, September 1997.
13. ISDN UserNetwork Interface, Layer 3, General Aspects, ITU-T Rec. Q.930, ITU
Geneva, March 1993.
14. ISDN UserNetwork Interface, Layer 3For Basic Call Control, ITU-T Rec.
Q.931, ITU Geneva, May 1998.
15. Interface Between Data Terminal Equipment DTE . and Data Circuit Terminating
Equipment DCE . for Terminals Operating in the Packet Mode and Accessing the
Packet-Switched Public Data Network through a Switched Telephone Network or an
Integrated Serices Digital Network or a Circuit-Switched Public Data Network,
ITU-T Rec. X.32, ITU Geneva, October 1996.
16. The International Public Telecommunications Numbering Plan, ITU-T Rec. E.164,
ITU Geneva, May 1998.
17. Roger L. Freeman, Reference Manual for Telecommunications Engineering, 3rd
edition, John Wiley & Sons, New York, to be published.
14
BUILDING AND CAMPUS WIRING
AND CABLING FOR DATA
COMMUNICATIONS
14.1
BACKGROUND AND OBJECTIVE
The enterprise network has come of age. Its performance is only as good as
its underlying building wiring and cabling. The principal signal lead is the
twisted pair and will remain so for quite some time. Longer runs are based
on fiber-optic cable.
Ideally, a wiring and cabling plan for signal and ground leads and cabling
should be incorporated in a building design layout prior to construction or
renovation. Included in the plan is the strategic placement of
wiringrequipment closets, signal entry locations, and convenient cross-connects Ref. 1..
The principal objective of this chapter is to describe and suggest layouts of
structured cabling systems for commercial buildings. Both single-tenant and
multitenant structures are covered. Our concern, of course, is the cabling
that will carry telecommunication signals with emphasis on data and enterprise networks.
Two subsidiary topics are also covered: 1. the effects of electromagnetic
interference EMI. and their mitigation and 2. grounding, bonding, and
shielding. The chapter leans heavily on information provided by EIArTIA568A, Commercial Building Telecommunications Cabling Standard Ref. 2..
Issues dealing with survivability and improved availability are flagged
where appropriate.
The following is a list of common services and systems that should be
considered in the design of a commercial building cabling system:
1. Voice communications and associated PABX
2. Data communications including LANs and WANs
527
528
BUILDING AND CAMPUS WIRING AND CABLING FOR DATA COMMUNICATIONS
3. Other building signaling systems

Common abbreviations: UTP, unshielded twisted pair; STP, shielded twisted
pair.
14.2 MAJOR ELEMENTS OF TELECOMMUNICATION
BUILDING LAYOUT
Figure 14.1 gives an overview of a campus infrastructure consisting of four
buildings that are interconnected, showing the cabling system indicating its
major components. We discuss, in separate sections, the several system
architectural components that include horizontal cabling, backbone cabling,
work areas previously called work stations., telecommunication closets,
equipment rooms, and entrance facilities.
14.3
HORIZONTAL CABLING
Horizontal cabling connects work areas to the horizontal cross-connects. in a

telecommunication closet TC.. Consider that a work area has a telephone on
an employees desk connected by four-pair UTP and a workstation PC
connected separately . to cross-connects in the telecommunication closet by
two-pair STP or four pair UTP via wall connectors in each case. Of course,
cross-connects allow access to circuits for testing and ease of rearrangements.
EIArTIA recommends that distances from wall connectors to a TC should
be no greater than 90 m or 295 ft.
In metallic cable installations, the proximity of horizontal cabling to
electrical facilities that generate a high level of electromagnetic interference
EMI. should be taken into account. Motors and transformers, copiers, and
PCs are typical examples of these sources.
The topology of horizontal cabling is of the star type. Each work area is
served by a telecommunication closet on the same floor and should be
connected to a horizontal cross-connect installed in that closet. Bridged taps
and splices should not be permitted as part of the copper horizontal cabling.
It is recommended that a length limitation be enforced of no more than
6 m 20 ft. for cross-connect jumpers and patch cords in cross-connection
facilities.
Three types of transmission media are recommended for use in the
horizontal cabling system. These are:
a. Four-pair 100- UTP cables
b. Two-pair 150- STP-A cables
c. Two-fiber, 62.5r125- m optical fiber cable
The use of 50- cable should be deprecated.
HORIZONTAL CABLING
529
Figure 14.1. Campus wiringrcabling, interbuilding backbone cabling. (From Figure 1-1,
EIArTIA-568A, Ref. 2.)
14.3.1
Selection of Media
An individual work area should be provided two telecommunication

outletrconnectors. One outletrconnector is associated with voice telephony,
and the other is associated with data connectivity. The voice telephony circuit
should be supported by a four-pair 100- UTP cable, category 3 or higher.
The several categories of UTP cable are described in Section 14.7.1. The data
530
circuit should be supported by one of the following:
Four-pair 100- UTP cable, category 5

Two-pair 150- STP-A cable
Two-fiber, 62.5r125- m optical fiber cable
14.4
BACKBONE CABLING
Backbone cabling provides interconnection among telecommunication closets,

equipment rooms, and entrance facilities in the basic telecommunication
cabling structure shown in Figure 14.1. Backbone cabling also includes
intermediate and main cross-connects, mechanical terminations and patch
cords or jumpers used for backbone-to-backbone cross-connection. The
cabling among buildings in a campus environment is also included in backbone cabling.
14.4.1
Topology of Backbone Cabling
As shown in Figure 14.2, backbone cabling is based on a hierarchical star

topology. Here each horizontal cross-connect in a telecommunication closet
is cabled to a main cross-connect or an intermediate cross-connect and is
then cabled to a main cross-connect except for certain cabling between
telecommunication closets.. The hierarchy consists of no more than two
levels of cross-connects. From the horizontal cross-connect, no more than
one cross-connect should be passed through to reach the main cross-connect.
Therefore, interconnects between any two horizontal cross-connects should
pass through three or fewer cross-connects. Only a single cross-connect
should be passed through to reach the main cross-connect. This concept is
illustrated in Figure 14.2.
In certain situations it may be convenient to cable directly between
telecommunication closets. This may hold for bus or ring type LAN installations. Such cabling is in addition to the connections for the basic star
topology described in the preceeding text.
14.4.2
Selecting Cable Media
Transmission cable media should be selected from among the following:

a.
b.
c.
d.
100- UTP cable

150- STP-A cable
62.5r125- m optical fiber cable
single-mode optical fiber cable
BACKBONE CABLING
531
Figure 14.2. Backbone hierarchical star topology. (From Figure 5-1, EIArTIA-568A, Ref. 2.)
EIArTIA-568A sees future use of coaxial cable as deprecated. There may

be a notable exception if CATV broadband last-mile communication capability is desired. . Table 14.1 gives recommended cable types versus spectral
bandwidth versus maximum run distance.
The 90-m distances in Table 14.1 assume that 5 m 16 ft. is needed at each
end for jumpers connecting to the backbone.
Figure 5.2 of EIArTIA-568-A for voice telephony and lower-speed data
suggests greater distances than those shown in Table 14.1for example,
532
TABLE 14.1
Intra- and Interbuilding Distances
Cable Type
Category 3 UTP multipair
150- STP-A
Spectral Bandwidth
Maximum Distance
516 MHz
1020 MHz
20100 MHz
20300 MHz
90 m (295 ft)
90 m (295 ft)
90 m (295 ft)
90 m (295 ft)
between the horizontal cross-connect and intermediate cross-connect 500-m

1640-ft. maximum with UTP, STP-A, or fiber-optic cable; and the total
distance from horizontal cross-connect to main cross-connect, no more than
800 m 2624 ft..
14.5
TELECOMMUNICATION CLOSETS
The principal function of a telecommunication closet is to terminate horizontal cable distribution. It also is the location for the termination of backbone
cable. The cross-connection of horizontal and backbone cable terminations
using jumpers or patch cords allows flexible connectivity when extending
various services to telecommunication outletrconnectors. Connecting hardware, jumpers, and patch cords used for this purpose are collectively referred
to as horizontal cross-connect.
A telecommunication closet may also contain the intermediate cross-connect or the main cross-connect for different portions of the backbone cabling
system. Sometimes backbone-to-backbone cross-connections in the telecommunications closet are used to tie different telecommunication closets together in a ring, bus, or tree configuration as discussed above.
A telecommunication closet also provides a controlled environment to
house telecommunication equipment, connecting hardware, and splice closures serving a portion of the building.
14.5.1
Cabling Practices
Certain precautions should be observed such as the elimination of cable

stress caused by cable tension. Cables should not be routed in tightly cinched
bundles. Appropriate cable routing and dressing fixtures should be used for
effective organization and management of the different types of cables in
telecommunication closets. Horizontal and backbone building cables should
be terminated on connecting hardware that meets the requirements of
EIArTIA-568-A. These cable terminations should not be used to administer
cabling system moves, adds, and changes. All connections between horizontal
and backbone cables should be made through a horizontal cross-connect.
100- UNSHIELDED TWISTED-PAIR (UTP ) CABLING SYSTEMS
533
Equipment cables that consolidate several ports on a single connector

should be terminated on dedicated connecting hardware. Equipment cables
that extend a single port appearance may be either permanently terminated
or interconnected directly to horizontal or backbone terminations. Direct
interconnections reduce the number of connections in a link but may also
reduce flexibility.
Equipment rooms are considered distinct from telecommunication closets
because of the nature and complexity of the equipment they contain. Any or
all of the functions of a telecommunication closet may alternatively be
provided by an equipment room. Such a room provides a controlled environment to house telecommunication equipment, connecting hardware, splice
closures, grounding and bonding facilities, and protection apparatus where
applicable. From a cabling perspective, an equipment room contains either
the main cross-connect or the intermediate cross-connect used in the backbone cabling hierarchy.
14.6
ENTRANCE FACILITIES
An entrance facility serves as the entry points for cables, connecting hardware, protection devices, and other equipment needed to connect the outside
plant facilities to the premises cabling. Such components may be used for
public network services, private network customer premises services, or both.
The demarcation point between regulated carriersrservice providers and
customer premise cabling may be part of the entrance facility. The entrance
facilities include connections between cabling used in the outside environment and cabling authorized for in-building distribution. The connections
may be accomplished via splices or other means.
14.7
14.7.1
100- UNSHIELDED TWISTED-PAIR (UTP) CABLING SYSTEMS

UTP Category Definitions
The recognized categories of UTP cabling are as follows:

a. Category 3. This designation
ated connecting hardware
specified up to 16 MHz.
b. Category 4. This designation
c. Category 5. This designation
applies to 100- UTP cables and associwhose transmission characteristics are

534
14.7.2
Horizontal UTP Cable
The transmission characteristics discussed here apply to cables consisting of

four UTP of 24 AWG thermoplastic insulated solid conductors and enclosed
by a thermoplastic jacket. Four-pair, 22 AWG cables that meet or exceed
these requirements may also be used. The diameter of the insulated conductor should be 1.22-mm 0.048-in.. maximum. The pair twist length normally
are selected by the cable manufacturer to comply with the crosstalk requirements of EIArTIA-568-A.
The color codes of this cable are as shown in Table 14.2.
The diameter of the completed cable should be less than 6.35 mm
0.25 in...
14.7.2.1 Specific Transmission Characteristics. See Table 14.3.
Attenuation. See Table 14.5.
Near End Crosstalk (NEXT) Loss. See Table 14.6.
Propagation Delay. At 10 MHz, not to exceed 5.7 nsrm.
TABLE 14.2
Color Codes for Horizontal 100- UTP Cable
Conductor Identification
TABLE 14.3
Color Code
Abbreviation
Pair 1
White-blue, Note 1;
Blue, Note 2
(W-BL) (BL)
Pair 2
White-orange, Note 1;
Orange, Note 2
(W-O) (O)
Pair 3
White-green, Note 1;
Green, Note 2
(W-G) (G)
Pair 4
White-brown, Note 1;
Brown, Note 2
(W-BR) (BR)
Transmission Characteristics
Parameter
DC resistance
DC resistance unbalance
Mutual capacitance at 1 kHz
Value
9.38 per 100 m (328 ft)
No more than 5% between any two conductors
Not to exceed 6.6 nFr100 m (328 ft) for Category 3;
5.6 nFr100 m for Category 4 and Category 5
At 1 kHz any pair, not to exceed 330
pFr100 m (328 ft)
Capacitance unbalance: Pair to

ground
Characteristic impedance from 1 MHz 100 " 15%
to highest reference frequency
Structural return loss (SRL)
See Table 14.4
Source: Based on Ref. 2.
BACKBONE UTP CABLING
TABLE 14.4
535
Horizontal UTP Structural Return Loss (Worst Pair)
Frequency
(MHz)
Category 3
(dB)
Category 4
(dB)
Category 5
(dB)
110
1016
1620
20100
12
12 y 10 log( fr10)
21
21 y 10 log( fr10)
21 y 10 log( fr10)
23
23
23
23 y 10 log( fr20)
Source: From Table 10-2, Ref. 2.
TABLE 14.5
Horizontal UTP Cable Attenuation (per 100 m [328 ft] at 20C)
Frequency
(MHz)
Category 3
(dB)
Category 4
(dB)
Category 5
(dB)
0.064
0.256
0.512
0.772
1.0
4.0
8.0
10.0
16.0
20.0
25.0
31.25
62.5
100.0
0.9
1.3
1.8
2.2
2.6
5.6
8.5
9.7
13.1
0.8
1.1
1.5
1.9
2.2
4.3
6.2
6.9
8.9
10.0
0.8
1.1
1.5
1.8
2.0
4.1
5.8
6.5
8.2
9.3
10.4
11.7
17.0
22.0
14.8
Backbone UTP cabling consists of cables in pair sizes greater than four pairs
made up of 24 AWG thermoplastic insulated copper conductors that are
formed into one or more units of unshielded twisted pairs. The units are
assembled into binder groups of 25 pairs or part thereof following standard
industry color code. It should be noted that multipair 22 AWG cables that
meet the transmission requirements listed below are also acceptable for
backbone UTP application. The diameter over the insulation will be 1.22 mm
0.048 in.. maximum.
14.8.1
Transmission Performance Requirements
See Tables 14.3, 14.4, 14.5, and 14.6.
536
TABLE 14.6 Horizontal UTP Cable NEXT Loss

(Worst Pair Combination) [100 m (328 ft)]
Frequency
(MHz)
Category 3
(dB)
Category 4
(dB)
Category 5
(dB)
0.150
0.772
1.0
4.0
8.0
10.0
16.0
20.0
25.0
31.25
62.5
100.0
53
43
41
32
27
26
23
68
58
56
47
42
41
38
36
74
64
62
53
48
47
44
42
41
39
35
32
Note: 0.150 MHz is for reference purposes only.

Source: From Table 10.5, Ref. 2.
Note on Crosstalk Interference. In multipair cable, a given pair receives

crosstalk interference from other energized pairs sharing the same sheath.
The total crosstalk energy that a pair receives is specified as the power sum
crosstalk. The power sum crosstalk for uncorrelated disturbing pairs can be
calculated from the individual pair-to-pair crosstalk measurements at a given
frequency. Generally, the power sum crosstalk energy is dominated by the
couplings between pairs in close proximity and is relatively unaffected by
pairs in separate binder groups. Therefore, it is desirable to separate services
with different signal levels or services that are susceptible to impulse noise
into separate binder groups.
The propagation delay of any pair at 10 MHz should not exceed
5.7 nsrm.
14.8.2
Connecting Hardware for UTP Cable
Under this heading we include telecommunication outletrconnectors, patch

panels, transition connectors, and cross-connect blocks. It should be noted
that requirements for connector categories 3, 4, and 5 are not sufficient in
themselves to ensure required cabling system performance. Link performance also depends on cable characteristics including cross-connect jumpers
and patch cords., the total number of connections, and the care with which
they are installed and maintained.
It is desirable that hardware used to terminate UTP cables be of the
insulation displacement contact IDC. type.
537
Connection hardware for the 100- UTP cabling system is installed at the
following:
a.
b.
c.
d.
e.
Main cross-connect
Intermediate cross-connect
Horizontal cross-connect
Horizontal cabling transition points
Telecommunication outletrconnectors
Typical cross-connect facilities consist of cross-connect jumpers or patch

cords and terminal blocks or patch panels that are connected directly to
horizontal or backbone cabling.
Attenuation. Worst-case attenuation of any pair within a connector should
not exceed the values listed in Table 14.7 at each specified frequency for a
given performance category.
NEXT Loss. Table 14.8 shows the NEXT loss for the connecting hardware
for the three categories of interest.
Return Loss. Connector return loss is a measure of the degree of impedance
matching between the UTP cable and connector and is derived from swept
frequency voltage measurements on short lengths of 100- twisted-pair test
leads before and after inserting the connector under test.
Because the return loss characteristics of category 3 connecting hardware
are not considered to have a significant effect on the link performance of
category 3 UTP cabling, return loss requirements are not specified for
category 3 connectors.
TABLE 14.7 Attenuation of Connecting Hardware Used for

100- UTP Cable
Frequency
(MHz)
Category 3
(dB)
Category 4
(dB)
Category 5
(dB)
1.0
4.0
8.0
10.0
16.0
20.0
25.0
31.25
62.5
100.0
0.4
0.4
0.4
0.4
0.4
0.1
0.1
0.1
0.1
0.2
0.2
0.1
0.1
0.1
0.1
0.2
0.2
0.2
0.2
0.3
0.4
Source: Table 10-8, Ref. 2.
538
TABLE 14.8
UTP Connecting Hardware NEXT Loss
Frequency
(MHz)
Category 3
(dB)
Category 4
(dB)
Category 5
(dB)
1.0
4.0
8.0
10.0
16.0
20.0
25.0
31.25
62.5
100.0
58
46
40
38
34
65
58
52
50
46
44
65
65
62
60
56
54
52
50
44
40
DC Resistance. The DC resistance between the input and output connections of the connecting hardware not including the cable stub, if any. used
for 100- cabling should not exceed 0.3-.
14.9
150- SHIELDED TWISTED-PAIR CABLING SYSTEMS
This section covers the requirements for 150- STP-A inside cable for use in
the horizontal cabling system. The cable is restricted to two-pair size consisting of 22 AWG thermoplastic insulated solid conductors enclosed by a shield
and an overall thermoplastic jacket. The pair twist lengths are selected by the
manufacturer to ensure compliance with the crosstalk requirements of
EIArTIA-568-A. The pair 1 color code is red-green, and for pair 2 it is
orange-black.
14.9.1
Transmission Performance Requirements
DC Resistance. The resistance of any conductor should not exceed 5.71

r100 m 328 ft..
DC Resistance Unbalance. The resistance unbalance between the two conductors of any pair should not exceed 4%.
Capacitance Unbalance: Pair to Ground. The capacitance unbalance to
ground at 1 kHz of any pair should not exceed 100 pFr100 m 328 ft..
Balanced Mode Attenuation. The cable balanced mode attenuation is given
in Table 14.9.
Common Mode Attenuation. The common mode attenuation of any pair, at
or corrected to a temperature of 25 " 3C and measured in accordance with
OPTICAL FIBER CABLING SYSTEMS
539
TABLE 14.9 Cable Balanced Mode Attenuation

(Horizontal STP-A Cable)
Frequency
(MHz)
Maximum Attenuation
[dBr100 m (328 ft)]
0.0096
0.0384
4.0
8.0
10
16
20
25
31.25 a
62.5
100
300
0.30
0.50
2.2
3.1
3.6
4.4
4.9
6.2
6.9
9.8
12.3
21.4
Value not clear in reference

ASTM D 4566, shall not exceed:

95.0
f
50
dBrkm,
for all frequencies from 50 MHz to 600 MHz,
where f s frequency in MHz. The attenuation measurements from 50 MHz

to 600 MHz shall be performed on cable lengths of 100 m 328 ft. to 305 m
1000 ft..
NEXT Loss. The NEXT loss between the two pairs within a cable should not
exceed:
q 58.0 dB at 9.6 kHz terminated in 270
q58.0 dB at 38.4 kHz terminated in 185
q58.0 dB for all frequencies between 0.1 MHz and 5 MHz terminated in
150 and
q58.0 y 15 log fr5. dB for all frequencies from 5 MHz to 300 MHz
terminated in 150 where f is expressed in MHz.
Propagation Delay. The propagation delay of any pair at 10 MHz should not
exceed 5.7 nsrm.
14.10
OPTICAL FIBER CABLING SYSTEMS
62.5r125- m optical fiber cable is recommended for horizontal cable, and

both 62.5r125- m and single-mode optical fiber cables may be used for
backbone cables.
540
For the 62.5r125- m cable, at 850-nm wavelength, budget 3.75 dBrkm

loss and a minimum information capacity of 160 MHz-km; at 1300 nm,
budget 1.5 dBrkm loss and a minimum information transmission capacity of
500 MHz-km.
14.10.1
Backbone Optical Fiber Cable
The optical fiber construction consists of 62.5r125- m optical fibers or

single-mode optical fibers, or both, typically formed into groups of 6 or 12
fibers each. These groups and individual fibers should be identifiable in
accordance with EIArTIA-598. These groups are assembled to form a single
compact core, and the core is covered by a protective sheath. The sheath
consists of an overall jacket and may contain an underlying metallic shield
and one or more layers of dielectric material applied over the core.
Customer premises optical fiber backbone cabling has been and continues
to be primarily 62.5r125- m multimode fiber based because this fiber can
use LED transmitters. With the rapidly growing bit rate requirements,
consideration should be given to installing single-mode optical fiber in
addition to the multimode fiber. Single-mode fiber systems inherently have
higher bit rate capacities and longer distance capabilities than the conventional multimode optical fiber.
14.11 GROUNDING AND BONDING INFRASTRUCTURE FOR
COMMERCIAL BUILDINGS
14.11.1
Rationale
A well-designed and installed grounding and bonding infrastructure in a

commercial building in support of telecommunications can pay off in two
ways:
It can be a major step in assuring electrical efficiency.

It can negate or mitigate the effects of EMI on the telecommunication
system as a whole.
These goals can be accomplished by providing a ground reference for

telecommunication systems within the telecommunication entrance facility,
telecommunication closets., and equipment rooms.. To further these objectives, the bonding and connecting of pathways, cable shields, conductors, and
hardware at telecommunication closets, equipment rooms, and entrance
facilities will be required.
Figure 14.3 shows a typical layout for bonding and grounding of fairly
large, multistory commercial buildings with multiple backbones. The telecommunications grounding and bonding infrastructure originates with a connection to the service equipment power. ground and extends throughout the
GROUNDING AND BONDING INFRASTRUCTURE FOR COMMERCIAL BUILDINGS
541
Figure 14.3. Bonding and grounding layout for a large commercial building. (From Figure
2.1-1, EIArTIA-607, Ref. 3.)
542
Figure 14.4. Connectivity to the service equipment (power) ground. (From Figure 5.2-1,
EIArTIA-607, Ref. 3.)
building. It comprises five major components:

a.
b.
c.
d.
e.
Bonding conductor for telecommunications

Telecommunications main grounding busbar TMGB.
Telecommunications bonding backbone TBB.
Telecommunications grounding busbar TGB.
Telecommunications bonding backbone interconnecting bonding conductor TBBIBC.
All bonding conductors should be insulated and of copper. Bonding conductors should not be placed in ferrous metallic conduit. Bonding conductors
should be marked green in color.
The bonding conductor for telecommunications bonds the TMGB to the
service equipment power. ground as shown in Figure 14.4. As a minimum,
the bonding conductor should be the same size as the TBB.
14.11.2
The Telecommunications Bonding Backbone (TBB)
A TBB is a conductor that interconnects all TGBs with the TMGB. A TBBs
basic function is to reduce or equalize potential differences between telecommunication systems bonded to it. A TBB is not intended to serve as the only
conductor providing a ground fault return path.
A TBBs. originates at the TMGB, extends throughout the building using
the telecommunication backbone pathways, and connects to the TGBs. in all
telecommunication closets and equipment rooms. The interior water piping
system of the building should not be used as a TBB, nor should the metallic
cable shield.
543
Bonding and Sizing the TBB. A TBB is an insulated copper conductor. Its
minimum conductor size is No. 6 AWG; a larger diameter is preferable, even
3r0 AWG.
Whenever two or more vertical TBBs are used within a multistory building, the TBBs are bonded together with a TBB interconnecting bonding
conductor TBBIBC. at the top floor and at a minimum of every third floor
in between. A TBB may be spliced, provided that all applicable requirements
of Section 5 of EIArTIA-607 are met. The TBB is connected to the TMGB
as specified in Section 5.4.7.1 of EIArTIA-607.
The TBB conductors are installed without splices, where practicable. If
splices are necessary, there should be a minimum number which are accessible and located in telecommunication spaces. Joined segments of a TBB
should be connected using irreversible compression-type connectors, exothermic welding, or the equivalent.
14.11.3
The Telecommunication Main Grounding Busbar (TMGB)
The TMGB serves as a dedicated extension of the building grounding

electrode system for the telecommunications infrastructure. The TMGB also
serves as the central attachment point for telecommunication bonding backbones TBB. and equipment, and it is located such that it is accessible to
telecommunication personnel. Extensions of the TMGB i.e., other busbars .
are telecommunication grounding busbars TGB.. Typically, there should be
a single TMGB per building.
The ideal location of the TMGB is in the telecommunication entrance
room or space. However, the TMGB should be located such as to minimize
the length of the bonding conductor for telecommunication.
TMGB Description. The TMGB is a predrilled copper busbar with standard
NEMA bolt hole sizing and spacing for the type of connectors to be used. It
should be sized in accordance with the immediate requirements of the
application and with consideration of future growth. Its minimum dimensions
are 6 mm thick = 100 mm wide and variable in length. For reduced contact
resistance, the busbar should be electrotin-plated.
Connections to the TMGB. The connections to the bonding conductor for
telecommunications and the TBBs to the TMGB should use listed 2-hole
compression connectors, exothermic-type welded connections, or the equivalent. The connection of conductors for bonding telecommunication equipment to the TMGB may use one-hole lugs or equivalent; however, 2-hole
compression connectors are preferable.
All metallic raceways for telecommunication cabling located within the
same room or space as the TMBG should be bonded to the TMGB.
A practical location for the TMGB is to the side of the panelboard where
provided.. The vertical location of the TMGB should take into consideration
544
whether the grounding and bonding conductors are routed in an access floor
or overhead cable tray.
14.11.4
The Telecommunications Grounding Busbar (TGB)
The TGB is the common central point of grounding connection for telecommunication systems and equipment in the location served by that telecommunication closet or equipment.
Description of the TGB. The TGB consists of a predrilled copper busbar
provided with standard NEMA bolt hole sizing and spacing for the type of
connectors to be used. It will have the minimum dimensions of 6 mm
thick = 50 mm wide and vary in length to meet the applications requirement
with consideration for future growth.
The TGB should be electrotin-plated for reduced contact resistance. If not
plated, the busbar should be cleaned prior to fastening the conductors to the
busbar.
Bonds to the TGB. TBBs and other TGBs within the same space are bonded
to the TGB by means of at least a No. 6 AWG conductor, preferably with a
3r0 AWG conductor. The bonding conductor between a TBB and a TGB
should be continuous and routed in the shortest possible straight-line path.
Where a panelboard for telecommunications is located within the same
room or space as the TGB, the panelboards ACEG bus when equipped. or
the enclosure should be bonded to the TGB. The TGB should be as close to
the panelboard as practicable and should be installed to maintain clearances
required by applicable electrical codes.
Where a panelboard for telecommunications is not located within the
same room or space as the TGB, consideration should be given to bonding
the panelboards ACEG bus when equipped. or the enclosure to the TGB.
The TGB should be bonded to the TBBIBC as required. All metallic
raceways for telecommunication cabling located within the same room or
space as the TGB should be bonded to the TGB.
Connections of TBBs to the TGB should use listed 2-hole compression
connectors.
A practical location for the TGB is to the side of the panelboard where
provided.. The vertical location of the TGB should take into consideration
whether the grounding and bonding conductors are routed in an access floor
or overhead cable tray.
14.11.5
Bonding to the Metal Frame of a Building
All bonding conductors and connectors for bonding the metal frame of a
building will be listed for the purpose intended and approved by an NRTL.
In buildings where metal frames structural steel. are effectively grounded,
545
each TGB should be bonded to the metal frame within the room using a No.
6 AWG conductor. Where the metal frame is external to the room and
readily accessible, the metal frame should be bonded to the TGB and TMGB
separately with a No. 6 AWG conductor.
When practical because of shorter distances and other considerations, and
where horizontal steel members are permanently bonded to vertical column
members, TGBs may be bonded to these horizontal column members in lieu
of vertical column members.
14.11.6
Telecommunications Entrance Facility (TEF)
The TEF includes the entrance point where telecommunications services

enter. There is generally a room or other space devoted to this. It is also
where the joining of inter- and intrabuilding backbone facilities takes place
and where the proper grounding and bonding of these facilities is accomplished. The TEF may also contain antenna entrances and electronic equipment serving telecommunication functions.
The TEF is the desirable location for the TMGB. The TMGB may serve
as the TGB for collocated equipment in the TEF as appropriate.
14.11.6.1 Placement of the TMGB. Placement of the TMBG should be
such that it results in the straightest route considering the total length of the
bonding conductor from the telecommunication primary protectors to the
TMGB. This bonding conductor is intended to conduct lightning and ac fault
currents from the telecommunication primary protectors. At least a 300-mm
1-ft. separation should be maintained between this insulated conductor and
any dc power cables, switchboard cable, or high-frequency cables, even when
isolated by metallic conduit or EMT electrical metallic tubing..
The TMGB is the common point in the TEF to which grounding connections for that room are made. It should be as close to the panelboard for
telecommunications as practicable and should be installed to maintain clearances required by applicable electrical codes. When a panelboard for
telecommunications is not installed in the TEF, the TMGB should be located
near the backbone cabling and associated terminations. In addition, the
TMGB should be placed such that the bonding conductor for telecommunications is as short and as straight as possible.
When the TEF pathway incorporates an isolation gap, the pathway on the
building side of the gap should be bonded to the TMGB. In buildings where
the backbone cable incorporates a shield or metallic member, this shield or
metallic member should be bonded to the TMGBrTGB.
The TMGB is intended to be the location for connecting grounding bars
of telecommunication equipment located in the TEF. This typically may be
multiplex equipment e.g., DS1 channel banks. or optical fiber termination
equipment.
546
14.11.7
Telecommunication Closets and Equipment Rooms
Each telecommunication closet and equipment room should contain a TGB.

It should be located inside the closetrroom and be insulated from its
support. A 50-mm 2-in.. separation is recommended. Multiple TGBs may be
installed within the same closet to minimize conductor lengths and terminating space. In all cases, multiple TGBs within a closet should be bonded
together with at least a No. 6 AWG conductor.
Where a panelboard for telecommunications is not installed in the
telecommunications closet, the TGB should be placed near the backbone
cabling and associated terminations. In addition, the TGB should be located
so that the grounding conductors are as short and straight as possible.
14.12 CUSTOMER-OWNED OUTSIDE PLANT (OSP)

INFRASTRUCTURE
14.12.1
OSP Cabling Infrastructure Defined
The outside plant cabling infrastructure provides interconnections between

building entrance facilities, structures on a campus or other multibuilding
environment, or telecommunication pedestals or cabinets. The customerowned OSP consists of backbone cables, splices, terminations, and patchcords or jumpers used for backbone-to-backbone interconnection.
14.12.2
OSP Topology
A star configuration is the recommended topology for customer-owned OSP

cabling infrastructure. Figure 14.5 illustrates a campus with a star backbone
topology. In the example, building A is the center or hub of the star with
backbone cables extending to other campus buildings. B, C, D, E, F. and
an outdoor telecommunication pedestal G.. The example also shows an
optical fiber backbone cable passing from building A to building F
through an intermediate building E..
In the example, Figure 14.5, building A provides a point of service for a
microwave communication link to a second campus. The backbone cables can
be used for distributing these applications from A to all or to selected
buildings. Of course, A was selected only as an example.
Campus telecommunication applications require use of both intrabuilding
and interbuilding backbone cabling. Figure 14.6 shows the relationship between the campus star backbone and the intrabuilding backbones of building
E. This illustrates the intrabuilding cabling topology from an individual
work area through the intrabuilding backbone cabling to the campus backbone main interconnect facility in building A.
A simple star topology is not always feasible, particularly on very large
campuses. The distances between buildings may exceed the maximum allow-
CUSTOMER-OWNED OUTSIDE PLANT (OSP) INFRASTRUCTURE
547
Figure 14.5. An example of a campus star topology showing outside plant installation. (From
Figure 3, TIArEIA-758, Ref. 4.)
able cable lengths. This problem can often be solved by turning to a

hierarchical star topology. Each campus segment may connect to a hub
location that would support the area as a simple star topology. These hub
locations may be connected with other topologies to support equipment
and technologies normally used for wide area applications. Here SONET
over cable or LOS microwave may answer the interconnectivity issues.
548
Figure 14.6. Example of campusrbuilding cable topology. (From Figure 4, EIArTIA-758,

Ref. 4.)
14.12.3
Recognized Cable Media and Media Selection
Customer-owned OSP cabling must support a wide range of services and site
sizes. Therefore, more than one transmission medium is recognized. This
standard specifies recognized transmission media that may be used individually or in combination. The recognized media include:
50r125- m optical fiber cable

62.5r125- m optical fiber cable
REFERENCES
549
Single-mode optical fiber cable

100- twisted-pair cable
75- coaxial
The specific performance characteristics for recognized cables, associated

connecting hardware, cross-connect jumpers, and patch cords are specified in
this text.
Customer-owned OSP cabling specified by EIArTIA-758 Ref. 4. is applicable to a wide range of user requirements. Media choices must be made
depending upon the characteristics of the applications, and distance. Where
a single cable type may not satisfy all user requirements, it will be necessary
to use more than one media type in the OSP cabling. Where possible, the
different media should use the same physical pathway architecture and space
for connecting hardware. In making this choice, factors to be considered
include:
Flexibility with respect to supported services

Required useful life of backbone cabling
Site size and user population
REFERENCES
1. Fred J. McClimans, Communications Wiring and Interconnection, McGraw-Hill,
New York, 1992.
2. Commercial Building Telecommunications Cabling Standard, ANSIrEIArTIA-568A,
Telecommunication Industry Association, Washington DC, October 1995.
3. Commercial Building Grounding and Bonding Requirements for Telecommunications,
ANSIrEIArTIA-607, Telecommunication Industry Association, Washington DC,
August 1994.
4. Customer-Owned Outside Plant Telecommunications Cabling Standard,
ANSIrEIArTIA-758, Telecommunication Industry Association, Washington DC,
April 1999.
15
BROADBAND DATA
TRANSPORT TECHNIQUES
15.1
CHAPTER OBJECTIVE
Data is one of several binary media that is transmitted electrically. Voice and
video can also be represented in a binary format for electrical transmission.
The demand for bandwidth* to accommodate these media with ever-increasing transmission rates needed to be satisfied. The general theme is transmission over long distances. Long-distance may be just across the street or
around the world.
The chapter covers three methods of data transport:
Fiber optics, with particular emphasis here

Line-of-sight microwave
Satellite communications
The question arises: How broad is broad?

Line-of-sight LOS. microwave and satellite communications can provide
in the order of 500 MHz, the bandwidth being limited by international treaty.
Fiber-optics transmission can provide 75,000 gigahertz or more, being limited
by the laws of physics. Present technology permits 10-Gbps transmission on
one wavelength on a single fiber with 40-Gbps operation on the drawing
boards. Through shrewd bit-packing techniques, LOS radio can achieve 655
Mbps on a single radio frequency RF. carrier; satellite communications has
the capability to achieve the same bit rate.
It also became apparent that the DS1rE1 digital hierarchy described in
Chapter 7 had transmission rate limitations and lacked flexibility. We can
define format as a way of organizing the bits in a serial bit stream so that the
resulting representation will be meaningful to a digital receiver.
*A better expression might be bit rate capacity.
551
552
BROADBAND DATA TRANSPORT TECHNIQUES
To overcome the bit rate limitations of the DS1rE1 family of the

plesiochronous digital hierarchy PDH., two new format structures were developed:
Synchronous optical network SONET.

Synchronous digital hierarchy SDH.
The former was developed and is widely deployed in North America, whereas
the latter is a European development and is deployed to some lesser degree
in Europe and in those countries of European hegemony.
SONET was developed in the United States, and its standards are issued
by Telcordia and ANSI. SDH is of European origin and its standards are
issued by the ITU-T Organization and by ETSI European Telecommunication Standardization Institute .. The second part of this chapter covers these
advanced digital multiplexing schemes.
We hasten to add that there is an implication in terminology that these
two formats are only compatible with optical fiber transmission. This statement is patently untrue. Given the necessary operational bandwidth, SONET
and SDH operate with LOS microwave and satellite communications just as
well as with optical fiber. With a 40-MHz bandwidth assignment, the 622Mbps rates may be transported. This rate is denominated STS-12 by SONET
and STM-4 by SDH. Even higher rates can be accommodated in the higher
frequency bands such as 28 and 40 GHz. However, the length of these links is
severely limited by rainfall considerations. Typical link lengths may be in the
range of 35 km depending on the desired time availability Ref. 1..
Some fiber-optics applications implement links from less than 1 m to 10 m
or more. We should expect to find these extremely short links inside digital
switches, for example. On the other side of the coin, nearly all transoceanic
communication is done by optical fiber today. It is the medium of choice.
Such transport systems as discussed in this chapter carry immense amounts
of data. The data are in the formats discussed in previous chapters such as IP
and frame relay. They may also have an ATM format, which is covered in
Chapter 16.
15.2
15.2.1
INTRODUCTION TO FIBER-OPTICS TRANSMISSION

What Is So Good About Fiber-Optic Transmission?
An optical fiber is a circular waveguide of tiny dimensions, about the

diameter of a human hair. We generally think of optical fiber being made of
glass. Most fiber is made of glass; however, other materials such as plastic
may also be used. There are also new advances in halides and bromides. It is
expected that once these materials appear on the scene in the form of optical
553
waveguide, we should expect to see dramatic improvements in the performance of transmitting data. For example, early tests have shown reductions
in loss per unit length of at least one order of magnitude, possibly two
Ref. 2..
I tell my students in seminars that optical fiber has nearly an infinite
bandwidth. If we consider the operational band between 1550 and 1635 nm,
the available bandwidth is around 10 THz teraHz. or 10,000 GHz. The
usable radio-frequency spectrum is from about 10 kHz to 100 GHz, or,
rounded off, 100 GHz of bandwidth for the total spectrum. So we can fit
about 100 total RF spectrums in the 1550 nm band. Thats pretty close to
infinity in my book.
Another point is the optical fiber needs no equalization as coaxial cable
does.
When we discuss light, we will be using wavelength rather than frequency
to describe a certain emission location in the frequency spectrum. Chronologically, as fiber development proceeded, we started with the shorter wavelengths of around 820 nm 0.820 m. and then progressively moved to longer
wavelengths of around 1330 nm 1.330 m. and 1550 nm 1.550 m.. We can
convert frequency F in Hz. to wavelength in meters., and vice versa, with
the following formula:
F s 3 = 10 8 mrs
Remember, F is measured in Hz and is measured in meters. For example,
1550 nm s 1550 = 10y9 m. If you wish to convert this to frequency, which
you may be a bit more familiar with, use this formula. For example, using
1550-nm wavelength:
FHz s 3 = 10 8r1550 = 10y9
s 1.93548 = 10 14 Hz or 193.548 = 10 12 Hz s 193.548 THz
15.2.2
Advantages of Optical Fiber
Light and Small. Fiber cable is very light and small in size. Over 200 fibers
can be placed in 0.5-in. cable. Compare this to copper pair or coaxial tube
cables. Another example is the wiring in a large U.S. aircraft carrier. It is
estimated that if all signal leads were changed over from copper to fiber, a
200-ton saving of topside weight would accrue.
Repeater Spacing. One of the greatest impairments to digital transmission is
jitter. Jitter accumulation is a function of the number of regenerative repeaters there are in tandem on a certain connectivity. Fiber-optic systems
require about 1r100th the number of repeaters compared to comparable
coaxial cable systems.
554
Electromagnetic Compatibility (EMC). Fiber-optic systems do not radiate RF

energy and in themselves are immune from radiation from other sources. For
this reason, power companies run aerial fiber cable to support their telecommunication systems on the same tower line as their high-tension transmission
lines. Fiber is immune to the rampant electrical noise of high-tension lines.
Crosstalk. On metallic cable systems, crosstalk can be a limiting impairment.
There is no possibility of crosstalk on fiber-optic cable systems because there
is no optical coupling from one fiber to another.
Temperature and Humidity. Fiber-optic systems are relatively unaffected by
temperature and humidity extremes, unlike their metallic counterparts. This
may not be the case for active light components, which, however, can be kept
in a protected environment such as in repeater casings.
Cost. The cost of fiber-optic cable and related electronics has been eroding
considerably. In many circumstances, fiber-optic cable is much more economic than equivalent coaxial cable, especially considering related electronics e.g., there are fewer repeaters per unit distance for fiber cable..
Frequency Allocation and Licensing. Unlike microwave and satellite radio
systems, fiber-optic cable systems and metallic cable systems. require no
frequency allocations, nor must they go through extensive and expensive
licensing procedures.
15.2.3
Overview of an Optical Fiber Link
As illustrated in Figure 15.1, a fiber-optic link consists of a light source i.e.,

transmitter ., a fiber-optic cable with connectors andror splices, and a detector i.e., receiver.. Long fiber-optic links require intermediate repeaters that
Figure 15.1. A conceptual block diagram of a fiber-optic link.
555
are regenerative in the case of digital transmission.. The distance between

repeaters is much greater than with coaxial cable systems, and very much
greater than when dealing with wire-pair systems. The distance is a function
of source and detector parameters, the bit rate and BER desired, and, of
course, fiber loss. That distance is greatly extended with the use of erbiumdoped amplifiers.
There are two basic impairments to fiber-optic transmission: fiber loss and
dispersion. Standard glass fiber experiences losses as low as 0.25 dB per
kilometer at the longer wavelengths. When monomode fiber is used, we
expected that dispersion could be neglected. This is true for the hundreds of
Mbps transmission range. However, at the higher bit rates e.g., ) 1 Gbps.,
chromatic dispersion becomes a factor.
15.2.4
Optical Fiber Transmission
15.2.4.1 Optical Fiber Composition. As shown in Figure 15.2, an optical

fiber is composed of a central cylindrical core of higher refractive index n1 .
and a concentric cladding of lower refractive index n 2 .. The most common
fibers for communication links have glass or fused silica core and cladding,
although plastic-clad glass or all-plastic construction is used on applications
which allow less bandwidth andror more attenuation. Glass fiber is usually
covered by an outer protective plastic coating which has no effect on light
propagation. The coating protects the glass fiber from mechanical damage
and from strength and transmission degradation due to exposure to air. It
protects the fiber and gives it pulling strength.
Optical Fiber as a Waveguide. Fibers act as tubular mirrors, with the
claddingcore interface confining light to the fiber core. Light coupled into
an optical fiber is prevented from escaping by inward reflection within
the core.
Snells Law. Snells law deals with the change of direction of light rays as
they pass from a medium of one refractive index to another. Understanding
of two terms is important to the discussion of refraction. The normal is the
Figure 15.2. Basic composition of an optical fiber.
556
Figure 15.3. Reflection, refraction, and critical angle. Note that the critical angle is where the
light ray just grazes the surface as shown in part B.
imaginary line perpendicular to the interface of the two materials; in the case
of fiber, it is perpendicular to the interface between the core and cladding.
The angle of incidence is the angle between the normal and the incident ray.
Referring to Figure 15.3 for the definition of angles 1 and 2 , Snells
law states
n1 sin 1 s n 2 sin 2
where n1 and n 2 represent media of higher and lower indices of refraction,
respectively.
The refracted ray entering the lower index area is bent away from the
normal, as shown in Figure 15.3A. At the critical angle, the refracted ray is
bent 90 from the normal and just grazes the interface 2 s 90., as
illustrated in Figure 15.3B. The critical angle of incidence can be described
by
sin c s n 2rn1 ,
n1 ) n 2
When the angle of incidence exceeds the critical angle, the light ray is totally
reflected at the interface, a characteristic called total internal reflection see
Figures 15.3C and 15.4.. At angles less than the critical angle, most of the
energy of the light ray escapes, as shown in Figure 15.3A. The discussion that
follows concerns only rays passing through the axis of the fiber meridional
rays.. Skew rays propagate without passing through the axis of the fiber and
are not dealt with here, due to the mathematical complexity of their behavior. For most purposes, it is sufficient to assume that skew rays will behave in
much the same way as meridional rays Ref. 2..
557
Figure 15.4. Total internal reflection within a fiber.
Numerical Aperture (NA). The angle of incidence for total reflection

depends on the relative values of the indices of refraction of the core and the
cladding. Total internal reflection takes place only for rays striking the
interface at angles greater than or equal to the critical angle. This gives rise
to the acceptance cone shown in Figure 15.5. It is described by the acceptance
angle A , which is the maximum angle with respect to the fiber axis at which
an entering ray will experience total internal reflection. The sine of the
acceptance cone half-angle is called the numerical aperture NA. and is a
measure of the light-gathering ability of the fiber. NA is directly related to
the refractive index of the core and cladding. Applying Snells law, and
assuming air refractive index of 1. as the external medium, the following
equation can be derived, showing that NA is a function of the refractive
Figure 15.5. Acceptance cone.
558
indices of the core and cladding in a step-index fiber:

NA s sin A s n12 y n22 .
1r2
Note: Rays of light are shown entering the center of the fiber in Figure
15.5. Similar light cones obeying the same analysis are incident at an infinite
number of points on the coreair interface. This does not affect the calculation of NA Ref. 3..
15.2.4.2 Distortion and Dispersion
Distortion and Dispersion Differentiated. Distortion is an inherent property
of all optical fibers. It causes pulse broadening in digital transmission as
shown in Figure 15.6. When pulses broaden to the point where they begin to
merge, intersymbol interference isi. begins to occur. The upper limit of
bandwidth-distance factor is a function of distortion.
Figure 15.6. Multimode distortion, pulse broadening, and intersymbol interference.
559
Why is there this broadening? As we can see in the upper drawing of

Figure 15.6, light launched into multimode fiber contains several transmission modes: low order and high order. We send a pulse down this pipe, and
it contains signal energy of the low-order mode and signal energy in the
high-order mode. As can be seen in the figure, the low-order mode has fewer
reflections as it travels down the pipe; thus to reach the distant receiver this
energy arrives first because it travels less distance than the energy in the
high-order mode. Such indeed is what causes pulse broadening. When there
are several pulses, then, broadened energy of the first pulse spills into the
second pulse position. This confuses the receivers decision circuit: Was it a 1
or a 0 that was received?
Distortion can be the determining factor limiting link length in very high
bit rate systems. The higher the bit rate on a system, the more dispersion and
distortion have to be controlled. Distortion is usually measured in picoseconds per kilometer.
DISTORTION VERSUS DISPERSION. We tend to use the terms distortion and
dispersion synonomously. Reference 3 MIL-HDBK-415. states that this can
cause a great deal of confusion. The IEEE dictionary calls distortion temporal and dispersion spectral Ref. 4..
DISCUSSION OF DISPERSION AND DISTORTION
Multimode Distortion. In multimode fibers, light can travel different paths

within the fiber core. These paths, or propagation modes, vary in length.
Because of this variance, light will exit a fiber over a time interval slightly
longer than that over which it entered. This phenomenon, known as multimode distortion, causes an undesirable broadening of the signal waveform as
it progresses along with fiber as shown in Figure 15.6. Multimode distortion
is most prominent in multimode step-index fibers, where it is the major form
of distortion. It is less evident in graded-index fibers and nonexistent in
single-mode fibers.
Dispersion. Within a given fiber, different wavelengths of light have different propagation characteristics. This variance with wavelength is called
dispersion and, like multimode distortion, can cause pulse broadening. Unlike
multimode distortion, dispersion affects all types of fiber, even single-mode.
Three types of dispersion fall within the scope of this chapter: material
dispersion, waveguide dispersion, and chromatic dispersion.
Material Dispersion. The speed of light through a fiber varies with refractive index, which itself is wavelength-dependent. Light having a wavelength at
which the refractive index is lower will travel faster than light at a wavelength
at which the refractive index is higher. Even the light source with the
narrowest spectral width emits over a finite band of wavelengths. As illustrated in Figure 15.7, choice of either a source with narrow spectral width or
560
Figure 15.7. Effect of source spectral width and fiber index variations on material dispersion.
of a fiber material with reduced variation of index with wavelength will lessen
the effects of material dispersion. Material dispersion is a major contributor
to total dispersion on fiber-optic systems using light-emitting diodes LEDs.
for the light source transmitter . because of the wide spectral width LEDs
display.
Waeguide Dispersion. Waveguide dispersion is attributed to the dependence of phase and group velocities on the geometric characteristics of the
waveguide. Because light wave peaks occur at different locations in the
waveguide with changing frequency, intensity differences occur when all
reflections are added together. This causes broadening, but waveguide dispersion usually has much less effect than material dispersion and can be
ignored in most cases.
Chromatic Dispersion. Chromatic dispersion affects only fiber-optic links
carrying very high rate digital systems e.g., ) 1 Gbps.. Such systems will use
laser diode sources with monomode fiber. A laser diode has a narrow
emission line width, perhaps several nanometers wide. With distributed
feedback lasers, this line width can be narrowed to 0.5-nm widths and less.
Even this narrow width carries many emission wavelengths. Some of these
wavelengths travel slightly faster than others. As the bit rate increases, the
pulse width decreases, as one might expect. We are interested in half-pulse
width where the detector will make a 1 or 0 decision. The durations of these
half-pulse widths start approaching the time difference of arrival of signal
energy giving rise to chromatic dispersion. Again it is the problem of the
slower signal energy spilling into the subsequent bit position confusing the
receiver.
Zero Dispersion Near 1300 Nanometers. In fused-silica single-mode fibers,
there is sometimes an operating wavelength at which material dispersion and
waveguide dispersion are equal in amplitude but opposite in sign. This results
561
in total dispersion approaching zero. The exact frequency at which this

occurs is a function of core diameter and refractive index profile. The range
extends from about 1300 nm for a 10- m core to about 1600 nm for a 4- m
core. Choice of materials and core diameter make it possible to bring the
wavelengths of minimum dispersion and minimum attenuation to coincide.
15.2.4.3 Attenuation. Attenuation absorption and scattering . of light in a
fiber is caused by material variations, microbends induced by mechanical
stress, and surface imperfections at the corecladding interface.
Absorption and Scattering. Absorption, whereby light is converted to heat in
the fiber core, and scattering, whereby energy is ejected from the core into
surrounding areas, have the same result: a decrease in the energy reaching
the detector. Absorption is caused by impurities and hydroxyl OHy. ions
present even in the purest glass fibers. Broad infrared and ultraviolet
absorption regions contribute additional absorption. Finally, Rayleigh scattering due to variations in density and composition in the core material causes
further attenuation. The overall loss resulting from all of these factors is least
in wavelength windows around 1300 and 1500 nm. Conveniently, dispersion
in single-mode fibers can be minimized at these wavelengths as explained
above. Figure 15.8 shows optical fiber attenuation versus wavelength.
Figure 15.8. Optical fiber attenuation versus wavelength. (From Figure 11.3, page 723,
Ref. 2.)
562
15.2.5
Types of Optical Fiber
One method of classifying optical fibers is by their refractive index profile

and the number of modes propagated. Classifying them in this manner
derives three types of fiber:
1. Multimode, step-index
2. Multimode, graded-index
3. Single-mode fiber
15.2.5.1 Multimode Fibers. The core diameter of multimode fibers allows
several modes of light to propagate, producing multimode distortion. Core
sizes of glass multimode fibers range from 50 to 200 m or even larger.
Multimode Step-Index Fiber. Conceptually, the simplest fiber is the step-index fiber, with a core of high refractive index and a concentric lower index
cladding, with a sharp interface between the two as shown in Figure 15.9A.
Step-index fiber produces high multimode distortion.
Multimode Graded-Index Fiber. In a graded-index fiber, refractive index
decreases continuously with radial distance from the center of the core,
giving the fiber a nearly parabolic index profile. This is illustrated in Figure
15.9B. Light propagation occurs through refraction, with light rays turned
back toward the core axis less sharply than the reflection angles in a
step-index fiber. Light rays travel a wavelike course down the fiber.
Graded-index fiber displays less multimode distortion than does step-index fiber. Light rays travel a longer path further from the axis, but they travel
faster due to decreasing index of refraction further from the axis. A specialized type of graded-index fiber, called a depressed cladding or W-index fiber
for its characteristic index profile, exhibits lower attenuation than does
conventional graded-index fiber. This fiber has a core and thin cladding
surrounded by an external higher index region. Such a unique composition
prevents leaky modes from the cladding from reentering the core.
Multimode fiber is tending to become obsolete in favor of single-mode
fiber.
15.2.5.2 Single-Mode Fiber. The core of a fiber can be reduced in size
typically 210 microns. where only the axial mode can propagate. This is
called single-mode fiber, and its propagation characteristic is shown in Figure
15.9C. Single-mode fiber, sometimes called monomode fiber, eliminates
multimode distortion because only one transverse mode can propagate,
giving these fibers the broadest bandwidth. A type of material dispersion,
however, does affect single-mode fibers, particularly at very high data rates,
with pulse components of different wavelengths taking different times to
traverse the fiber. This is referred to as chromatic dispersion. See Section
15.2.4.2..
563
Figure 15.9. Index profiles and modes of propagation.
15.2.6
Fiber-Optic Cable
The delicate optical fibers are cabled for protection against mechanical and
environmental abuse during installation and operation. Fibers are cabled
using many of the same materials used to cable metallic media, but employing different cabling techniques. Unlike metallic conductors, optical fibers do
not contribute to the strength of the cable. In fact, the fiber itself must be
decoupled from tensile, thermal, and vibrational loads.
15.2.6.1 Cable Composition. Strengthened fiber-optic cables are used in
all outdoor and most indoor applications. Unstrengthened cables are used in
low-cost multiple-line links in controlled environments, such as within a rack.
The following is a list of typical components of fiber-optic cable. It should be
564
pointed out that not every cable will contain each constituent, and configurations will vary.
1. Coated optical fiber with plastic or silicon elastomer coating to protect
the fiber from scratches and microbending.
2. Single-fiber buffer tubes or multifiber ribbons to cushion fibers and
isolate them from longitudinal stresses during installation and from
microbending during operation. Buffer tubes are not to be confused
with the protective plastic or silicon fiber coating, also called a fiber
jacket.
3. Central and distributed strength members composed of either metallic
or dielectric materials to ensure adequate tensile strength during installation and to resist kinking and crushing of the cable.
4. Polyester binder tape to hold buffer tubes in place.
5. Moisture-barrier compounds.
6. Outer and inner nonconducting jackets composed of low-density
polyethylene resistant to abrasion, chemicals, fungus, moisture, and
decomposition by ultraviolet radiation. High-density polyethylene is
used in armored cable.
7. Corrugated steel tape for protection from rodents.
8. Messenger cable for support of some aerial cables.
15.2.7
Fiber-Optic Transmitters
A conventional fiber-optic transmitter consists of an electrical signal driver, a

light source, and a pigtail that connects the light source to a fiber in the
optical fiber cable, as shown in Figure 15.10. A transmitter in a fiber-optic
communication system converts electrical energy into optical energy and
launches the converted signal into the optical fiber transmission medium.
Conventional fiber-optic light sources are either light-emitting diodes LEDs.
or the higher-performance laser diodes LDs.. Of course, the operating
wavelength of the transmitter must be compatible with its companion receiver.
Figure 15.10. Block diagram of a typical fiber-optic transmitter.
565
Figure 15.11. Optical power versus drive current

for typical light sources.
15.2.7.1 Light-Emitting Diodes. An LED is the preferred light source

when less stringent system requirements permit its use. An LED source tends
to have a longer life, is generally less expensive, and requires lower drive
current than the higher-performance laser diode LD.. An LED is stable over
a long lifetime. LEDs are suitable for analog applications because of their
linearity; optical power output is a nearly linear function of drive current, as
shown in Figure 15.11. However, because of an LEDs broad spectral width
of its emission and because the emitted area of light is larger than the core,
the LED suffers from low coupling efficiency. LEDs have a broad spectral
linewidth, as shown in Figure 15.12, resulting in increased material dispersion. Spectral linewidth is the wavelength interval between half-power 3 dB.
points. In addition, modulation rates are limited. Table 15.1 compares LEDs
and LDs.
Figure 15.12. Typical spectral linewidths for LED and LD. Note that the peak intensities have
been normalized to the same value. The actual peak intensity of an LD is much greater than
that of an LED.
566
TABLE 15.1
Comparison of Typical LEDs and LDs
Characteristic
LED
Coupled power ( W) (max)

(into 50- m core)
Radiant power (mW) (max)
Spectral linewidth (nm)
Wavelength (nm) (max)
Drive current (mA)
Modulation bandwidth (GHz) (max)
Life (hours) (estimated, not guaranteed)
Cost
a
LD
) 50
3000
20
30 to 90
1550
10 to 200
1
10 5 to 10 6
Lower
20
- 1 to 5
1550
10 to 200
10
10 4 to 10 5
Higher
Power remaining after coupling losses have been taken into account.
15.2.7.2 Laser Diodes. Laser diodes LDs. are used for light sources
when a fiber-optic link requires a higher-performance transmitter. Because
of fast risetime, LDs are capable of higher modulation rates. They have
narrower spectral linewidths than do LEDs, as shown in Figure 15.12. As a
result, an LD produces less material dispersion within the fiber than does an
LED. However, LDs require an auxiliary high-voltage power supply for
external biasing to achieve the lasing effect. LDs have a shorter life, are
temperature sensitive, and are less stable than LEDs. They are best suited
for digital transmission. In links where the LD light source overloads the
detector, an attenuator neutral density filter. can be placed between the
light source and the fiber to lessen the light coupled into the fiber, yet allow
the light source to operate within optimum range.
15.2.8
Receivers
A fiber-optic receiver consists of the fiber pigtail or coupler, light detector,

opto-electronic circuitry, and preamplifier as shown in Figure 15.13.
Positive-intrinsic-negative PIN. photodiodes and avalanche photodiodes
APDs. are the two common types of light detectors used in fiber-optic
communication systems. The receiver converts received optical power into an
electrical signal.
15.2.8.1 PIN Photodiodes. Positive-intrinsic-negative PIN. photodiodes
have a longer life and are less subject to thermal instability than APDs. Peak
responsivity is in the 800- to 900-nm range and 1300- to 1600-nm range. They
have a wide dynamic range and good linearity. PINs have no internal gain,
and therefore they exhibit less sensitivity than do APDs. PIN diodes introduce shot noise from photocurrent and leakage current noise. PIN diodes
and APDs are compared in Table 15.2. PINs have a large dynamic range and
567
Figure 15.13. Simplified block diagram of a fiber-optic receiver.
TABLE 15.2
Comparison of Typical PIN Photodiodes and APDs
Characteristic
Sensitivity (dBm for BER or SNR)
Spectral response range (nm)
Noise, equivalent power (Wr'Hz )
Cost
PIN Photodiode
APD
y30 to y45
200 to 1700
y40 to y50
200 to 1700
1 = 10 y 10 to 27 = 10 y 14
Lower
1 = 10 y 14
Higher
have bandwidths exceeding 2 GHz. There is quantum noise associated with

the photodetection process. Thermal noise in the electronic preamplifier is
the controlling noise of the receiver output.
15.2.8.2 Avalanche Photodiodes. An avalanche photodiode APD. has
internal gain, thereby providing a sensitivity approximately 520 dB greater
than the PIN while generating some internal noise. APDs require an external
bias voltage. Peak responsivity is in the 800- to 900-nm range and in the 1300to 1600-nm range. APDs introduce quantum noise i.e., shot noise from
multiplied photocurrent . and surface leakage noise i.e., shot noise from
nonmultiplied leakage current .. They require temperature compensation to
prevent damage and maintain gain. They are suited to long-haul systems and
for high data rates.
15.2.8.3 Noise. Noise appears at the detector and is due to quantum or
photon noise arising in the optical beam itself and thermal noise arising in
the resistances associated with the front-end amplifier. Quantum noise or
photon noise is more evident in highly sensitive detectors such as APDs,
whereas thermal noise is more evident with PIN diodes. An important aspect
of the spectral density of thermal noise is that it is quite constant with
frequencies up to near 10 12 Hz. Above this, thermal noise density begins to
decline and drops off rapidly above 10 13 Hz, so that, although thermal noise
568
is common at microwave frequencies, it is essentially absent at optical

frequencies i.e., above 10 14 Hz.. On the other hand, quantum noise becomes
evident in the 10 12- to 10 13-Hz range right where thermal noise is dropping
off, and it increases directly with frequency. Quantum noise becomes altogether dominant at optical frequencies and sometimes is referred to as blue
noise because of its spectral density characteristics. With current optical
systems, the conversion to electronic signals before detection is not carried
out, and thus the detector front-end amplifier will exhibit both types of noise.
This will change when heterodyne systems begin to be implemented.
15.2.9
Repeaters
Repeaters reshape, retime, and regenerate digital signals by converting

optical signals to electrical signals and back to optical signals again. Signal
conditioning occurs in the electrical domain. Current technology does not
permit such signal conditioning in the optical domain.. A repeater may be
viewed as a receiver back-to-back with a transmitter, with necessary control
and signal conditioning equipment. Some repeaters also contain diagnostic
equipment to transmit alarm status, monitor bit error rate BER., and
provide orderwire functions. To allow for various combinations of emitters
and detectors, some repeaters use plug-in module technology. Repeaters are
usually powered by a serving switch or have an auxiliary power feed. Other
repeaters are powered by a dedicated power supply using local prime power
along with backup battery.
Repeater spacing varies from 15 to over 150 km and depends on the type
of fiber, type of transmitter and detector used, and the bit rate being
transmitted. Figure 15.14 is a functional block diagram of a repeater.
15.2.10
Fiber-Optic Amplifiers
The most common type of amplifier found in fiber-optic systems today is the
erbium-doped fiber amplifier EDFA.. These amplifiers operate in the 1550nm band, which is propitious because it is the band of lowest loss. The
amplifiers bandwidth is on the order of 100s to 1000s of GHz range, which
can support most WDM wavelength division multiplex. operation. Typical
Figure 15.14. A functional block diagram of a fiber-optic cable repeater.
569
amplifier gains are from 30 to 50 dB, and power output gains are from q10
to q20 dBm.
EDFA amplification is based on a pumping principle, where the most
common pumping wavelengths are 1480 or 980 nm. The pump output of from
50 to 100 mW is coupled into a length of fiber from 25 to 100 m in length,
where the longer length displays the greatest gain. Over 100 m, the gain
starts to tail off. For a more detailed description of EDFAs, consult Refs. 1
and 2.
EDFAs are usually installed at a light source output and at the light
detector input. In such a situation with two amplifiers, the gain accrued,
assuming 40 dB per amplifier, is an 80-dB net gain. If we allowed just 0.35
dBrkm including splice loss on a link without amplifiers, and now installed
two amplifiers as we have indicated with a net gain of 80 dB, we can add to
the distance between repeaters some 80r0.35 km or an additional 228 km.
This provides many benefits: less active devices with improved reliability i.e.,
less probability of failure., fewer repeaters thus reduced jitter, some improvement in error performance, and so on. Amplifiers make DWDM dense
wavelength division multiplex. feasible on fiber-optic systems. The amplifiers
compensate for the rather large insertion losses of DWDM systems.
15.2.11
The Joining of Fibers: Splices and Connectors
15.2.11.1 Splices. Splices must provide precise alignment of fiber ends

when completed. Alignment of the prepared ends may be achieved through
the use of an alignment tube, with the fiber ends sealed with epoxy once
alignment has been achieved. More positive alignment may be achieved with
a V-groove guide or a square tube, so that fibers introduced at a slight angle
to the axis of the tube or groove are forced to line up face-to-face. Epoxy is
then applied. Sometimes an index-matching fluid is added to minimize
Fresnel reflection and scattering loss. An alternate method is the use of
fusion splicing. Here the fiber ends are manipulated under a microscope or
by monitoring optical throughputs, and the ends are then fused by an electric
arc or other source of heat. Fusion splice losses of 0.150.25 dB per splice
are common.
15.2.11.2 Connectors. Connectors require efficient signal transfer across
the joined fiber ends, and they are designed to be mated and demated many
times. Insertion losses of connectors vary from as low a 0.5 dB to as high as
1.5 dB. Connectors are either lens- or butt-coupled. Some butt-coupled
connectors use index-matching fluid to reduce loss, whereas others rely on
optical contact Ref. 1..
15.2.12
Modulation and Coding
Digital rather than analog modulation is better suited to most fiber-optic

transmission systems because of the relatively nonlinear characteristics of
570
fiber-optic components. Currently, both analog and digital modulation are

accomplished by intensity modulation of the light source. The most common
application for fiber-optic transmission is telecommunications using PCM
see Chapter 8., or on LANs such as FDDI covered in Section 10.1.4. using
other digital modulation schemes. Analog transmission is widely used in the
cable TV CATV. arena.
15.2.12.1
Digital Signals
Polarity. This section covers bipolar signals only which some call just polar ..
Bipolar signals have two nonzero polarities and are usually symmetrical
about the zero axis.
Coding. Digital signals are also classified according to whether the line
signal state returns to or maintains the zero level during each bit interval, or
uses polarity changes to signify a designated logic element. This is referred to
as coding. Signal codes commonly used in fiber-optic communication systems
are illustrated in Figure 15.15.
Figure 15.15. Digital coding (line formats). (From Ref. 2.)
HIGHER-ORDER DIGITAL MULTIPLEXING FORMATS: SONET AND SDH
15.2.12.2
571
Optical Line Coding and Transmission Rate
Coding. RZ return-to-zero . bipolar and bipolar AMI alternate mark inversion. are ternary, three-state electrical signals. The optical regime in digital
applications, however, supports only binary, two-state signals. Therefore, the
ternary signals must be converted to binary for applications to optical
transmitters. This is done by assigning a two-digit binary code i.e., 00, 01, 10.
to each of the three states. NRZ non-return-to-zero . polar, biphase level,
and conditioned diphase are binary signals and can be applied directly to the
optical transmitter.
Rate. Each data unit interval in ternary coding can contain a maximum of
two states. Because each state is represented by two digits after conversion to
optical line code, the optical line rate is four times the electrical code bit
rate. Biphase level and conditioned diphase coding depend on mid-interval
transitions. Consequently, the optical line rate is twice the electrical code bit
rate. NRZ is the only code of those treated here for which the optical line
rate is the same as the electrical bit rate Ref. 3..
15.3 HIGHER-ORDER DIGITAL MULTIPLEXING FORMATS:
SONET AND SDH
15.3.1
Synchronous Optical Network (SONET)
15.3.1.1 Introduction. SONET provides a digital multiplex format for the

wide bandwidth of optical fiber cable. Prior to the implementation of
SONET, multiplex formats had capacities in the range of 100500 Mbps.
SONET also incorporates some innovations such as slipless operation and
the ability to accommodate digital broadcast-quality TV. SONET also supports B-ISDN.
BellCore describes SONET as a transporter of standard digital. hierarchical signals DS1, DS1C, DS2, and DS3.. It can also transport E-1 hierarchical
signals and ATM. See Figures 15.31 and 15.32 at the end of this section for
system block diagrams. Just because SONET has the word optical in its
name does not mean that its use is exclusively on optical networks. If the
transmission medium has the bandwidth and other parameters such as group
delay meeting transmission requirements., it most certainly can be used to
carry SONET line signals. Millimeter-wave radio is an example.
15.3.1.2 SONET Rates and Formats. SONET has its own digital hierarchy supporting various digital rates for user transport. The SONET hierarchy
is built upon a basic signal of 51.840 Mbps and a byte interleaved multiplex
scheme that results in a family of digital rates and formats defined at a rate
of N times 51.840 Mbps. The basic signal has a portion of its capacity
devoted to overhead, and the remaining portion carries payload.
572
TABLE 15.3
Current Allowable OC-N Line Rates
OC Level
Line Rate (Mbps)
OC-1
OC-3
OC-9
OC-12
OC-18
OC-24
OC-36
OC-48
OC-192
51.840
155.520
466.560
622.080
933.120
1244.160
1866.240
2488.320
9953.28
SONETs basic modular signal is called STS-1 with its 51.840-Mbps data
rate. STS-1s optical counterpart is called OC-1 OC stands for optical
carrier.. OC-1 is directly converted from STS-1 after frame synchronous
scrambling.
Higher-level SONET signals are obtained by synchronously multiplexing
lower-level signals. Higher-level signals are integer multiples of the basic
STS-1 rate and are denoted STS-N and OC-N, where, again, N is an integer.
The current values of N are 1, 3, 9, 12, 24, 36, 48, and 192 Ref. 5.. These
data rates are shown in Table 15.3.
15.3.1.3 STS-1 Frame Structure. The STS-1 frame is shown in Figure
15.16. It consists of 90 columns and 9 rows of 8-bit bytes for a total of 810
bytes. With a frame length of 125 microseconds derived from 8000 frames
per second., STS-1 has a bit rate of 51.840 Mbps. The order of transmission
Figure 15.16. STS-1 frame structure.
573
of bytes is row-by-row, from left to right. The most significant bit is transmitted first, and the bits are numbered 18 Ref. 6..
As shown in Figure 15.16, the first three columns are the transport
overhead, containing overhead bytes of sections and line layers. Of the 27
bytes assigned, 9 are assigned to line overhead and 18 to section overhead.
The remaining capacity, consisting of 87 columns, constitutes the STS-1
envelope capacity.
Figures 15.17, 15.18, and 15.19 show the STS-1 synchronous payload
envelope SPE.. It consists of 87 columns and 9 rows of bytes for a total of
783 bytes. Column 1 is called the STS path overhead POH. and contains 9
bytes with the remaining 774 bytes available for payload. The path overhead
provides the facilities, such as alarm and performance monitoring, required
to support and maintain the transportation of the SPE between end locations
known as path terminations. where the SPE is either assembled or disassembled. The STS-1 SPE may begin anywhere in the STS envelope capacity.
Typically, it begins in one frame and ends in the next or can be wholly
contained in one frame. A key innovation in SONET is the payload pointer.
The STS-1 payload pointer is contained in the transport overhead. It designates the location of the byte where the STS-1 SPE begins.
SPE Assembly Process. The concept of a tributary signal such as a DS3
signal being assembled into an SPE, to be transported end-to-end across a
Figure 15.17. Synchronous payload envelope (SPE).
574
Figure 15.18. STS-1 SPE with STS-1 POH and STS-1 payload capacity. Based on Ref. 6.
Figure 15.19. STS-1 SPE in the interior of an STS-1 frame. Based on Ref. 6.
575
Figure 15.20. SPE assembly process. (From Ref. 7. Courtesy of Hewlett-Packard Company.)
synchronous network, is fundamental to SONET operation. The process of

assembling the tributary signal into an SPE is referred to as payload mapping.
To provide uniformity across all SONET transport capabilities, the payload capacity provided for each individual tributary signal is always slightly
greater than that required by the tributary signal. Thus, the essence of the
mapping process is to synchronize the tributary signal with the payload
capacity provided for transport. This is achieved by adding extra stuffing bits
also called justification bits. to the signal bit stream as part of the mapping
process. Thus, for example, a DS3 tributary signal at a nominal rate of 44
Mbps needs to be synchronized with a payload capacity of 49.54 Mbps
provided by the STS-1 SPE. Addition of the POH completes the assembly of
the STS-1 SPE and increases the bit rate of the composite signal to 50.11
Mbps. This process is shown in Figure 15.20 Ref. 7..
SPE Disassembly Process. At the point of exit from the synchronous
network, the tributary signal that has been transported over the network
needs to be recovered from the SPE which provided the transportation
facilities. This process of disassembling the tributary signal from the SPE is
referred to as payload demapping.
The SPE comprises the POH, the tributary signal, and additional stuffing
bits that have been added in order to synchronize the data transmission rate
of the tributary signal to the payload capacity available for transportation. In
essence the demapping process is to desynchronize the tributary signal from
the composite SPE signal and to reproduce this tributary signal, as nearly as
possible, in its original form. For example, as STS-1 SPE carrying a mapped
DS3 payload arrives at the tributary disassembly location with a signal rate of
50.11 Mbps. Stripping the POH and the stuffing bits from the SPE results in
a discontinuous signal representing the transported DS3 signal with an
average data rate of 44.74 Mbps. These timing discontinuities are reduced by
576
Figure 15.21. SPE disassembly process. (From Ref. 7. Courtesy of Hewlett-Packard

Company.)
Figure 15.22. STS-N frame.
means of a desynchronizing phase-locked loop PLL. in order to produce a

continuous DS3 signal at the same average data rate. Figure 15.21 shows this
process.
15.3.1.4 STS-N Frame Structure. Figure 15.22 shows the STS-N frame
structure. It is made up of N STS-1 byte interleaved signals. Frame alignment of the transport overhead channels of individual STS-1 signals is
carried out prior to byte interleaving. Because of the unique payload pointer,
associated STS SPEs do not require alignment.
SONET provides super rate services that require multiples of the STS-1
rate. Typical of such a rate is the ISDN H4 channel. For these services, the
super rates are mapped into an STS-Nc SPE and transported as a concatenated STS-Nc whose constituent STS-1s are kept together. The STS-Nc is
577
multiplexed, switched, and transported through the network as a single

entity. A concatenation indicator, which is used to show that the STS-Nc is to
be kept together, is contained in the STS-1 payload pointer. Figure 15.23
shows the STS-3c transport overhead assignments, and Figure 15.24 shows
the STS-Nc SPE.
In the STS-Ne SPE, only one set of STS POH is required. The STS-Ne
SPE is carried within the STS-Ne in such a way that the STS POH always
appears in the first of the N-STS-1s that make up the STS-Ne Ref. 10..
15.3.1.5 The Structure of Virtual Tributary. The sub-STS-1 payloads are
transported and switched with the virtual tributary VT. structure. There are
four sizes of VTs VTx.: VT1.5 1.728 Mbps., VT2 2.304 Mbps., VT3 3.456
Mbps., and VT6 6.912 Mbps.. These VT sizes are shown in Figure 15.25. In
the 9-row structure of the STS-1 SPE, these VTs occupy 3 columns, 4
columns, 6 columns, and 12 columns, respectively.
The VT-structured STS-1 SPE is divided into seven VT groups which
efficiently accommodate mixes of VTs. Each VT group occupies 12 columns
of the 9-row structure and may contain 4, 3, 2, or 1 VTs.. The size x . of the
VT determines the number of VTs contained in that particular VT group.
VT groups can carry 4, 3, 2, or 1 VT accommodating 4 VT1.5s, 3 VT2s,
Figure 15.23. Transport overhead assignment, OC-3 carrying an STS-3c SPE. (From Telcordia GR-253-CORE, Issue 2, Revision 2, Figure 3-8, page 3-9. Copyright Telcordia, Jan. 1999.)
578
Figure 15.24. STS-Nc SPE. (From Telcordia GT-253-CORE, Issue 2, Revision 2, Figure 3-7,
page 3-8. Copyright Telcordia, Jan. 1999.)
2 VT3s, or 1 VT6, respectively. However, only one size is allowed in a

particular VT group, but a different size is allowed for each VT group in an
STS-1.
Floating and locked are the two possible modes of operation of the VT
structure. The locked mode minimizes interface complexity in distributed
64-kbps switching. The floating mode minimizes delay for distributed VT
switching.
Floating VT Mode. Four consecutive 125- s frames of the STS-1 SPE are
organized into a 500- s superframe in the floating VT mode. The phase of
the superframe is indicated by the multiframe indicator byte H4. in the STS
POH. This defines a 500- s structure for each of the VTs, which is called the
VT superframe. The VT payload pointer and the VT SPE are contained in the
VT superframe. Four bytes of the superframe are for VT pointer use V1,
V2, V3, and V4.. The remaining bytes define the VT envelope capacity,
which is different for each size, as shown in Figure 15.26.
V1 is the first byte in the superframe V2 through V4 appear as the first
byte in the succeeding VT frames, regardless of size.
The VT payload pointer provides a flexible and dynamic tool for alignment of the VT SPE with the VT envelope capacity, independent of other
VT SPEs. Each VT SPE contains one byte V5. of VT POH, and the
remaining bytes constitute the VT payload capacity, which is different for
each VT size.
579
Figure 15.25. VT sizes. (From Ref. 6. Figure 3-9, page 3-4. Copyright Bellcore, reprinted with
permission.)
580
Figure 15.26. VT superframe and envelope capacity of a VT superframe. Designations:

V1, VT Pointer 1, V2, VT Pointer 2, V3, VT Pointer 3 (Action), V4, VT Reserved. (Adapted
from Ref. 6.)
Locked VT Mode. Here the VT structure contains synchronous payloads

that are locked to the STS-1 SPE. There are no pointers to process in this
mode because the tributary information is fixed and immediately identifiable
with respect to the STS-1 pointer.
Because the locked VTs remain fixed with respect to the STS-1 SPE, the
V1 through V4 pointer bytes and the 500- s VT superframe are not used.
The multiframe indicator byte H4. in this mode defines the 3-millisecond
superframe for DS0 transmission. Also in the VT locked mode there is no
VT POH byte. The two bytes of the 125- s locked VT frame that correspond
to the VT pointer and V5 bytes are reserved. In each 125- s period, the
581
Figure 15.27. Interface layers. (From Ref. 7. Courtesy of Hewlett-Packard Company.)
remaining N y 2 bytes of information capacity is equivalent in the two

modes. A mix of the two modes within an STS-1 SPE is not permitted.
15.3.1.6 SONET Layers and Transport Functions. There are four
SONET layers: physical, section, line, and path. The layers have a hierarchical relationship and are considered from the top down. Figure 15.27 may be
used in showing how each layer requires the services of all lower-level layers
to perform its function.
Bellcore Ref. 6. gives the following example of this interdependability.
Suppose that two path layer processes are exchanging DS3s. The DS3 signal
and the STS path overhead POH. are mapped into an STS-1 SPE, which is
then given to the line layer. The line layer multiplexes several inputs from the
path layer frame and frequency aligning each one. and adds line overhead
e.g., required for protection switching.. Finally the section layer provides
framing and scrambling before optical transmission by the physical layer.
Figures 15.28 and 15.29 show section, line, and path definitions and will help
to place these terms in perspective.
Interface Layers
Physical Layer. The physical layer transports the bits as optical or electrical pulses across the physical medium. There is no overhead associated with
the physical layer. The principal function of this layer is to convert STS
signals and make them suitable for transmission on the selected medium,
which may be optical, metallic, or radio. The physical layer deals with such
things as pulse shape, power levels, and line code.
Section Layer. This layer prepares STS signals for delivery to the physical
layer. It is concerned with framing, scrambling, section error monitoring, and
582
Figure 15.28. Simplified diagram defining section, line, and path. (From Ref. 7. Courtesy of
Hewlett-Packard Company.)
Figure 15.29. More detailed section, line, and path definitions. DCS, digital cross-connect
system; ADM, add-drop multiplex; REG, regenerator; TM, terminal multiplex. (From Figure
2-2, page 2-2, Ref. 6. Copyright Telcordia, reprinted with permission.)
583
the communication of section level overhead, such as a local orderwire. The

section terminating equipment STE. interprets or creates the overhead
associated with the section layer.
Line Layer. This layer provides synchronization and multiplexing functions
for the path layer. The overhead associated with these functions includes that
required for maintenance and protection and is inserted in the line overhead
channels. The line terminating equipment LTE. interprets and modifies or
creates the line overhead. The LTE carries out section layer functions and
therefore the LTE is also an STE.
Path Layer. This layer deals with the transport of network services between SONET terminal multiplexing equipment. Such services include DS1,
DS3, and DS4NAs.
The path layer maps these services into the format required by the line
layer and communicates end-to-end with the POH. The path terminating
equipment interprets and modifies or creates the overhead defined for this
layer.
Layer Interaction. Figure 15.27 shows the interactions of the interface
layers. Similar to OSI open system interconnection ., each layer communicates horizontally to peer equipment in that layer. It also processes certain
information and passes it vertically to the next layer.
The path layer transmits horizontally to its peer entities the services and
the path layer overhead. The path layer maps the services and POH into
SPEs that it passes vertically to the line layer.
The line layer transmits SPEs with line layer overhead to its peer entities.
At this time the SPEs are synchronized and multiplexed. Then the STS-N
signal is passed to the section layer.
The section layer communicates with its peer entities, transmitting to them
STS-N signals and section layer overhead e.g., local orderwire.. This layer
maps STS-N and section overhead into bits that are handed over to the path
layer, which transmits optical or electrical pulses to its peer entities.
It should be noted that access to all four layers is not required of every
SONET NE.* As an example, an OC-N regenerator uses only the first two
layers physical and section layers.. Also, an NE that merely routes SPEs and
does not accept any new inputs from the path layer only uses the first three
layers. However, in this case these NEs may only monitor the overhead of
layers that they do not terminate Ref. 6..
15.3.2
Synchronous Digital Hierarchy
Synchronous digital hierarchy SDH. resembles SONET in many respects.

The SDH design is behind SONET in maturity. It represents the European
*NE stands for network element.
584
TABLE 15.4
SDH Bit Rates
SDH Level
Hierarchical Bit Rate (kbps)
1
4
16
64
155,520
622,080
2,488,320
9,953,280
Figure 15.30. Generalized SDH multiplexing structure. C-n stands for container-n. (From
Figure 2-1rG.708, page 3, ITU-T Rec. G. 708, Ref. 9.)
hegemony much like the E-1 hierarchy does. It is generally understood to be

more flexible than SONET and was designed more with the PTT government-owned telecommunications administrations . in mind. The SONET design, on the other hand, has more flavors of the private network.
15.3.2.1 SDH Standard Bit Rates. The standard operational SDH bit
rates are shown in Table 15.4. ITU-T Rec. G.707 Ref. 8. states that the
first level of the digital hierarchy shall be 155,520 kbps and that higher
synchronous digital hierarchy bit rates shall be obtained as integer multiples
of the first level bit rate.
15.3.2.2 Interface and Frame Structure of SDH. Figure 15.30 illustrates
the relationship between various multiplexing elements that are given below
and shows possible multiplexing structures. Figures 15.31, 15.32, and 15.33
show specific derived multiplexing methods.
Definitions
Synchronous Transport Module (STM). An STM is the information structure used to support section layer connections in the SDH. It is analogous
to STS in the SONET regime. STM consists of information payload and
section overhead SOH. information fields organized in a block frame
structure which repeats every 125 s. The information is suitably conditioned for serial transmission on selected media at a rate which is
585
Figure 15.31. Multiplexing method directly from container 1 using AU-4. Note: Unshaded
areas are phase-aligned. Phase alignment between the unshaded and shaded areas is
defined by the pointer (PTR) and is indicated by the arrow. (From Figure 6-2rG.707, page 7,
ITU-T Rec. G.707, Ref. 8.)
synchronized to the network. A basic STM STM-1. is defined at 155,520

kbps. Higher-capacity STMs are formed at rates equivalent to N times
multiples of this basic rate. STM capacities for N s 4 and N s 16 are
defined, and higher values are under consideration by ITU-T.
An STM comprises a single administrative unit group AUG. together
with the SOH. STM-N contains N AUGs together with SOH.
Container, C-n ( n s 1 to n s 4). This element is a defined unit of payload
capacity which is dimensioned to carry any of the levels currently defined
in Section 15.3.2.1 and may also provide capacity for transport of broadband signals which are not yet defined by CCITT ITU-T organization..
Virtual Container-n (VC-n ). A virtual container is the information structure
used to support path layer connection in the SDH. It consists of information payload and POH information fields organized in a block frame which
repeats every 125 or 500 s. Alignment information to identify VC-n
frame start is provided by the server network layer. Two types of virtual
container have been identified:
586
Figure 15.32. SDH multiplexing method directly from container 1 using AU-3. Note: Unshaded areas are phase-aligned. Phase alignment between the unshaded and shaded areas
is defined by the pointer (PTR) and is indicated by the arrow. (From Figure 6-3rG.707, page
8, ITU-T Rec. G.707, Ref. 8.)
Figure 15.33. Multiplexing method directly from container 3 using AU-3. Note: Unshaded
areas are phase-aligned. Phase alignment between the unshaded and shaded areas is
defined by the pointer (PTR) and is indicated by the arrow. (From Figure 6-4rG.707, page 9,
ITU-T Rec. G.707, Ref. 8.)
587
Lower-order irtual container-n, to VC-n n s 1, 2.. This element comprises a single C-n n s 1, 2., plus the basic virtual container POH
appropriate to that level.
Higher-order irtual container-n, to VC-n n s 3, 4.. This element comprises a single C-n n s 3, 4., an assembly of tributary unit groups TUG2s., or an assembly of TU-3s, together with virtual container POH appropriate to that level.
Administrative Unit-n (AU-n ). An administrative unit is the information
structure which provides adaptation between the higher-order path layer
and the multiplex section. It consists of an information payload the
higher-order virtual container. and an administrative unit pointer which
indicates the offset of the payload frame start relative to the multiplex
section frame start. Two administrative units are defined. The AU-4
consists of a VC-4 plus an administrative unit pointer which indicates the
phase alignment of the VC-4 with respect to the STM-N frame. The AU-3
consists of a VC-3 plus an administrative unit pointer which indicates the
phase alignment of the VC-3 with respect to the STM-N frame. In each
case the administrative unit pointer location is fixed with respect to the
STM-N frame Ref. 9..
One or more administrative units occupying fixed, defined positions in
an STM payload are termed an administratie unit group AUG.. An AUG
consists of a homogeneous assembly of AU-3s or an AU-4.
Tributary Unit-n, TU-n. A tributary unit is an information structure which
provides adaptation between the lower-order path layer and the higherorder path layer. It consists of an information payload the lower-order
virtual container. and a tributary unit pointer which indicates the offset of
the payload frame start relative to the higher-order virtual container frame
start.
The TU-n n s 1, 2, 3. consists of a VC-n together with a tributary unit
pointer.
One or more tributary units occupying fixed, defined positions in a
higher-order VC-n payload are termed a tributary unit group TUG.. TUGs
are defined in such a way that mixed-capacity payloads made up of
different-sized tributary units can be constructed to increase flexibility of
the transport network.
A TUG-2 consists of a homogeneous assembly of identical TU-1s or a
TU-2. A TUG-3 consists of a homogeneous assembly of TUG-2s or a
TU-3.
Container-n ( n s 14). A container is the information structure which forms
the network synchronous information payload for a virtual container. For
each of the defined virtual containers there is a corresponding container.
Adaptation functions have been defined for many common network rates
into a limited number of standard containers. These include standard
E-1rDS-1 rates defined in ITU-T Rec. G.702.
588
Figure 15.34. STM-N frame structure. (From Figure 6-6rG.707, page 11, ITU-T Rec. G.707,
Ref. 8.)
15.3.2.3 Frame Structure. The basic frame structure, STM-N, is shown in

Figure 15.34. The three main areas of the STN-N frame are section overhead, AU pointers, and information payload.
Section oerhead is shown in rows 13 and 59 of columns 19 = N of the
STM-N in Figure 15.34.
Administrative Unit (AU) Pointers. For columns 19 = N in Figure 15.34,
row 4 is available for AU pointers. The positions of the pointers of the AUs
for different organizations of the STM-1 payload are shown in Table 15.5.
See ITU-T Rec. G.709 for application of pointers and their detailed specifications.
The following taken from Ref. 8. summarizes the rules for interpreting
the AU-n pointers:
1. During normal operation, the pointer locates the start of the VC-n
within the AU-n frame.
2. Any variation from the current pointer value is ignored unless a
consistent new value is received three times consecutively or it is
preceded by one of the rules 3, 4, or 5. Any consistent new value
received three times consecutively overrides i.e., takes priority over.
rules 3 or 4.
3. If the majority of the I bits of the pointer word are inverted, a positive
justification operation is indicated. Subsequent pointer values are incremented by one.
4. If the majority of the D bits of the pointer word are inverted, a negative
justification operation is indicated. Subsequent pointer values are
decremented by one.
5. If the NDF new data flag. is set to 1001, then the coincident pointer
value replaces the current one at the offset indicated by the new
pointer value unless the receiver is in a state that corresponds to a loss
of pointer.
589
Administrative Units in the STM-N. The STM-N payload can support N

AUGs where each AUG may consist of either one AU-4 or three AU-3s.
The VC-n associated with each AU-n does not have a fixed phase with
respect to the STM-N frame. The location of the first byte of the VC-n is
indicated by the AU-n pointer. The AU-n pointer is in a fixed location in the
STM-N frame. This is shown in Figures 15.35 and 15.36.
The AU-4 may be used to carry, via the VC-4, a number of TU-ns
n s 1, 2, 3. forming a two-stage multiplex. An example of this arrangement
is illustrated in Figures 15.35a and 15.36a. The VC-n associated with each
TU does not have a fixed-phase relationship with respect to the start of the
VC-4. The TU-n pointer is in a fixed location in the VC-4, and the location
of the first byte of the VC-n is indicated by the TU-n pointer.
590
Figure 15.35. Administrative units in the STM-1 frame, X denotes AU-n pointer; AU-n denotes
AU-n pointer q VC-n. (From Figure 6-7rG.707, page 12, ITU-T Rec. G.707, Ref. 8.)
Figure 15.36. Two-stage multiplex. X denotes AU-n pointer; denotes TU-n pointer; AU-n
denotes AU-n pointer q VC-n; TU-n denotes TU-n pointer q VC-n. (From Figure 6-8rG.707,
page 12, ITU-T Rec. G.707, Ref. 8.)
The AU-3 may be used to carry, via the VC-3, a number of TU-ns
n s 1, 2. forming a two-stage multiplex. An example of this arrangement is
illustrated in Figures 15.35b and 15.36b. The VC-n associated with each
TU-n does not have a fixed-phase relationship with respect to the start of the
VC-3. The TU-n pointer is in a fixed location in the VC-3, and the location
of the first byte of the VC-n is indicated by the TU-n pointer.
15.3.2.4 Interconnection of STM-1s. The SDH has been designed to be
universal, allowing transport of a layer variety of signals such as those
specified in ITU-T Rec. G.702, including North American 1.544-Mbps and
European 2.048-Mbps regimes. However, different structures can be used
for the transport of virtual containers. The following interconnection rules
LINE-OF-SIGHT MICROWAVE AND MILLIMETER-WAVE TRANSMISSION SYSTEMS
591
Figure 15.37. SDH multiplexing structure. Note: G.702 tributaries associated with containers
C-x are shown. Other signals (e.g., ATM) can also be accommodated. (From Figure
6-1rG.707, page 6, ITU-T Rec. G.707, Ref. 8.)
are used:
1. The rule for interconnecting two AUGs based upon two different types
of administrative unit, namely AU-4 and AU-3, is to use the AU-4
structure. Therefore, the AUG based upon AU-3 is demultiplexed to
the TUG-2 or VC-3 level according to the type of the payload, and then
it is remultiplexed within an AUG via the TUG-3rVC-4rAU-4 route.
2. The rule for interconnecting VC-11s transported via different types of
tributary unit, namely TU-11 and TU-12, is to use the TU-11 structure.
VC-11, TU-11, and TU-12 are described in ITU-T Rec. G.709.
15.3.2.5 Basic SDH Multiplexing Structure. The SDH multiplexing
structure is shown in Figure 15.37.
15.4 LINE-OF-SIGHT MICROWAVE AND MILLIMETER-WAVE

TRANSMISSION SYSTEMS
15.4.1
Broadband Radio Systems
Radio systems operating in frequency ranges of 3.7 GHz and above generally
are assigned some 500 MHz of bandwidth. There are several reasons why one
user cannot use the entire 500-MHz pipe.In most circumstances, full-duplex
operation is required. Leaving aside interference considerations, this drops
the usage pipe width in half to 250 MHz i.e., 250 MHz in each direction..
Further, there is a high probability that there will not be uniform behavior
across bandwidths as wide as 250 MHz. In addition, such frequency bands
592
must be shared with other nearby users. Thus usable bandwidths are reduced
to 30 or 40 MHz per radio-frequency carrier for line-of-sight microwave and
100 MHz or less for satellite radio systems. Systems operating above 28 GHz
have bandwidths of 2.5 GHz, with some bandwidths in the mid-millimeter
range over 5 GHz wide. The problem with these higher-frequency systems is
cost. There is available technology, but production runs are low because of
low demand.
Why use radio at all when fiber-optic systems have almost unlimited
coherent bandwidths? Again it is cost. If it is cheaper, with equal performance and equally dependable, then use it. This proves out very often for
line-of-sight microwave radio systems in heavily urbanized areas and in areas
of difficult terrain. Typically, the area for the first application would be New
York City, and that for the second would be the jungled mountains of
Venezuela.
Satellite radio systems can have much broader applications, especially for
private networks. Again the driving factor is cost, performance, and dependability. In this regard, space segment charges become a cost factor; furthermore, the space segment can be a point failure if network availabilityrsurvivability is a primary consideration, and it may be a drawback.
15.4.2
An Overview of Line-of-Sight Microwave
15.4.2.1 Line of Sight. A microwave terminal consists of a radio transmitter, a receiver, and an antenna subsystem. The antenna is mounted on a
tower and is connected to the transmitter and receiver by means of a
transmission line, usually waveguide. Economic factors limit tower height to
some 100-m 300-ft. maximum.
This leads to the so-called line-of-sight LOS. concept. The maximum
distance from one microwave terminal to the next is limited by LOS, which is
a function of tower height. Radio waves at such high frequencies do not bend
around obstacles in the conventional sense; they are either diffracted or
blocked entirely. There is, however, some bending which we will explain.
The term line of sight gives the connotation that microwave radio rays
behave exactly like light waves. One difference is that radiowaves at these
microwave frequencies are somewhat bent because of variations in the
refractive index of the atmosphere through which they progress. In about
80% of circumstances we can get some 15% more range between microwave
terminals than we would get using light as the transmission medium.
A microwave radio ray beam is either diffracted or blocked by an obstacle
as we mentioned above. The most common obstacle is the horizon. If there
are no other obstacles, which microwave engineers call smooth earth, then
the distance to the radio horizon can be calculated for a given antenna tower
height h. by the following formulas:
d mi s
' 2 h .
593
Figure 15.38. Radio line of sight versus optical line of sight.
where d is in statute miles and h, the antenna tower. height, is measured in

feet, and
'
d km s 2.9 2 h .
where d is in kilometers and h is measured in meters Ref. 2..
If a tower is 100 m high, the radio distance to the horizon is 41 km. This
concept is shown in Figure 15.38. If we build a second tower directly in line
also 100 m high, the link would be 2 = 41 km long i.e., 82 km*., there would
be smooth earth, and there would be no other intervening obstacles with
standard refraction.
Of course, we can achieve much greater distances if we take advantage of
natural and man-made high spots. Use tops of mountains, ridges, cliffs, and
tops of high buildings, or lease space on high towers typically TV towers..
Such techniques can notably extend LOS distances.
15.4.2.2 Typical Line-of-Sight Microwave Installation. Figure 15.39 is a
simplified conceptual block diagram of a microwave terminal. The transmitter emits an analog carrier wave that, in our case, is modulated digitally. We
might transmit an OC-3 configuration SONET. or STM-1 bit rate of 155
Mbps. How can we get 155 Mbps in a 30- or 40-MHz bandwidth? The trick is
similar to those we described in Section 6.3.2, where we squeezed much
greater data rates into a voice channel than it appeared it could accommodate. In essence we converted to a higher-level digital signal using what is
called M-ary modulation techniques often of a hybrid nature. Typical of
*This calculation is only good to a first approximation.
594
Figure 15.39. A functional block diagram of a digital LOS microwave terminal. Typical PCM
line codes: AMI, polar, BNZS, HDB3rCMI. BB, baseband; , optional feature. (From Ref. 2.)
595
these was the quadrature amplitude modulation QAM., which is hybrid in

that it combined both amplitude and phase modulation.
Because digital transmission is notoriously wasteful of RF bandwidth,
regulatory authorities such as the U.S. Federal Communications Commission
FCC. require a certain minimum bit packing. In the case of the FCC, at least
4.5 bits per hertz of bandwidth are required. 64-QAM provides 6 bitsrHz
theoretically and about 4.5 bitsrHz practically. Thus 30 MHz can accommodate 30 MHz = 4.5 bps, or 135 Mbps.* For STM-1rSTS-3 rate, probably
128-QAM might be more appropriate.
Microwave transmitters generally have just 1-watt 0-dBW. outputs, and
10-watt traveling wave tube TWT. or solid-state amplifiers SSA. are available off-the-shelf if required. Often, in the case of millimeter wave installations, transmitter output may be measured in milliwatts.
Nearly all microwavermillimeter wave links suffer from fading. Simplistically, this is the variation in level of the received signal from its free space
value. Fading can cause intersymbol interference or insufficient signal-tonoise ratio, both affecting error performance. It can even cause loss of frame
alignment. We mitigate effects of fading by overbuilding the link. Here we
use larger antennas than required for an unfaded link; or we can increase
power output of the transmitter by using a 10-watt HPA; or we can improve
receiver noise performance by using low-noise amplifiers LNAs.. There are
still other techniques available to the link design engineer such as coding gain
devices and diversity.
Because of user congestion in the popular RF band from 1 to 10 GHz,
more and more new installations are being forced to use frequencies above
10 GHz. Frequency bands above this 10-GHz arbitrary borderline suffer from
excess attenuation over the free space value. due to heavy downpour.
rainfall. This will impact the time availability of a link. Time aailability is the
decimal value or percentage of time that a particular link meets performance
objectives. We can use similar techniques overbuilding a link. Usually this is
sufficient to meet an objectives criterion. In practice, in either case, we
generally recommend that antennas no greater than 3.5 m 12 ft. be used.
The reason, of course, is economicnot so much in antenna cost as in the
cost of mechanically stiffening towers to meet twist and sway maximums.
Large parabolic dish antennas act like a big sail surface on top of a tower,
and wind places twisting torques on the antenna and tower such that severe
fading or circuit outages result.
It is again emphasized that as time progresses, there will be more and
more pressure to use the higher frequencies i.e., ) 23 GHz. because of the
ever relentless demand for frequencies and greater bandwidth allocation.
*Without filter roll-off properties such as the raised cosine filter.
Also, ray beamwidths get progressively smaller as antenna gains increase, further magnifying the
twist and sway situation; it gets easier and easier for the transmit ray beam to fall out of the
capture range of the receive antenna, and vice versa.
596
Figure 15.40. A simplified diagram of a satellite communication payload.
There are several high-absorption bands due to the gaseous content of the
atmosphere. There is one particularly severe absorption band between 58
and 62 GHz. If we stay away from these absorption bands, then we can either
live with the rainfall problem or overbuild a link to mitigate. In this regard,
channel coding and interleaving can be very effective, if properly designed.
Thus these higher frequencies will be particularly attractive even for
STS-3rSTM-1 and higher SONETrSDH configurations.
15.4.3
Satellite Communications
15.4.3.1 Concept and Applications. Satellite communications is an extension of LOS microwave. In fact, in nearly every case it uses the same
frequency bands as LOS microwave. In other words, both services share the
same bands, which can be a drawback. Again we deal with 500-MHz
bandwidths. This means that the total allocation for the service in a particular band is 500 MHz. One exception to the 500-MHz rule is the 30r20-GHz
band which allocates 2.5 GHz of bandwidth.
The nominal 500 MHz of bandwidth is usually broken down into segments
of 36 MHz; however, in some cases the segments are 72 MHz and in a few
cases they are 100 MHz or slightly greater. A satellite is an RF repeater as
shown in Figure 15.40. The RF repeating device is called a transponder, and
there is a transponder tuned to the center frequency of each frequency
segment and with sufficient amplitude response frequency response. to
cover the 36-MHz or 72-MHz. bandwidth. A typical satellite might have 24
such transponders.
To increase satellite capacity, frequency reuse is implemented. Satellite
capacity can be doubled or tripled by this technique, reusing the same
frequencies. If there is doubling of capacity by frequency reuse, there is a
doubling of the number of transponders. Spot beam antennas with low
sidelobes plus good polarization isolation permit frequency reuse, cutting
interference from one system into the other down to a reasonable level.
597
15.4.3.2 Uplinks and Downlinks. An uplink carries the digitalrdata signal from an earth station to the satellite. In the satellite it is converted in
frequency and returned to a distant earth station on a downlink. With only
few exceptions, the downlink is always the weak link. There are several
reasons why, and these reasons give us insight into satellite communications.
On an uplink there are few limitations. We can easily spew all the RF
power necessary up into space, provided it is confined to reach only the
satellite of interest. There is no limitation on prime power generation and
little chance of causing interference, provided that earth station sidelobes are
held in check and the angular spread of the ray beam is contained. RF power
not used by the satellite in its receiving system spreads out into space, which
harms no one.
Quite different conditions hold for the downlink. The downlink power
must be limited for two important reasons. First, the satellite spews RF
power onto the earths surface in a band shared by LOS microwave. If the
power level is too great, harmful interference will occur on LOS links
encompassed by the satellite ray beam. Thus downlink power must be limited
to limit interference on LOS microwave service. Second, downlink power is
also very expensive for three reasons: i. Prime power on the satellite is
generated by solar cells; that is, more power requires more solar cells. ii. If
more power is required, larger batteries are needed for darkness periods and
solar eclipse. iii. If more power is needed, larger components such as power
supplies . are required on the satellite, and thus more weight is implied. The
cost of launching a satellite is really a function of its weight Ref. 2..
Figure 15.41 illustrates a satellite connectivity with uplinks and downlinks.
15.4.3.3 Satellite Orbits. Satellites are in elliptical orbits. As we remember, a circle is a special form of an ellipse, and most satellites used for
commercial communications in the western world are in circular orbits 35,786
km approximately 22,254 miles. above the equator. These orbits are called
Figure 15.41. A satellite connectivity showing

uplinks and downlinks.
598
geostationary where the satellite appears fixed over a point on the earths
surface. The primary advantage of such an orbit is that earth station antenna
tracking in many cases is not required. The major drawback is the resulting
propagation delay up to the satellite and down again, some 125 ms each way.
Secondarily is the increased free space loss incurred by the large distance.
Another class of satellites used for telecommunication are the low earth
orbit LEO. satellites such as the Iridium system operated by Motorola. Here
the altitude of the satellite is around 500 km, vastly reducing the delay and
reducing free space loss accordingly. Such satellites ordinarily would require
tracking from earth stations; however, quasi-omnidirectional antennas are
used on the satellites as well as on the earth terminals. The primary market
for Iridium is worldwide cellular and personal communication services PCS..
Another is the TELEDESIC system with hundreds of LEO satellites to be
used for the specialized fixed-service data transport market.
15.4.3.4 Digital Satellite System Operation and Access. There are three
generic access techniques available for satellite communications. By access
we mean how we can use these satellite resources. Such generic techniques
are:
Frequency division multiple access FDMA.

Time division multiple access TDMA.
Code division multiple access CDMA.
FDMA. An uplink user is assigned a frequency segment inside a transponder

bandwidth. Other users are assigned other segments. There are guard bands
between contiguous segments. These user frequency slots or segments are
transmitted uplink to the satellite where there is a frequency translation and
the segments are passed onwards on the downlink for reception at distant
earth stations. At a distant earth station, only those segments are removed
that are destined for it. The FDMA concept is shown in Figure 15.42a. Each
frequency slot or segment may carry a frequency division multiplex configuration for an analog network, a digital configuration for digital network
interconnectivity, or a pure data configuration for private data network
connectivity. A segment, for instance, might be 10 MHz wide and be used for
LAN interconnectivity. The modulation in this case might be quadrature
phase-shift keying QPSK., easily accommodating a 10-Mbps data signal.
TDMA. This access technique operates in the time domain where each user is
assigned a time slot, rather than a frequency slot as in FDMA. Each user, in
hisrher assigned time slot, uses the entire transponder bandwidth with a
data burst for the time slot duration. A user bursts and waits for hisrher next
turn, bursts and waits. Thus TDMA only operates in the digital domain
where traffic storage is practical to permit the burst and wait operation.
TDMA time slots can be varied in their duration in accordance with traffic
demand. Figure 15.42b is a conceptual diagram of TDMA.
REFERENCES
599
Figure 15.42. Conceptual drawings of ( a ) FDMA, ( b ) TDMA, and ( c ) CDMA in the timefrequency plane. (From Ref. 1.)
CDMA. Code division multiple access uses one of two forms of spread
spectrum wave form: frequency hop FH. or direct sequence DS.. Direct
sequence is also called pseudo-noise. In either case, users can be stacked one
on the other using the same center frequency. For FH each user has a
different hopping pattern; for DS each user has a different sequence determined by a key ariable. It can be argued that spread spectrum systems allow
more users per unit bandwidth than does TDMA or FDMA. Figure 15.42c is
a conceptual diagram of CDMA. Spread spectrum systems are ideally suited
for digital data transmission.
VSAT Networks. There is one very special and particularly attractive application of satellite communications for the data end user for a wide area
network WAN.. This is the very-small-aperture terminal VSAT. approach.
VSAT networks were discussed in Chapter 11, Section 11.7.
REFERENCES
1. Roger L. Freeman, Reference Manual for Telecommunications Engineering, 3rd
edition, John Wiley & Sons, New York, 2001.
2. Roger L. Freeman, Telecommunications Transmission Handbook, 4th edition, John
3. Design Handbook for Fiber Optic Communications Systems, MIL-HDBK-415, U.S.
Dept. of Defense, Washington DC, 1985.
600
4. The IEEE Standard Dictionary of Electrical and Electronics Terms, 6th edition,
IEEE, New York, 1996.
5. Synchronous Optical Network SONET . Common Generic Criteria, GR-253-CORE,
Issue 1, TelcordiarBellcore, Morristown, NJ, 1994.
6. Synchronous Optical Network SONET . Transport Systems: Common Generic Criteria, TR-TSY-GR-253-CORE, Issue 2, Revision 2, TelcordiarBellcore, Morristown, NJ, Jan. 1999.
7. Introduction to SONET, an H-P seminar, Hewlett-Packard Company, Burlington,
MA, October 1993.
8. Synchronous Digital Hierarchy Bit Rates, ITU-T Rec. G.707, ITU Geneva, March
1996.
9. Network Node Interface for the Synchronous Digital Hierarchy, ITU-T Rec. G.708,
10. Curtis A. Siller and Mansoor Shafi, SONETrSDH A Sourcebook of Synchronous
Networking, IEEE Press, New York, 1996.
16
BROADBAND ISDN (B-ISDN)
AND THE ASYNCHRONOUS
TRANSFER MODE (ATM)
16.1
WHERE ARE WE GOING?
Frame relay Chapter 12. was the beginning of the march toward an optimized* format for multimedia transmission voice, data, video, facsimile..
There were new and revolutionary concepts in frame relay. There was a
trend toward simplicity where the header was considerably shortened. A
header is pure overhead, so it was cut back as much as practically possible.
A header also implies processing. By reducing the processing, delivery time
could be speeded up. Frame relay is a low-latency system.
In the effort to speed up delivery, operation was unacknowledged at least
at the frame relay level.; there was no operational error correction scheme. It
was unnecessary because it was assumed that the underlying transport system
had excellent error performance better than 1 = 10y9 .. There was error
detection for each frame, and a frame found in error was thrown away. Now
that is something that we never did for those of us steeped in old-time data
communication. It is assumed that the higher OSI open system interconnection. layers would request repeats of the few frames missing i.e., thrown
away..
Frame relay also moved into the flow control arena with the BECN and
FECN bits and the CLLM. The method of handling flow control has a lot to
do with its effectiveness in this case. It also uses a discard eligibility DE. bit
that set a type of priority to a frame. If the DE bit were set, the frame would
be among the first to be discarded in a time of congestion.
The distributed queue dual bus DQDB. developed by the IEEE 802.6
committee provides a simple and unique access scheme. Even more impor*Compromise might be a more appropriate word.
601
602
BROADBAND ISDN (B-ISDN) AND THE ASYNCHRONOUS TRANSFER MODE (ATM)
tant, its data transport format is based on the cell that the IEEE calls a slot.
The DQDB slot or cell has a format very similar to that of the ATM cell,
which we discuss at length in this chapter. It even has the same number of
octets i.e., 53., 48 of which is the payload capacity. This is identical to ATM.
Also enter a comparatively new concept of the header check sequence HCS.
for detecting errors in the header. The DQDB has no error detection for the
body or info portion of the slot or cell. There is a powerful CRC32 in the
trailer of the IMPDU initial MAC protocol data unit., an upper layer of
DQDB. It also employs the BOM, COM, and EOM beginning of message,
continuation of message, and end of message, respectively . as well as the
message identifier MID. in the segmentation and reassembly SAR. process.
These fields are particularly helpful in message reassembly.
Switched multimegabit data service SMDS. came on the scene using
DQDB as its access protocol. DQDB uses a 5-octet ACFrheader and a
48-octet payload with no trailer, which is very similar to the ATM cell. SMDS
has a 44-octet payload, a 7-octet header, and a 2-octet trailer, for a total of 53
octets in what we will call a cell. Many of the strategies of DQDB followed
on down to SMDS. However, SMDS has a credit manager at the switching
system SS. which polices users. This policing concept came from frame relay.
It is carried onward into ATM.
We consider ATM to be a prime candidate to transport data. Data traffic
is bursty, and the short cells, which are the basic transport format for ATM,
can handle short bursts very efficiently. IP and frame relay, because of their
comparatively long frames, are less amenable to short burst traffic. It is most
attractive to the user when comparatively high speed circuits are available
such as DS3 or E3 see Chapter 7.. Service can also be provided down to DS1
and E1 circuits. Reference 1 states that ATM tends to be expensive.
Broadband-ISDN B-ISDN. provides the framework for the transport of
ATM cells. Its nomenclature is similar to that of ISDN Chapter 13..
Broadband is not a very well-defined term. In this context it is any digital
channel with a data rate greater than DS0 56 or 64 kbps. or E0 64 kbps..
Certainly the ISDN H0 channel falls in this category with its 384-kbps date
rate. We like to think of B-ISDN being that transport framework starting at
the ITU primary rate of 1.544 and 2.048 Mbps.
16.2
INTRODUCTION TO ATM
ATM is based on a 53-octet cell. A typical cell has a 5-octet header and a
48-octet information field. All traffic is transported in the information field
of that cell, whether voice, data, or image.
As we discussed, ATM is an outgrowth of the several data transmission
format systems covered previously. Admittedly, some may argue this point.
ATM cells can be mixed and matched as shown in Figure 16.1, where a cell
carrying a data payload may be contiguous with another cell carrying voice.
ATM cells can be transported on SONET, SDH Chapter 15., E1rDS1
INTRODUCTION TO ATM
603
Figure 16.1. ATM links simultaneously carry a mix of voice, data, and image information.
Chapter 7., and other popular digital formats. Cells can also be transported
contiguously without an underlying digital network format.
Philosophically, voice and data must be handled differently regarding time
sensitivity. Voice cannot wait for long processing and ARQ delays. Most
types of data can. So ATM must distinguish the type of service such as
constant bit rate CBR. and variable bit rate VBR. services. Voice service is
typical constant bit rate or CBR service.
Signaling is another area of major philosophical difference. In data communications, signaling is carried out within the header of a data frame or
packet.. As a minimum the signaling will have the destination address and
quite often the source address as well. And this signaling information will be
repeated over and over again on a long data file that is heavily segmented.
On a voice circuit, a connectivity is set up, and the destination address, and
possibly the source address, are sent just once during call setup. There is also
some form of circuit supervision to keep the circuit operational throughout
the duration of a telephone call. ATM is a compromise, stealing a little from
each of these separate worlds.
Like voice telephony, ATM is fundamentally a connection-oriented
telecommunication system. Here we mean that a connection must be established between two stations before data can be transferred between them. An
ATM connection specifies the transmission path, allowing ATM cells to
self-route through an ATM network. Being connection-oriented also allows
ATM to specify a guaranteed quality of service QoS. for each connection.
By contrast, most LAN protocols are connectionless. This means that LAN
nodes simply transmit traffic when they need to, without first establishing a
specific connection or route with the destination node.
In that ATM uses a connection-oriented protocol, bandwidth is allocated
only when the originating end user requests a connection. This allows ATM
604
to efficiently support a networks aggregate demand by allocating bit rate

capacity on demand based on immediate user need. Indeed it is this concept
which lies in the heart of the word asynchronous. An analogy would help.
New York City is connected to Washington, DC, with a pair of railroad
tracks for passenger trains headed south and another pair of tracks for
passenger trains headed north. On those two pairs of tracks we would like to
accommodate everybody we can when they would like to ride. The optimum
for reaching this goal is to have a continuous train of coupled passenger cars.
As the train enters Union Station, it disgorges its passengers and connects
around directly for the northward run to Pennsylvania Station. Passenger
cars are identical in size, and each has the same number of identical seats.
Of course at 2 A.M. the train will have very few passengers and many
empty seats. Probably from 7 to 9 A.M. the train will be full, no standees
allowed, so we will have to hold potential riders in the waiting room. They
will ride later; those few who try to be standees will be bumped. Others might
seek alternate transportation to Washington, DC.
Here we see that the railroad tracks are the transmission medium. Each
passenger car is a SONETrSDH frame. The seats in each car are our ATM
cells. Each seat can handle a person no bigger than 53 units. Because of
critical weight distribution, if a person is not 53 units in sizerweight, we will
place some bricks on the seat to bring the sizerweight to 53 units exactly.
Those bricks are removed at the destination. All kinds of people ride the
train because America is culturally diverse, analogous to the fact that ATM
handles all forms of traffic. The empty seats represent idle or unassigned
cells. The header information is analogous to the passengers tickets. Keep in
mind that the train can only fill to its maximum capacity of seats. We can
imagine the SONETrSDH frame as being full of cells in the payload, some
cells busy and some idlerunassigned. At the peak traffic period, all cells will
be busy, and some traffic passengers . may have to be turned away.
We can go even further with this analogy. Both Washington, DC, and New
York City attract large groups of tourists, and other groups travel to business
meetings or conventions. A tour group has a chief tour guide in the lead seat
cell. and an assistant guide in the last seat cell.. There may be so many in
the group that they extend into a second car or may just intermingle with
other passengers on the train. The tour guide and assistant tour guide keep
an exact count of people on the tour. The lead guide wears a badge that says
BOM, all tour members wear badges that say COM, and the assistant tour
guide wears a badge that says EOM.
Each group has a unique MID message ID.. We also see that service is
connection-oriented Washington, DC, to New York..
Asynchronous means that we can keep filling the seats on the train until
we reach its maximum capacity. If we look up the word, it means nonperiodic,
whereas the familiar E-1rT-1 are periodic i.e., synchronous.. One point that
seems to get lost in the literature is that the train has a maximum capacity.
Thus the concept bandwidth on demand is that we can use the bandwidth
USERNETWORK INTERFACE (UNI ), CONFIGURATION, AND ARCHITECTURE
605
until we fill to rated capacity. Again we have the unfortunate use of the word
bandwidth, because our capacity will be measured in octets, not hertz. For
example, SONETs STS-1 has a payload capacity of 87 = 9 octets see
Chapter 15., not 87 = 9 Hz.
16.3 USERNETWORK INTERFACE (UNI), CONFIGURATION,

AND ARCHITECTURE
ATM is the underlying message format of B-ISDN. At times in this section,
we use the terms ATM and B-ISDN interchangeably. Figures 16.2 and 16.3
interrelate the two. Figure 16.2 relates B-ISDN access reference configuration with the ATM usernetwork interface UNI.. Figure 16.3 is the traditional ITU-T Rec. I.121 Ref. 2. B-ISDN protocol reference model showing
the extra layer necessary for switched service. There is absolutely no relationship between this model and the OSI reference model.
The ATM Forum Ref. 3. provides the following definitions applicable to
Figure 16.3:
U-Plane. The user plane provides for the transfer of user application information. It contains physical layer, ATM layer, and multiple ATM adaptation layers required for different service users such as constant bit rate
service CBR. and variable bit rate service VBR..
Figure 16.2. B-ISDN usernetwork interface (UNI) configuration. The reference model is at
the top of the drawing. Note its similarity to the ISDN model, Figure 13.2. The R reference
point is where nonstandard ISDN equipment is made B-ISDN-compatible via B-TA. In this
case, the TA functionality is limited to physical layer conversion.
606
Figure 16.3. B-ISDN protocol reference model.
C-Plane. The control plane protocols deal with call establishment and call
release and other connection control functions necessary for providing
switched services. The C-plane structure shares the physical and ATM
layers with the U-plane, as shown in Figure 16.3. It also includes ATM
adaptation layer AAL. procedures and higher-layer signaling protocols.
M-Plane. The management plane provides management functions and the
capability to exchange information between the U-plane and the C-plane.
The M-plane contains two sections: layer management and plane management. Layer management performs layer-specific management functions,
while the plane management performs management and coordination
functions related to the complete system.
We return to Figure 16.3 and B-ISDNrATM layering and layer descriptions in Section 16.6.
16.4
16.4.1
THE ATM CELL: KEY TO OPERATION

ATM Cell Structure
The ATM cell consists of 53 octets; 5 of these make up the header, and 48
octets carry the payload or info portion of the cell. The basic structure is
shown in Figure 16.4, where contiguous cells flow on a circuit from left to
right. Figure 16.5 shows the detailed structure of the cell header at the UNI.
607
Figure 16.4. Basic ATM cell structure is shown where a stream of contiguous cells flow from
left to right.
Figure 16.5. Header structure at the UNI (left); header structure at the NNI (right). GFC,
generic flow control; VPI, virtual path identifier; VCI, virtual channel identifier; PTI, payload
type; CLP, cell loss priority, HEC, header error control.
Now let us examine each of the bit fields that make up the ATM cell
header.
Generic Flow Control (GFC). These first 4 bits are either part of the VPI, if
the cells is traveling between two switches, or make up a reserved field if the
cell is traveling between an end-node and a switch. In this case it has local
significance only and can be used to provide standardized local flow control
function on the customer site.
VPI r VCI Fields (Routing Field). Twenty-four bits are available for routing a
cell from a customer site. There are 8 bits for virtual path identifier VPI.
and 16 bits for virtual channel identifier VCI.. Preassigned combinations of
VPI and VCI values are given in Table 16.1. The VCI value of zero is not
available for user virtual channel identification. The bits within the VPI and
608
TABLE 16.1
Combinations of Preassigned VP, VCI, PT, and CLP Values at the UNI
Use
Unassigned cell
Invalid
Metasignaling
(see Rec. I.311)
General broadcast signaling
(see Rec. I.311)
Point-to-point signaling
(see Rec. I.311)
Segment OAM F4 flow cell
(see Rec. I.610)
End-to-end OAM F4 flow cell
(see Rec. I.610)
VP resource management cell
(see Rec. I.371)
Reserved for future VP functions
(Note 6)
Reserved for future functions
(Note 7)
(Note 7)
Segment OAM F5 flow cell
(see Rec. I.610)
VPI
00000000
Any VPI value
other than 0
XXXXXXXX
(Note 1)
XXXXXXXX
(Note 1)
XXXXXXXX
(Note 1)
Any VPI value
Any VPI value
Any VPI value
Any VPI value
Any VPI value
Any VPI value
Any VPI value
End-to-end OAM F5 flow cell

(see Rec. I.610)
Any VPI value
VC Resource management cell

(see Rec. I.371)
Any VPI value
Reserved for future VC

functions
Any VPI value
VCI
(Note 8)
PTI
CLP
00000000 00000000
00000000 00000000
Any value
Any value
0
B
00000000 00000001
(Note 5)
00000000 00000010
(Note 5)
00000000 00000101
(Note 5)
00000000 00000011
(Note 4)
00000000 00000100
(Note 4)
00000000 00000110
(Note 10)
00000000 00000111
(Note 10)
00000000 000SSSSS
(Notes 2 and 10)
00000000 000TTTTT
(Note 3)
Any VCI value other than
00000000 00000000,
00000000 00000011,
00000000 00000100,
00000000 00000110, or
00000000 00000111
00000000 00000000,
00000000 00000011,
00000000 00000100,
00000000 00000110, or
00000000 00000111
00000000 00000000,
00000000 00000011,
00000000 00000100,
00000000 00000110, or
00000000 00000111
00000000 00000000,
00000000 00000011,
00000000 00000100,
00000000 00000110, or
00000000 00000111
0AA
0AA
0AA
0A0
(Note 11)
0A0
(Note 11)
110
(Note 9)
0AA
(Note 11)
0AA
0AA
100
101
110
111
A
A
A
The GFC field is available for use with all of these combinations.
A Indicates that the bit may be 0 or 1 and is available for use by the appropriate ATM layer function.
B Indicates the bit is a dont care bit
C Indicates the originating entity shall set the CLP bit to 0. The value may be changed by the
network.
Tables notes continued on page 609
609
VCI fields used for routing are allocated using the following rules:
The allocated bits of the VPI field are contiguous.

The allocated bits of the VPI field are the least significant bits of the
VPI field, beginning at bit 5 of octet 2.
The allocated bits of the VCI field are contiguous.
The allocated bits of the VCI field are the least significant bits of the
VCI field, beginning at bit 5 of octet 4.
Comment: Consider a very high-speed backbone carrying millions of messages. The split between VPI and VCI saves routers in the backbone from
requiring that their call-mapping database keep track of millions of individual
calls. In this situation, the backbone routers use only the VPI portion of the
call identifier. Because thousands of VCs might be going on the same VP,
switches inside can treat all the VCs for that VP as a unit.
Outside the backbone, the switches treat the VPIrVCI together as one
combined field. The term VP-switching refers to switches that are looking at
610
only the VPI portion of the field. VC-switching refers to switches that are
looking at the entire field.
Payload Type (PT) Field. The first bit s 0 indicates data rather than ATM
control information. For data, the middle bit indicates congestion experienced, and the last bit is used by AAL5 see Section 16.6.2.2. to indicate last
cell of a packet. Payload Type Identifier PTI. values are given in Table 16.2.
The main purpose of the PTI is to discriminate between user cells i.e., cells
carrying user information. and nonuser cells. The first four code groups in
the table 000-011. are used to indicate user cells. Within these four, 2, and 3
010 and 011. are used to indicate congestion experienced. The fifth and sixth
code groups 100 and 101. are used for VCC-level management functions.
Any congested network element, upon receiving a user data cell, may
modify the PTI as follows: Cells received with PTI s 000 or PTI s 010 are
transmitted with PTI s 010. Cells received with PTI s 001 or PTI s 011 are
transmitted with PTI s 011. Noncongested network elements should not
change the PTI.
Cell Loss Priority (CLP) Field. Depending on network conditions, cells where
the CLP is set i.e., CLP value is 1. are subject to discard prior to cells where
the CLP is not set i.e., CLP value is 0.. The concept here is identical with
that of frame relay and the DE bit. ATM switches may tag CLP s 0 cells
detected by the usage parameter control UPC. to be in violation of the
traffic contract by changing the CLP bit from 0 to 1.
Header Error Control (HEC) Field. The HEC is an 8-bit field, and it covers
the entire cell header. The code used for this function is capable of either
TABLE 16.2
PTI Coding
PTI Coding
Bits
4 3 2
0 0 0
0 0 1
0 1 0
0 1 1
1
1
1
1
0
0
1
1
0
1
0
1
Interpretation
User data cell, congestion not experienced.
ATM-user-to-ATM-user indication s 0
User data cell, congestion not experienced.
User data cell, congestion experienced.
User data cell, congestion experienced.
OAM F5 segment associated cell
OAM F5 end-to-end associated cell
Resource management cell
Source: Table appearing on page 6 in Section 2.2.4, ITU-T Rec. I.361 (11r95 ),
Ref. 4.
611
Figure 16.6. HEC receiver modes of operation.
single-bit error correction or multiple-bit error detection. Briefly, the transmitting side computes the HEC value. The receiver has two modes of
operation, as shown in Figure 16.6. In the default mode there is the
capability of single-bit error correction. Each cell header is examined; if an
error is detected, one of two actions takes place. The action taken depends
on the state of the receiver. In the correction mode, only single-bit errors can
be corrected and the receiver switches to the detection mode. In the detection mode, all cells with detected header errors are discarded. When a
header is examined and found not to be in error, the receiver switches to the
correction mode. The term no action in Figure 16.6 means that no
correction is performed and no cell is discarded.
Figure 16.7 is a flow chart showing the consequences of errors in the ATM
cell header. The error protection function provided by the HEC provides for
a. recovery from single-bit errors and b. a low probability of delivery of cells
with errored headers under bursty error conditions. ITU-T Rec. I.432.1 Ref.
5. states that error characteristics of fiber-optic transmission systems appear
to be a mix of single-bit errors and relatively large burst errors. Thus, for
some transmission systems the error correction capability might not be
invoked.
16.4.2
Header Error Control Sequence Generation
The transmitter calculates the HEC value across the entire ATM cell header
and inserts the result in the appropriate header field.
The notation used to describe header error control is based on the
property of cyclic codes. For example, code vectors such as 1000000100001
it has 13 elements or bits, count them. can be represented by a polynomial
P x . s X 12 q X 5 q 1. The elements of an n-element code word are thus
the coefficients of a polynomial of the order n y 1. In this application,
these coefficients can have the value 0 or 1 and the polynomial operations
are performed using modulo 2 operations. The polynomial representing the
612
Figure 16.7. Consequences of errors in an ATM cell header. (From Figure 4rI.432.1, page 4,
ITU-T Rec. I.432.1, Ref. 5.)
content of a header excluding the HEC field is generated using the first bit of
the header as the coefficient of the highest-order term.
The HEC field is an 8-bit sequence. It is the remainder of the division
modulo 2. by the generator polynomial X 8 q X 2 q X q 1 of the product
X 8 multiplied by the content of the header excluding the HEC field.
At the transmitter, the initial content of the register of the device computing the remainder of the division is preset to all 0s and is then modified by
division of the header excluding the HEC field by the generator polynomial
as described above.; the resulting remainder is transmitted as the 8-bit HEC.
To improve the cell delineation performance in the case of bit-slips, the
following is recommended:
The check bits calculated by the use of the check polynomial are added
modulo 2. to an 8-bit pattern before being inserted in the last octet of
the header.
The recommended pattern is 01010101 the left bit is the most
significant bit..
CELL DELINEATION AND SCRAMBLING
TABLE 16.3
Header
pattern
613
Header Pattern for Idle Cell Identification

Octet 1
Octet 2
Octet 3
Octet 4
Octet 5
00000000
00000000
00000000
00000001
HEC s Valid code

01010010
The receiver must subtract equal to add modulo 2. the same pattern
from the 8 HEC bits before calculating the syndrome of the header.
This operation in no way affects the error detectionrcorrection capabilities

of the HEC.
As an example, if the first 4 octets of the header were all zeros, the
generated header before scrambling would be 00000000 00000000 00000000
00000000 01010101. The starting value for the polynomial check is 0s
binary..
16.4.3
Idle Cells
Idle cells cause no action at a receiving node except for cell delineation
including HEC verification. Idle cells are inserted and discarded for cell rate
decoupling. Idle cells are identified by the standardized pattern for the cell
header shown in Table 16.3.
16.5
16.5.1
CELL DELINEATION AND SCRAMBLING

Delineation and Scrambling Objectives
Cell delineation allows identification of cell boundaries. In other parts of this

book we simply called this alignment. The cell HEC field achieves cell
delineation. Keep in mind that the ATM signal must be self-supporting in
that it has to be transparently transported through every network interface
without any constraints from the transmission systems used.
Scrambling is used to improve security and robustness of the HEC cell
delineation mechanism discussed below. In addition, it helps the randomizing
of data in the information field for possible improvement in transmission
performance.
Any scrambling operation must not alter the ATM header structure,
header error control, or cell delineation algorithm.
614
Figure 16.8. Cell delineation state diagram. Note: The correct HEC means that the header
has no bit error (syndrome is zero) and has not been corrected. (From Figure 5rI.432.1, page
6, ITU-T Rec. I.432.1, Ref. 5.)
16.5.2
Cell Delineation Algorithm
Cell delineation is performed by using the correlation between the header

bits to be protected 32 bits or 4 octets. and the HEC octet, which are the
relevant control bits 8 bits. introduced in the header using a shortened cyclic
code with the generating polynomial X 8 q X 2 q X q 1.
Figure 16.8 shows the state diagram of HEC cell delineation method.
A discussion of the figure follows:
1. In the HUNT state, the delineation process is performed by checking
bit by bit for the correct HEC i.e., syndrome equals zero. for the
assumed header field. For the cell-based* physical layer, prior to
scrambler synchronization, only the last 6 bits of the HEC are used for
cell delineation checking. For the SDH-based interface, all 8 bits are
used for acquiring cell delineation. Once such an agreement is found, it
is assumed that one header has been found, and the method enters the
PRESYNC state. When octet boundaries are available within the
receiving physical layer prior to cell delineation as with the SDH-based
interface, the cell delineation process may be performed octet by octet.
2. In the PRESYNC state, the delineation process is performed by checking cell by cell for the correct HEC. The process repeats until the
correct HEC has been confirmed Delta times consecutively. If an
incorrect HEC is found, the process returns to the HUNT state.
*Only cell-based and SDH-based interfaces are covered by current ITU-T recommendations.
Besides these, we cover cells riding on other transport means at the end of this chapter.
ATM LAYERING AND B-ISDN
615
Figure 16.9. In-sync time versus bit error probability. (From Figure I.1rI.432.1, page 13,
ITU-T Rec. I.432.1, Ref. 5.)
3. In the SYNCH state, the cell delineation will be assumed to be lost if

an incorrect HEC is obtained Alpha times consecutively.
The parameters Alpha and Delta are chosen to make the cell delineation
process as robust and secure as possible while satisfying QoS requirements.
Robustness depends on Alpha when it is against false misalignments due to
bit errors. And robustness depends on Delta when it is against false delineation in the resynchronization process.
For the SDH-based physical layer, values of Alpha s 7 and Delta s 6 are
suggested by the ITU-T organization Rec. I.432, Ref. 5.; and for cell-based
physical layer, values of Alpha s 7 and Delta s 8. Figures 16.9 and 16.10
give performance information of the cell delineation algorithm in the presence of random bit errors, for various values of Alpha and Delta.
16.6
The B-ISDN reference model is given in Figure 16.3, and its several planes
are described therein. We reiterate that B-ISDN layering has no relationship
whatsoever with the OSI reference model.
616
Figure 16.10. Acquisition time versus bit error probability. (From Figure I.2rI.432.1, page 14,
ITU-T Rec. I.432.1, Ref. 5.)
16.6.1
Functions of Individual ATMr

r B-ISDN Layers
Figure 16.11 illustrates B-ISDNrATM layering and sublayering of the protocol reference model. It identifies the functions of the physical layer, the ATM
layer and the AAL and related sublayers.
16.6.1.1 Physical Layer. The physical layer consists of two sublayers. The
physical medium PM. sublayer includes only PM-dependent functions. The
transmission convergence TC. sublayer performs all functions required to
transform a flow of cells into a flow of data units i.e., bits. which can be
transmitted and received over a physical medium. The service data unit
SDU. crossing the boundary between the ATM layer and the physical layer
is a flow of valid cells. The ATM layer is unique meaning independent of the
underlying physical layer.. The data flow inserted in the transmission system
payload is PM-independent and self-supported. The physical layer merges
the ATM cell flow with the appropriate information for cell delineation,
according to the cell delineation mechanism described above, and carries the
operations and maintenance OAM. information relating to this cell flow.
617
Figure 16.11. B-ISDNrATM functional layering. CS, convergence sublayer; PM, physical
medium; SAR, segmentation and reassembly (sublayer); TC, transmission convergence.
The physical medium sublayer provides bit transmission capability including bit transfer and bit alignment as well as line coding and electrical optical
transformation. Of course, the principal function is the generation and
reception of waveforms suitable for the medium, the insertion, and extraction
of bit timing information and line coding where required. The primitives
identified at the border between the PM and TC sublayers are a continuous
flow of logical bits or symbols with this associated timing information.
Transmission Convergence Sublayer Functions. Among the important functions of this sublayer are the generation and recovery of the transmission
frame. Another function is transmission frame adaptation which includes the
actions necessary to structure the cell flow according to the payload structure
of the transmission frame transmit direction. and to extract this cell flow out
618
of the transmission frame receive direction.. The transmission frame may be

a cell equivalent i.e., no external envelope is added to the cell flow., an
SDHrSONET envelope, an E-1rT-1 envelope, and so on.
In the transmit direction, the HEC sequence is calculated and inserted in
the header. In the receive direction, we include cell header verification. Here
cell headers are checked for errors and, if possible, header errors are
corrected. Cells are discarded where it is determined that headers are
errored and are not correctable.
Another transmission convergence function is cell rate decoupling. This
involves the insertion and removal of idle cells in order to adapt the rate of
valid ATM cells to the payload capacity of the transmission system. In other
words, cells must be generated to exactly fill the payload of SDHrSONET
for example. whether the cells are idle or busy.
Section 16.12 gives several examples of transporting cells using the convergence sublayer.
16.6.1.2 The ATM Layer. Table 16.4 shows the ATM layer functions
supported at the UNI U-plane.. The ATM layer is completely independent
of the physical medium. One important function of this layer is encapsulation.
This includes cell header generation and extraction. In the transmit direction,
the cell header generation function receives a cell information field from a
higher layer and generates an appropriate ATM cell header except for the
HEC sequence. This function can also include the translation from a service
access point SAP. identifier to a virtual path VP. and virtual circuit VC.
identifier.
In the receive direction, the cell header extraction function removes the
ATM cell header and passes the cell information field to a higher layer. As in
the transmit direction, this function can also include a translation of a VP
and VC identifier into an SAP identifier.
In the case of the NNI, the generic flow control GFC. is applied at the
ATM layer. The flow control information is carried in assigned and unassigned cells. Cells carrying this information are generated in the ATM layer.
TABLE 16.4
ATM Layer Functions Supported at the UNI
Functions
Multiplexing among different
ATM connections
Cell rate decoupling
(unassigned cells)
Cell discrimination based on
predefined header field values
Payload type discrimination
Loss priority indication and
selective cell discarding
Traffic shaping
Parameters
VPI r VCI
Preassigned header field values
Preassigned header field values
PT field
CLP field,
network congestion state
Traffic descriptor
619
In a switch, the ATM layer determines where the incoming cells should be
forwarded to, resets the corresponding connection identifiers for the next
link, and forwards the cell. The ATM layer also handles traffic management
functions and buffers incoming and outgoing cells. It indicates to the next
higher layer the AAL. whether or not there is congestion during transmission. The ATM layer monitors both transmission rates and conformance to
the service contract; this is called traffic shaping and traffic policing.
Cell Rate Decoupling. Cell rate decoupling includes insertion and suppression of idle cells in order to adapt the rate of valid ATM cells to the payload
capacity of the transmission system Ref. 6, ITU-T Rec. I.321.. According to
the ITU, this function is carried out in the TC transmission convergence.
portion of the physical layer. The ATM Forum provides the following
statement: The cell rate decoupling function at the sending entity adds
unassigned cells to the assigned stream i.e., cells with valid payloads. to be
transmitted, transforming a noncontinuous stream of assigned cells into a
continuous stream of assigned and unassigned cells. At the receiving entity
the opposite operation is performed for both unassigned and invalid cells.
The rate at which the unassigned cells are insertedrextracted depends on the
bit rate and rate variation. of the assigned cell andror the physical layer
transmission rate. The ATM Forum carries out the function in the ATM
layer and uses different terminology i.e., assigned and unassigned cells.,
whereas ITU-T carries out the function in the TC sublayer and uses the term
idle cells. Note the possible incompatibility here. Also note the unassigned cell coding in Table 16.1, and note idle cell coding in Table 16.3.
Physical layers that have synchronous cell time slots generally require cell
rate decoupling typically SONETrSDH, DS3rE3, etc.., whereas physical
layers that have asynchronous cell time slots do not require this function
because no continuous flow of cells need to be provided.
Cell Discrimination Based on Predefined Header Field Values. Certain
predefined header field values at the UNI are given in Table 16.1. The
functions of several of these are described as follows:
Meta-Signaling cells are used by the meta-signaling protocol for establishing and releasing signaling virtual channel connections. For virtual
channels allocated permanently PVC., meta-signaling is not used.
General Broadcast signaling cells are used by the ATM network to broadcast signaling information independent of service profiles.
F4 OAM cells. The virtual path connection VPC. operation flow F4 flow.
is carried via specially designated OAM cells. F4 flow OAM cells have
the same VPI value as the user-data cells transported by the VPC but
are identified by two unique preassigned virtual channels within this
VPC. At the UNI, the virtual channel identified by a VCI value of 3 is
620
used for VP-level management functions between ATM nodes on both

sides of the UNI i.e., single VP link segment., while the virtual channel
identified by a VCI value of 4 can be used for VP-level end-to-end
user-to-user . management functions.
B-ISDN operation and maintenance OAM. is a responsibility of the
physical and ATM layers. It is discussed in Section 16.13.
16.6.2
The ATM Adaptation Layer (AAL)
The basic purpose of the AAL is to isolate the higher layers see Figure 16.3.
from the specific characteristics of the ATM layer by mapping the higher-layer
protocol data units PDUs. into the information field of the ATM cell and
vice versa.
16.6.2.1 Sublayering of the AAL. To support services above the AAL, a
number of independent functions are required of the AAL. These functions
are organized in two logical sublayers: the convergence sublayer CS. and the
segmentation and reassembly sublayer SAR.. The principal functions of
these sublayers are:
SAR. Here the functions are a. segmentation of higher-layer information
into a size suitable for the information field at an ATM cell and b.
reassembly of the contents of ATM cell information fields into higherlayer information.
CS. Here the prime function is to provide the AAL service at the
AAL-SAP. This sublayer is service-dependent.
16.6.2.2 Service Classification for the AAL. The service classification is
based on the following parameters:
Timing relation between source and destination this refers to urgency

of traffic .: required or not required
Bit rate: constant or variable
Connection mode: connection-oriented or connectionless
When we combine these parameters, four service classes A, B, C, and D.

emerge as shown in Table 16.5. There are five different AAL types or
categories. The simplest is AAL-0. It just transmits cells down a pipe. That
pipe is commonly a fiber-optic link. Ideally, we would like the bit rate to be
some multiple of 53 = 8 or 424 bits. For example, 424 Mbps would handle 1
million cells per second.
TABLE 16.5
Service Classification for AAL
Parameters
Timing
compensation
Bit rate
Connection
mode
Example
AAL Type
621
Class A
Class B
Class C
Class D
Required
Required
Not required
Not required
Constant
Connectionoriented
Circuit emulation,
voice, video
Variable
Connectionoriented
Variable bit rate
video or audio
Variable
Connectionless
Type 1
Type 2
Variable
Connectionoriented
Connectionoriented
data transfer
Type 3, Type 5
Connectionless
data transfer
Type 4
AAL-1. AAL-1 is used to provide transport for synchronous bit streams.

Its primary application is to adapt ATM cell transmission to circuits
representative of E1rT1 and SDHrSONET. Typically, AAL-1 is for voice
communications POTSplain old telephone service.. AAL-1 checks that
mis-sequencing of information does not occur by verifying a 3-bit sequence
counter and allows for regeneration of the original clock timing of the data
received at the far end of the link. The SAR-PDU format of AAL-1 is shown
in Figure 16.12. The 4-bit sequence number SN. is broken down into a 1-bit
convergence sublayer indicator CSI. and sequence count. The SNP sequence number protection. contains a 3-bit CRC and a parity bit. Clock
recovery is via a synchronous residual time stamp SRTS. and common
network clock by means of a 4-bit residual time stamp extracted from CSI of
Figure 16.12. SAR-PDU format for AAL-1. SN, sequence number; SNP, sequence number
protection.
622
cells with odd sequence numbers. The residual time stamp is transmitted
over 8 cells. Alarm indication in this adaptation layer is via a check of the
ones density. When the ones density of the received cell stream becomes
significantly different from the density used for the particular PCM line
coding scheme in use, it is determined that the system has lost signal and
alarm notifications are given Ref. 7..
AAL-2. AAL-2 handles the VBR scenario such as MPEG* video. Functions
performed by AAL-2 include:
Segmentation and reassembly SAR. of user information

Handling of cell delay variation
Handling of lost and misinserted cells
Source clock frequency recovery at the receiver
Recovery of the source data structure at the receiver
Monitoring of AAL-PCI
Handling of AAL-PCI bit errors
Monitoring of user information field for bit errors and possible corrective action
AAL-2 is still in ITU-T definitive stages. However, an example of an

SAR-PDU format for AAL-2 is given in Figure 16.13. The following informa-
Figure 16.13. Example of an SAR-PDU format for AAL-2. The need of each of these fields, the
position of these fields, and their size require further study by the ITU-T Organization. SN is
the sequence number, used to detect lost or misinserted cells. A specific value of the
sequence number may indicate a specified purpose. IT is information type, used to indicate
beginning of message (BOM), continuation of message (CM), and end of message (EOM),
provides timing information, and also is a component of the audio or video signal; LI is length
indicator, checks to indicate that the number of octets of the CS-PDU are included in the
SAR-PDU payload field. CRC is cyclic redundancy check, used to correct up to two correlated
bit errors. (From Figure 1rI.363, page 4, ITU-T Rec. I.363, Ref. 8; see also Ref. 9.)
*MPEG is a set of digital video compression schemes. MPEG stands for Motion Picture Experts
Group.
PCI stands for protocol control information.
623
tion is passed from the user plane to the management plane:
Errors in the transmission of user information

Loss of timingrsynchronization
Lost or misinserted cells
Cells with errored AAL-PCI
Buffer underflow or overflow
AAL-2 Conergence Sublayer (CS). This sublayer performs the following

functions:
1. Clock recovery for variable bit rate audio and video services is performed by means of the insertion of a time stamp or real-time synchronization word in the CS-PDU.
2. Sequence number processing is performed to detect the loss or misinsertion of ATM-SDUs. The handling of lost and misinserted ATMSDUs is also performed in this sublayer.
3. For audio and video services, forward error correction may also be
performed.
The following is a comment by Radia Perlman taken out of context, Ref.
10, page 174.: Recently a new AAL was proposed for multiplexing several
voice channels into a single cell. Instead of being called AAL-6, its being
called AAL-2 and is under development.
AAL-3r4. Initially, in CCITT Rec. I.363 1991, Ref. 8., there were two
separate AALs, one for connection-oriented variable bit rate data services
AAL-3. and one for connectionless service AAL-4.. As the specifications
evolved, the same procedures turned out to be necessary for both of these
services, and the specifications were merged to become the AAL-3r4 standard. It is used for ATM transport of SMDS, CBDS connectionless broadband data services, an ETSI initiative., and frame relay.
AAL-3r4 has been designed to take variable-length framesrpackets maximum 65,535 octets. and segment them into cells. The segmentation is done
in a way that protects the transmitted data from corruption if cells are lost or
mis-sequenced.
Radia Perlman again comments taken out of context, Ref. 10.: AAL-3r4
attempts to provide service that allows you to send a reasonable sized
datagram. It was done in a complex and inefficient way, and people stepped
in and designed AAL-5, which is better in all ways, and AAL-3r4 will
eventually go away.
Modes of Serice. AAL-3r4 offers two different modes of service: Message
mode and streaming mode.
624
Figure 16.14. Formation of an ATM cell from a user frame, showing terminology used.
With the message mode, a CPCS-PDU is passed across the CPCS interface in its entirety in exactly one CPCS-IDU IDU, interface data unit.. In
other words, it handles a complete user frame as one piece. In the case of the
streaming mode, parts of a user frame can be sent at different points in time,
or a long frame may be sent on to a receiver before the entire frame has been
received by the transmitter Ref. 11..
Figure 16.14 depicts the steps required to accept a variable length user
data frame or packet and to develop ATM cells for transmission. These steps
are numbered 1 through 4:
1. Variable length packets, typically from frame relay, are padded to an
integral word length and encapsulated with a header and a trailer to
form what is called the conergence sublayer PDU CPCS..
2. Padding is added to make sure that fields align themselves to 32-bit
boundaries, allowing the efficient implementation of the operations in
625
Figure 16.15. Common part convergence sublayer (CPCS) PDU format.
Figure 16.16. SAR-PDU format for AAL-3r4.
hardware at lower layers. The CPCS adds a 4-octet header and a

4-octet trailer to form the CPCS PDU. The CPCS header and trailer
contain identifiers Btag and Etag. and length indicators BAsize and
Length. that help the receiving end allocate appropriate buffer space,
reassemble the CS data unit, and do basic error checks see Figure
16.15..
3. The CPCS-PDU is now broken down into 44-octet segments by the
SAR sublayer, which, in its turn, adds a 2-octet header and a 2-octet
trailer to each of the segments, forming what is referred to as an SAR
Protocol Data Unit SAR-PDU. see Figure 16.16.. This SAR-PDU has
a total of 48 octets including its headers and trailers.
4. These 48 octets now become the information field of a standard ATM
cell.
DISCUSSION OF FIGURES
16.15
AND
16.16
Discussion of Figure 16.15. The user data Figure 16.14. is encapsulated

with a 4-octet header and a 4-octet trailer. The header consists of the
following fields:
Common Part Indicator CPI field, 1 octet.. This field is used to interpret
subsequent fields for the CPCS functions in the header and trailer. The
626
counting units for the values specified in the BAsize and Length fields may
be indicated; other users are under study by the ITU-T Organization.
Beginning Tag (Btag) Field 1 octet.. This field allows the association of the
CPCS-PDU header and trailer. The sender inserts the same value in the
Btag and the Etag in the trailer for a given CPCS-PDU and changes
the values of each successive CPCS-PDU. The receiver checks the value of
the Btag in the header with the value of the Etag in the trailer. It does not
check the sequence of the BtagrEtags in successive CPCS-PDUs.
As an example, a suitable mechanism is as follows: The sender increments
the value placed in the Btag and Etag field for each successive CPCS-PDU
sent over a given MID value. Btag values are cycled up to modulo 256.
Buffer Allocation Size (BAsize) Indication Field 2 octets.. This field indicates to the receiving peer entity the maximum buffering requirements to
receive the CPCS-SDU. In message mode, the BAsize value is encoded
equal to the CPCS-PDU payload length. In streaming mode, the BAsize
value is encoded equal to or greater than the CPCS-PDU payload length.
The buffer allocation size is binary encoded as number of counting units.
The size of the counting units is identified by the CPI field.
Padding (PAD) Field 0 to 3 octets.. Between the end of the CPCS-PDU
payload and the 32-bit aligned CPCS-PDU trailer, there will be from 0 to
3 unused octets. These unused octets are called the Padding PAD. field.
They are strictly used as filler octets and do not convey any information. It
may be set to 0, and its value is ignored at the receiving end. This
padding field complements the CPCS-PDU payload to an integral multiple
of 4 octets.
Alignment (AL) Field 1 octet.. The function of the alignment field is to
achieve 32-bit alignment in the CPCS-PDU trailer. The alignment field
complements the CPCS-PDU trailer to 32 bits. This unused octet is strictly
used as a filler octet and does not convey any information. The alignment
field is set to zero.
End Tag (Etag) Field 1 octet.. For a given CPCS-PDU, the sender inserts
the same value in this field as was inserted in the Btag field in the header
to allow the association of the trailer with its header.
Length Field 2 octets.. The length field is used to encode the length of the
CPCS-PDU payload field. This field is also used by the receiver to detect
the loss or gain of information. The length is binary encoded as number of
counting units. The size of the counting units is identified by the CPI field.
Note: The length of the CPCS-PDU payload is limited to the maximum
value of the length field multiplied by the value of the counting unit.
627
Discussion of Figure 16.16

Segment Type (ST) Field 2 bits.. The segment-type indication identifies an
SAR-PDU as containing a Beginning of Message BOM. coded 10, a
Continuation of Message COM. coded 00, an End of Message EOM.
coded 01, or a Single Segment Message SSM. coded 11.
Sequence Number (SN) Field 4 bits.. The 4 bits allocated to this field allow
the stream of SAR-PDUs of a CPCS-PDU to be numbered modulo 16.
Each SAR-PDU belonging to an SAR-PDU and hence associated with
a given MID value. has its sequence number incremented by one relative
to its previous sequence number. The receiver checks the sequence of the
sequence number field of SAR-PDUs derived from one SAR-SDU and
does not check the sequence number of the sequence number field of the
SAR-PDUs derived from successive SAR-SDUs. As the receiver does not
check the sequence number continuity between SAR-SDUs, the sender
may set the sequence number field to any value from 0 to 15 at the
beginning of each SAR-SDU.
Multiplexing Identification (MID) Field 10 bits.. This field is used for
multiplexing. If no multiplexing is used, this field is set to zero.
In connection-oriented applications, it may be used to multiplex multiple SAR connections on a single ATM connection.
In connectionless and connection-oriented applications, all SAR-PDUs
of an SAR-SDU will have the same MID field value. The MID field is
used to identify SAR-PDUs belonging to a particular SAR-SDU. The
MID field assists in the interleaving of SAR-PDUs from different SARSDUs and reassembly of these SAR-SDUs.
SAR-PDU Payload Field 44 octets.. The SAR-SDU information is left
justified within the SAR-PDU payload field. The remaining octets of the
SAR-PDU payload field may be set to 0 and are ignored at the receiving
end.
Length Indication (LI) Field 6 bits.. The length indication field is binary
encoded with the number of octets of SAR-SDU information that are
included in the SAR-PDU payload field. The permissible values of the LI
field, depending on the coding of the ST field, are BOM, COM 44, EOM
4 . . . 44, 63, and SSM 8 . . . 44. The 63 value for an EOM is used in the
Abort-SAR-PDU.
CRC Field 10 bits.. The CRC-10 is used to detect bit errors in the SAR-PDU.
It is the remainder of the division modulo 2. by the generator polynomial
of the product of x 10 and the content of the SAR-PDU, including the
SAR-PDU header, SAR-PDU payload, and the LI field of the SAR-PDU
trailer. Each bit of the concatenated fields mentioned above are consid-
628
ered coefficients modulo 2. of a polynomial of degree 373. The CRC-10

generator polynomial is
G x . s 1 q x q x 4 q x 5 q x 9 q x 10
The result of the CRC calculation is placed with the least significant bit
right-justified in the CRC field.
AAL-5. This type of AAL was designed specifically to carry data traffic
typically found in todays LANs. AAL-5 evolved after AAL-3r4, which was
found to be too complex and inefficient for LAN traffic. Thus AAL-5 got the
name SEAL for simple and efficient AAL layer. Only a small amount of
overhead is added to the CPCS-PDU, and no extra overhead is added when
the AAL-5 segments them into SAR-PDUs. There is no AAL-level cell
multiplexing. In AAL-5 all cells belonging to an AAL-5 CPCS-PDU are sent
sequentially.
As with AAL-3r4, AAL-5 provides both message mode service and
streaming mode service. Both modes offer nonassured operation peer-to-peer.
The following points should be considered:
Integral CPCS-SPU may be delivered, lost or corrupted.

Lost and corrupted CPCS-SDUs are not corrected by retransmission.
There is a corrupted data delivery option.
Flow control may be provided as an option.
The CPCS has the following service characteristics:
Nonassured data transfer of user data frames with any length measured
in octets from 1 to 65,535 octets
The CPCS connection is established by management or by the control
plane
Error detection and optional indication bit error and cell loss or gain.
CPCS-SDU sequence integrity on each CPCS connection
CPCS-PDU Structure and Coding. The CPCS-PDU format is illustrated in

Figure 16.17. The CPCS functions require an 8-octet CPCS-PDU trailer. The
trailer is always located in the last 8 octets of the last SAR-PDU of the
CPCS-PDU. Therefore, a padding field provides for a 48-octet alignment of
the CPCS-PDU. The trailer together with the passing field and the CPCSPDU payload comprise the CPCS-PDU. The coding of the CPCS-PDU is
discussed in the following text. Refer to Figure 16.17..
CPCS-PDU Payload. The payload is used to carry to CPCS-SDU. The
field is octet-aligned and can range from 1 to 65,535 octets in length.
629
630
Padding PAD. Field. Between the end of the CPCS-PDU payload and
the CPCS-PDU trailer, there will be from 0 to 47 unused octets. These
unused octets are called the padding PAD. field. They are strictly used
as filler octets and do not convey any information. Any coding is
acceptable. This padding field complements the CPCS-PDU including
CPCS-PDU payload, padding field, and CPCS-PDU trailer . to an
integral multiple of 48 octets.
CPCS User-to-User Indication CPCS-UU . Field. The CPCS-UU field is
used to transparently transfer CPCS user-to-user information.
Common Part Indicator CPI . Field. One of the functions of the CPI field
is to align the CPCS-PDU trailer to 64 bits. Other functions are for
further study by the ITU-T Organization. When only the 64-bit alignment function is used, this field is coded as zero.
Length Field. The length field is used to encode the length of the CPCSPDU payload field. The length field value is also used by the receiver to
detect the loss or gain of information. The length is binary encoded as
number of octets. A length field coded as zero is used for the abort
function.
CRC Field. The CRC-32 is used to detect bit errors in the CPCS-PDU.
The CRC field is filled with the value of a CRC calculation which is
performed over the entire contents of the CPCS-PDU, including the
CPCS-PDU payload, the PAD field, and the first 4 octets of the
CPCS-PDU trailer. The CRC field shall contain the 1 complement of
the sum modulo 2. of:
1. The remainder of x k x 31 q x 30 q qx q 1. divided modulo 2. by
the generator polynomial, where k is the number of bits of the
information over which the CRC is calculated and
2. The remainder of the division modulo 2. by the generator polynomial
of the product of x 32 by the information over which the CRC is
calculated.
The CRC-32 generator polynomial is
G x . s x 32 q x 26 q x 23 q x 22 q x 16 q x 12 q x 11 q x 10 q x 8
q x7 q x 5 q x 4 q x 2 q x q 1
The result of the CRC calculation is placed with the least significant bit
right justified in the CRC field.
As a typical implementation at the transmitter, the initial content of the
register of the device computing the remainder of the division is preset to all
631
1s and is then modified by division by the generator polynomial on the

information over which the CRC is to be calculated; the ones complement of
the resulting remainder is put into the CRC field.
As a typical implementation at the receiver, the initial content of the
register of the device computing the remainder of the division is preset to all
1s. The final remainder, after multiplication by x 32 and then division modulo
2. by the generator polynomial of the serial incoming CPCS-PDU, will be
in the absence of errors.
C x . s x 31 q x 30 q x 26 q x 25 q x 24 q x 18 q x 15 q x 14 q x 12 q x 11
q x 10 q x 8 q x 6 q x 5 q x 4 q x 3 q x q 1
SAR-PDU Structure and Coding. The SAR sublayer utilizes the ATM-Userto-ATM-User indication AUU. parameter of the ATM layer primitives the
relationship between the AUU parameter and the ATM layer PTI encoding
is defined in Section 2.2.4 of ITU-T Rec. I.361. to indicate that an SAR-PDU
contains the end of an SAR-SDU. An SAR-PDU where the value of the
AUU parameter is 1 indicates the end of an SAR-SDU; the value of 0
indicates the beginning or continuation of an SAR-PDU. The structure of the
SAR-PDU is shown in Figure 16.18.
Figure 16.19 illustrates the formation segmentation. of AAL-5 ATM cells
from CPCS user data.
The following are Radia Perlmans comments on AAL-5: So indeed
AAL5 is superior in every way to AAL3r4. Because it supports variable bit
rate, it can also do everything anyone would have wanted to do with AAL2.
So almost certainly the only AALs that will survive are AAL1 and AAL5.
Figure 16.18. SAR-PDU format for AAL-5. (From Figure 4rI.363.5, page 9, ITU-T Rec. I.363.5,
Ref. 13.)
632
Figure 16.19. Segmentation and reassembly of a CPCS user PDU. (From Figure B.3rI.363.5,
page 20, ITU-T Rec. I.363.5, Ref. 13.)
SERVICES: CONNECTION-ORIENTED AND CONNECTIONLESS
16.7
633
The issues such as routing decisions and architectures have a major impact
on connection-oriented services, where B-ISDNrATM end nodes have to
maintain or get access to lookup tables which translate destination addresses
into circuit paths. These circuit path lookup tables that differ at every node
must be maintained in a quasi-real-time fashion. This will have to be done by
some kind of routing protocol.
One way to resolve this problem is to make it an internal network problem
and use a connectionless service as described in ITU-T Rec. I.364 Ref. 14..
We must keep in mind that ATM is basically a connection-oriented service.
Here we are going to adapt it to provide a connectionless service.
16.7.1
Functional Architecture
The provision of connectionless data service in the B-ISDN is carried out by

means of ATM switches and connectionless service functions CLSF.. ATM
switches support the transport of connectionless data units in the B-ISDN
between specific functional groups where the CLSF handles the connectionless protocol and provides for the adaptation of the connectionless data units
into ATM cells to be transferred in a correction-oriented environment. As
shown in Figure 16.20, CLSF functional groups may be located outside the
B-ISDN, in a private connectionless network, or in a specialized service
provider or inside the B-ISDN itself.
The ATM switching is performed by the ATM nodes ATM switchrcrossconnect. that are a functional part of the ATM transport network. The CLSF
functional group terminates the B-ISDN connectionless protocol and includes functions for the adaptation of the connectionless protocol to the
intrinsically connection-oriented ATM layer protocol. These latter functions
are performed by the ATM adaptation layer Type 3r4 AAL-3r4., while the
CLSF group terminations are carried out by the services layer above the
AAL called the connectionless network access protocol CLNAP.. The CL
protocol includes functions such as routing, addressing, and QoS selection. In
order to perform the routing of CL data units, the CLSF has to interact with
the controlrmanagement planes of the underlying ATM network.
The general protocol structure for the provision of connectionless data
service is shown in Figure 16.21. Figure 16.22 illustrates the protocol architecture for supporting connectionless layer service. The CLNAP layer uses
the Type 3r4 AAL unassured service and includes the necessary functionality to provide the connectionless layer service. The connectionless layer
service provides for transparent transfer of variable size data units from a
source to one or more destinations in a manner such that lost or corrupted
data units are not retransmitted. The transfer is performed using a connectionless technique, including embedding destination and source addresses
into each data unit.
634
Figure 16.20. Reference configuration for connectionless data service in the B-ISDN. CLSF,
connectionless service functions. P, M, S, and T are reference points. (From Figure 1rI.364,
page 2, ITU-T Rec. I.364, Ref. 14.)
Figure 16.21. General protocol structure for connectionless data service in the B-ISDN.
CLNAP, connectionless network access protocol; CLSF; connectionless service functions.
635
Figure 16.22. Protocol architecture for supporting connectionless service.
16.7.2
CLNAP Protocol Data Unit (PDU) Structure and Encoding
Figure 16.23 shows detailed structure of the CLNAP-PDU, which contains

the following fields:
Destination Address and Source Address. These 8-octet fields each contain a
4-bit address type subfield, followed by the 60-bit address subfield. The
address type subfield indicates whether the address subfield contains a
publicly administered 60-bit individual address or a publicly administered
60-bit group address. The address subfield indicates to which CLNAP
entityies. the CLNAP-PDU is destined; and in the case of the source
address, it indicates the CLNAP entity that sourced the CLNAP-PDU.
The encoding of the address type and address subfields are shown in
Figures 16.24 and 16.25. The address is structured in accordance with
ITU-T Rec. E.164 Ref. 15..
Higher-Layer Protocol Identifier (HLPI). This 6-bit field is used to identify
the CLNAP user layer entity which the CLNAP-SDU is to be passed to at
the destination node. It is transparently carried end-to-end by the network.
PAD Length. This 2-bit field gives the length of the PAD field 03 octets..
The number of PAD octets is such that the total length of the user-information field and the PAD field together is an integral multiple of 4 octets
32 bits..
CRC Indication Bit (CIB). This 1-bit field indicates the presence if CIB s 1.
or absence if CIB s 0. of a 32-bit CRC field.
Header Extension Length (HEL). This 3-bit field can take on any value from
0 to 5 and indicates the number of 32-bit words in the header extension
field.
Reserved. This 16-bit field is reserved for future use. Its default value is 0.
Header Extension. This variable-length field can range from 0 to 20 octets.
Its length is indicated by the value of the header extension length field
see above.. In the case where the HEL is not equal to zero, all unused
octets in the header extension are set to zero. The information carried in
636
Figure 16.23. Structure of the CLNAP-PDU. ( nb is the length of a field, n, in bits.) (From
Figure 5rI.364, page 7, ITU-T Rec. I.364, Ref. 14.)
Figure 16.24. Destination address field.
Figure 16.25. Source address field.
ASPECTS OF A B-ISDNrATM NETWORK
637
the header extension is structured into information entities. An information entity element. consists of in this order. element length, element
type, and element payload, described as follows:
Element Length. This is a 1-octet field and contains the combined lengths
of the element length, element type, and element payload in octets.
Element Type. This is also a 1-octet field and contains a binary coded
value which indicates the type of information found in the element
payload field.
Element Payload. This is a variable-length field and contains the information indicated by the element type field.
User Information. This is a variable-length field up to 9188 octets and is
used to carry the CLNAP-SDU.
PAD. This field is 0, 1, 2, or 3 octets in length and is coded as all zeros.
Within each CLNAP-PDU the length of this field is selected such that the
length of the resulting CLNAP-PDU is aligned on a 32-bit boundary.
CRC. This optional 32-bit field may be present or absent as indicated by the
CIB field. The field contains the result of a standard CRC-32 calculation
performed over the CLNAP-PDU with the reserved field always treated
as if it were coded as all zeros.
16.8
16.8.1
ASPECTS OF A B-ISDNr
r ATM NETWORK
ATM Routing and Switching
An ATM transmission path supports VPs virtual paths., and inside VPs are
VCs virtual channels . as shown in Figure 16.26.
As we discussed in Section 16.4.1, each ATM cell contains a label in its
header to explicitly identify the VC to which the cell belongs. This label
consists of two parts: a virtual channel identifier VCI. and a virtual path
identifier VPI.. See Figure 16.5.
16.8.1.1 Virtual Channel Level. Virtual channel VC. is a generic term
used to describe a unidirectional communication capability for the transport
of ATM cells. A VCI identifies a particular VC link for a given virtual path
Figure 16.26. Relationship between the VC, VP, and the transmission path.
638
connection VPC.. A specific value of VCI is assigned each time a VC is

switched in the network. A VC link is a unidirectional capability for the
transport of ATM cells between two consecutive ATM entities where
the VCI value is translated. A VC link is originated or terminated by the
assignment or removal of the VCI value.
Routing functions of virtual channels are done at a VC switchrcross-connect.* The routing involves translation of the VCI values of the incoming VC
links into the VCI values of the outgoing VC links.
Virtual channel links are concatenated to form a virtual channel connection VCC.. A VCC extends between two VCC endpoints or, in the case of
point-to-multipoint arrangements, more than two VCC endpoints. A VCC
endpoint is the point where the cell information field is exchanged between
the ATM layer and the user of the ATM layer service.
At the VC level, VCCs are provided for the purpose of useruser,
usernetwork, or networknetwork information transfer. Cell sequence integrity is preserved by the ATM layer for cells belonging to the same VCC.
16.8.1.2 Virtual Path Level. Virtual path VP. is a generic term for a
bundle of virtual channel links; all the links in a bundle have the same
endpoints.
A VPI identifies a group of VC links, at a given reference point, that share
the same VPC. A specific value of VPI is assigned each time a VP is switched
in the network. A VP link is a unidirectional capability for the transport of
ATM cells between two consecutive ATM entities where the VPI value is
translated. A VP link is originated or terminated by the assignment or
removal of the VPI value.
Routing functions for VPs are performed at a VP switchrcross-connect.
This routing involves translation of the VPI values of the incoming VP links
into the VPI values of the outgoing VP links. VP links are concatenated to
form a VPC. A VPC extends between two VPC endpoints; however, in the
case of point-to-multipoint arrangements, there are more than two VPC
endpoints. A VPC endpoint is the point where the VCIs are originated,
translated, or terminated. At the VP level, VPCs are provided for the
purpose of useruser, usernetwork, and networknetwork information
transfer.
When VPCs are switched, the VPC supporting the incoming VC links are
terminated first and a new outgoing VPC is then created. Cell sequence
integrity is preserved by the ATM layer for cells belonging to the same VPC.
Thus cell sequence integrity is preserved for each VC link within a VPC.
*A VC cross-connect is a network element which connects VC links. It terminates VPCs and
translates VCI values and is directed by management plane functions, not by control plane
functions.
ASPECTS OF A B-ISDNrATM NETWORK
639
Figure 16.27. Representation of the VP and VC switching hierarchy. (From Figure 4rI.311,
page 5, ITU-T Rec. I.311, Ref. 16.)
640
Figure 16.27 is a representation of the VP and VC switching hierarchy

where the physical layer is the lowest layer composed of, from bottom up, a
regenerator section level, digital section level, and transmission path level.
The ATM layer resides just above the physical layer and is composed of the
VP level and, just above that, the VC level.
16.9
SIGNALING REQUIREMENTS
16.9.1
Setup and Release of VCCs
The setup and release of VCCs at the UNI can be performed in various ways:
Without using signaling procedures. Circuits are set up at subscription

with permanent or semipermanent connections.
By using metasignaling procedures where a special VCC is used to
establish or release a VCC used for signaling. Metasignaling is a simple
protocol used to establish and remove signaling channels. All information interchanges in metasignaling are carried out via single cell messages.
By employing user-to-network signaling procedures such as a signaling
VCC to establish or release a VCC used for end-to-end connectivity.
By employing user-to-user signaling procedures such as a signaling VCC
to establish or release a VCC within a preestablished VPC between two
UNIs.
16.9.2
Signaling Virtual Channels
16.9.2.1 Requirements for Signaling Virtual Channels. For a point-topoint signaling configuration, the requirements for signaling virtual channels
are as follows:
1. One virtual channel connection in each direction is allocated to each
signaling entity. The same VPIrVCI value is used in both directions. A
standardized VCI value is used for point-to-point signaling virtual
channel SVC..
2. In general, a signaling entity can control, by means of associated
point-to-point SVCs, user-VCs belonging to any of the VPs terminated
in the same network element.
3. As a network option, the user-VCs controlled by a signaling entity can
be constrained such that each controlled user-VC is in either upstream
or downstream VPs containing the point-to-point SVCs of the signaling
entity.
641
For point-to-multipoint signaling configurations, the requirements for signaling VCs are as follows:
1. Point-to-Point Signaling Virtual Channel. For point-to-point signaling,
one VC connection in each direction is allocated to each signaling
entity. The same VPIrVCI value is used in both directions.
2. General Broadcast Signaling Virtual Channel. The general broadcast
signaling virtual channel GBSVC. may be used for call offering in all
cases. In cases where the point does not implement service profiles
or where the multipoints do not support service profile identification,
the GBSVC is used for call offering. The specific VCI value for general
broadcast signaling is reserved per VP at the UNI. Only when metasignaling is used in a VP is the GBSVC activated in the VP.
3. Selectie Broadcast Signaling Virtual Channels. Instead of the GBSVC, a
VC connection for selective broadcast signaling SBS. can be used for
call offering, in cases where a specific service profile is used. No other
uses for SBSVCs are foreseen.
16.9.3
Metasignaling
16.9.3.1 Metasignaling Functions at the User Access. In order to establish, check, and release point-to-point and selective broadcast signaling
virtual channel connections, metasignaling procedures are provided. For each
direction, metasignaling is carried in a permanent virtual channel connection
having a standardized VCI value. This channel is called the metasignaling
virtual channel. The metasignaling protocol is terminated in the ATM layer
management entity.
The metasignaling function will be required to:
Manage the allocation of capacity to signaling channels

Establish, release, and check the status of signaling channels
Provide a means to associate a signaling endpoint with a service profile
if service profiles are supported
Provide the means to distinguish between simultaneous requests
It may be necessary to support metasignaling on any VP. Metasignaling can

only control signaling VCs within its VP.
16.9.3.2 Relationship Between Metasignaling and the User Access
Signaling Configuration. A point-to-multipoint signaling configuration exists when the network supports more than one signaling entity at the user
side. In this configuration, terminals must use the metasignaling protocol to
request allocation of their individual point-to-point signaling virtual channels.
642
A point-to-point signaling configuration exists when the network supports

only one signaling entity at the user side. When this configuration is known,
terminals can use the specific VCI value reserved for the point-to-point
signaling virtual channel. In this case no broadcast signaling virtual channel
will be provided.
In a user-to-user signaling configuration, the metasignaling protocol can
optionally be used over a user-to-user VPC to manage a user-to-user signaling virtual channel. It is recommended but it is a user choice. to use the
standardized VCI value for the user-to-user metasignaling channel. In this
case, metasignaling shall not have an impact on the network.
The metasignaling protocol can be used to manage signaling virtual
channel between a user and another network over the same user access. In
this case, VPIs other than VPI s 0 are used and the VCI value is the
standardized one.
Normally, the VPI value equal to zero is used to manage signaling virtual
channels to the local exchange. For the case where communication of a user
to an alternative local exchange over the same user access is required,
another VPI value different from zero will be used.
16.9.3.3 Metasignaling Functions in the Network. metasignaling is not
used to assign signaling channels between two network signaling endpoints.
Therefore, every VP within the network has one VCI value reserved for
point-to-point signaling and activated in case signaling is used on this VP.
16.9.3.4 Metasignaling Requirements
a. Scope of Metasignaling. A metasignaling virtual channel is able to
manage signaling virtual channels only within its own VP pair. In
VPI s 0, the metasignaling virtual channel is always present and has a
standardized VCI value.
b. Initiation of Metasignaling for SVCI Assignment. Metasignaling VC may
be activated at VP establishment. Other possibilities are for further
study. The signaling virtual channel SVC. should be assigned and
removed when necessary.
c. Metasignaling VCI and VP. A specific VCI value for metasignaling is
reserved per VP at the UNI. For a VP with point-to-multipoint
signaling configuration, metasignaling is required and the metasignaling VC within this VP will be activated. For a VP in point-to-point
signaling configuration, the use of metasignaling is for further study.
d. SVC Bandwidth. The user should have the possibility to negotiate the
bandwidth parameter value. The bandwidth parameter values are for
further study.
e. Metasignaling-Virtual Channel MSVC . Bandwidth. MSVC has default
bandwidth value. The bandwidth can be changed by mutual agreement
between network operator and user. The default value is for further
study.
16.9.4
643
Practical Signaling Considerations
There are several network interfaces, each with its own signaling variant. The
following interfaces must be taken into account:
Usernetwork interface UNI.

Interface with the public ATM network
Interface with a private network
Network-to-network interface NNI. e.g., interface between ATM
switches . Note: This interface is often referred to as the B-ICI or
broadband intercarrier interface. It loosely resembles a Signaling
System No. 7 SS7. interface, and the signaling protocols derive from
ISDN SS7 procedures.
Great care must be taken when specifying interfaces. There are two
recognized standards-making bodies:
ITU
ATM Forum
There are various incompatibilities.

When we start to dig into practical signaling procedures, we will ask:
Havent we been over this ground before? The answer is a resounding yes. In
particular, in Chapter 12 dealing with frame relay, many of the protocol
elements are nearly identical; certainly sets of procedures are identical, such
as setting up PVCs and SVCs.
The ATM signaling protocol at the UNI is derived from the narrowband
ISDN call control procedures described in ITU-T Rec. Q.931. The applicable
protocol for ATM is Q.2931, which defines Digital Subscriber Signaling
System No. 2 DSS2. Ref. 17.. The equivalent ATM Forum Specification is
UNI 3.0. Subsequently, the ATM Forum published version 3.1 to fix the
disparities there were with Q.2931. However, they are still not identical.
The P-NNI private networknetwork interface . is exclusively defined by
the ATM Forum because it is out of scope for the ITU. It is called the
Interim Interswitch Signaling Protocol or IISP Ref. 18.. It uses UNI 3.1 with
some minor procedural differences. There is no routing protocol specified.
Instead, routing information is configured manually and statically at each
private network station.
At the B-ICI signaling interface, ITU-T Recs. Q.2761 and Q.2764 Refs.
19 and 20. refer. The ITU protocol at this interface is commonly known as
BISUP, where the name is taken from SS7 ISDN User Part or ISUP. The
ATM Forum has defined an equivalent specification called the B-ICI Specification Version 2.0 Ref. 21..
16.9.4.1 ATM Addressing. Similar to frame relay signaling, there are two
recognized address formats: NSAP and E.164. We are familiar with E.164
from frame relay; this is the recommended international telephone address-
644
ing scheme. An address consists of two parts: network prefix 13 bytes. and
the user part 7 bytes.. The addressing formats are shown and compared in
Figure 16.28. Some of the standards involved use other terminology. The first
part is called the international domain identifier IDI., and the second part is
the domain-specific part DSP..
The AFI is the first octet digit. and is common to all three. AFI stands for
Authority and Format Identifier and indicates which address format is being
used. The DCC data country code. specifies the country to which the
address is registered as defined in ISO 3166.
Within each one of these domains, there is the domain-specific part or
DSP i.e., the remainder of the address .. The remaining formats are identical
for both NSAP address types whether DCC or ICD. The DSP format
identifier DFI. specifies the meaning of the remainder of the address. The
Administrative Authority field identifies the organization that is responsible
for administering the remainder of the address. This may be a carrier U.S.
terminology., private network, or manufacturer. The remainder of the DSP is
identical for all domains.
The routing domain RD. identifies a unique domain within the IDI
format. This is further subdivided using a 2-octet AREA field. The end
system identifier ESI. and SELector SEL. portions of the DSP are identical
for all IDI Ref. ISO 10589.. The ESI can be globally unique such as an
IEEE 802 MAC address 48 bits.. The SELector SEL. field is not used for
routing but may be used by end-users or end systems ES..
The E.164 numbering plan is in BCD code and is padded with zeros on
the left-hand side to the standard 15 digits to correspond to the standards
Figure 16.28. ATM addressing plans, comparison. AFI, authority and format identifier; DCC,
data country code; IDI, international domain identifier; DFI, DSP format identifier; RD, routing
domain; ESI, end system identifier; SEL, SELector; DSP, domain-specific part. derived from
ATM Forum, version 3.0 specification (Ref. 22.)
645
requirements. This is the international ISDN number. This is broken down to

a country code CC. of one to three digits. The remainder of the address is a
nationally significant number NSN.. The NSN breaks down to a national
destination code NDC. and a subscriber number SN..
Addresses are registered throughout an ATM network using the interim
local management interface ILMI.. It is used to convey configuration and
management information across the UNI over a dedicated virtual channel
VPI s 0, VCI s 16.. This protocol uses the Simple Network Management
Protocol SNMP. with a management information base MIB. describing the
ATM UNI. For SNMP, see Chapter 18.. The ILMI procedures are used to
register end-station ATM addresses. Each end system e.g., ATM network
interface card, ATM-LAN bridge, and so forth. is expected to know its own
ESI value.
16.9.4.2 Circuit Setup Using ATM Signaling. The procedures described
below are very similar to those discussed in Section 12.12.2. Figure 16.29
shows a typical exchange of messages for call establishment. The discussion
below is based on ITU-T Q.2931 Ref. 12. and Ref. 13..
Connection Establishment. A call is initiated at the end-user side by sending
a SETUP message across the UNI. A SETUP message has 24 information
Figure 16.29. Connection establishment.
646
elements. Several of the key information elements are briefly described as

follows:
AAL parameters: AAL type, CBR bit rate, CPCS-SDU size, source
clock frequency
Broadband bearer capability: traffic type, timing requirements, CBR, or
VBR
Called party number: addressing information described above. NSAP or
E.164
Calling party number using ATM address of user originating the call
End-to-end transit delay: indicates nominal maximum delay acceptable
Quality of service QoS.: indicates QoS class backwards and forwards
Transit network selection: indicates the requested public carrier network
for call e.g., Sprint, MCI, BT, AT & T, etc...
When the SETUP message is received by the network, it is acknowledged

with a CALL PROCEEDING message. This message contains the connection identifier employed to notify the user which VPIrVCI will be used for
the connection. The SETUP message continues to pass through the network,
eventually arriving at the destination terminal with all the information
elements IEs. the originating user specified, plus the connection identifier.
Keep in mind that the network always chooses the VPIrVCI values for the
connection.
Once the called user accepts the call, it replies sending a CONNECT
message. It follows back across the network from UNI, thence from switch to
switch through a UNI to the originator. Note in Figure 16.29 that there is
also a CONNECT ACKNOWLEDGE exchange. To identify this call from
other calls that may be in progress, a unique call reference value is assigned.
There is equivalent NNI signaling between switches on the routing of the
call. When signaling messages cross the public UNI, the network switch
continues the calls progress by sending the appropriate NNI messages. The
protocol involved here is known as the BISUP. Breaking out the acronym, we
have B for broadband as in B-ISDN. and ISUP, which is an SS7 term for
ISDN User Part. BISUP messages have a different format than equivalent
UNI signaling messages, although with similar objectives. BISUP messages
have parameters whereas UNI messages have information elements. Routing
through the network is determined by the SETUP message with two of its
IEs information elements.: transit network selection and called party number. The originating switch then sends an IAM* initial address message.
which sets up the path for traffic.
*Note the similarity to Signaling System No. 7.
QUALITY OF SERVICE (QoS)
16.10
16.10.1
647

ATM Service Quality Review
A basic performance measure for any digital data communication system is

bit error rate BER.. Well-designed fiber-optic links predominate in the
national network all the way down to the local serving exchange. We may
expect BERs from such links between 1 = 10y10 and 1 = 10y12 ; with end-toend performance better than 1 = 10y9. Thus other performance issues may
dominate the scene. These may be called ATM-unique QoS items, namely:
Cell transfer delay

Cell delay variation
Cell loss ratio
Mean cell transfer delay
Cell errored ratio
Severely error cell block ratio
Cell misinsertion rate
16.10.2
Definitions
Cell Event
1. A cell exit event occurs when the first bit of an ATM cell has
completed transmission out of an end-user device to a private ATM
network element across the private UNI measurement point, or out of
a private ATM network element to a public ATM network element
across the public UNI measurement point, or out of an end-user device
to a public ATM network across the public UNI measurement point.
2. A cell entry event occurs when the last bit of an ATM cell has
completed transmission into an end-user device from a private ATM
network element across the private UNI measurement point, or into a
private ATM network from a public ATM network element across the
public UNI measurement point, or into an end-user device from a
public ATM network element across the public UNI measurement
point.
ATM Cell Transfer Outcome. The following are possible cell transfer outcomes between measurement points for transmitted cells ITU-T definitions .:
1. Successful Cell Transfer Outcome. The cell is received corresponding to
the transmitted cell with a specified time Tmax . The binary content of
the received cell conforms exactly to the corresponding cell payload,
648
2.
3.
4.
5.
and the cell is received with a valid header field after header error
control procedures are completed.
Errored Cell Outcome. The cell is received corresponding to the transmitted cell within a specified time Tmax . The binary content of the
received cell payload differs from that of the corresponding transmitted
cell or the cell is received with an invalid header field after the header
error control procedures are completed.
Lost Cell Outcome. No cell is received corresponding to the transmitted
cell within a specified time Tmax examples: never showed up or
late..
Misinserted Cell Outcome. This involves a received cell for which there
is no corresponding transmitted cell.
Seerely Errored Cell Block Outcome. This is when M or more lost cell
outcomes, misinserted cell outcomes, or errored cell outcomes are
observed in a receiver cell block of N cells transmitted consecutively on
a given connection.
16.10.3
Cell Transfer Delay
Cell transfer delay is defined as the elapsed time between a cell exit event at
measurement point 1 e.g., at the source UNI. and the corresponding cell
entry event at measurement point 2 e.g., the destination UNI. for a particular connection. The cell transfer delay between two measurement points is
the sum of the total inter-ATM node transmission delay and the total ATM
node processing delay between MP1 and MP2 .
In addition to the normal delay that one would expect for a cell to traverse
a network, extra delay is added in the ATM network at each ATM switch.
One cause of this delay is asynchronous digital multiplexing. Where this
method is employed, two cells directed toward the same output port of an
ATM switch or cross-connect can result in contention.
One or more cells are held in a buffer until the contention is resolved.
Thus the second cell suffers additional delay. Delay of a cell depends on the
traffic intensity within a switch which influences the probability of contention.
The asynchronous path of each ATM cell also contributes to cell delay.
Cells can be delayed one or many cell periods, depending on traffic intensity,
switch sizing, and the transmission path taken through the network.
16.10.4
Cell Delay Variation
ATM traffic, by definition, is asynchronous, magnifying transmission delay.

Delay is also inconsistent across the network. It can be a function of time
i.e., a moment in time., network designrswitch design such as buffer size.,
and traffic characteristics at that moment of time. The result is cell delay
variation CDV..
649
CDV can have several deleterious effects. The dispersion effect, or

spreading out, of cell interarrival times can impact signaling functions or the
reassembly of cell user data. Another effect is called clumping. This occurs
when the interarrival times between transmitted cells shorten. One can
imagine how this could affect the instantaneous network capacity and how it
can impact other services using the network.
There are two performance parameters associated with CDV: 1-point
CDV and 2-point CDV.
The 1-point CDV describes variability in the pattern of cell arrival events
observed at a single measurement point with reference to the negotiated
peak rate 1rT as defined in ITU-T Rec. I.371 Ref. 23..
The 2-point CDV describes variability in the pattern of cell arrival events
as observed at the output of a connection portion MP2 . with reference to the
pattern of the corresponding events observed at the input to the connection
portion MP1 ..
16.10.5
Cell Loss Ratio
Cell loss may not be uncommon in an ATM network. There are two basic
causes of cell loss: error in cell header or network congestion.
Cells with header errors are automatically discarded. This prevents misrouting of errored cells, as well as the possibility of privacy and security
breaches.
Switch buffer overflow can also cause cell loss. It is in these buffers that
cells are held in prioritized queues. If there is congestion, cells in a queue
may be discarded selectively in accordance with their level of priority. Here
enters the cell loss priority CLP. bit discussed in Section 16.4. Cells with this
bit set to 1 are discarded in preference to other, more critical cells. In this
way buffer fill can be reduced to prevent overflow.
Cell loss ratio is defined for an ATM connection as
Lost cells . r Total transmitted cells .

Lost and transmitted cells counted in severely errored cell blocks should be
excluded from the cell population in computing cell loss ratio.
16.10.6
Mean Cell Transfer Delay
Mean cell transfer delay is defined as the arithmetic average of a specified

number of cell transfer delays for one or more connections.
16.10.7
Cell Error Ratio
Cell error ratio is defined as follows for an ATM connection:
Errored cells . r Successfully transferred cells q Errored cells .
650
Successfully transferred cells and errored cells contained in cell blocks

counted as severely errored cell blocks should be excluded from the population used in calculating cell error ratio.
16.10.8
Severely Errored Cell Block Ratio
The severely errored cell block ratio for an ATM connection is defined as
Severely errored cell blocks . r Total transmitted cell blocks .

A cell block is a sequence of N cells transmitted consecutively on a given
connection. A severely errored cell block outcome occurs when more than M
errored cells, lost cells, or misinserted cell outcomes are observed in a
received cell block.
For practical measurement purposes, a cell block will normally correspond
to the number of user information cells transmitted between successive OAM
cells. The size of a cell block is to be specified.
16.10.9
Cell Misinsertion Rate
The cell misinsertion rate for an ATM connection is defined as
Misinserted cells . r Time interval .

Severely errored cell blocks should be excluded from the population when
calculating the cell misinsertion rate. Cell misinsertion on a particular connection is most often caused by an undetected error in the header of a cell
being transmitted on a different connection. This performance parameter is
defined as a rate rather than a ratio. because the mechanism producing
misinserted cells is independent of the number of transmitted cells received
on the corresponding connection.
Section 16.10 is based on ITU-T Rec. I.356, Ref. 24.
16.11
16.11.1
TRAFFIC CONTROL AND CONGESTION CONTROL

Generic Functions
The following functions form a framework for managing and controlling

traffic and congestion in ATM networks and are to be used in appropriate
combinations:
1. Network Resource Management NRM .. Provision is used to allocate
network resources in order to separate traffic flows in accordance with
service characteristics.
651
2. Connection Admission Control CAC .. This is defined as a set of actions

taken by the network during the call setup phase or during the call
renegotiation phase in order to establish whether a VC or VP connection request can be accepted or rejected, or whether a request for
reallocation can be accommodated. Routing is part of connection
admission control actions.
3. Feedback Controls. These are a set of actions taken by the network and
by users to regulate the traffic submitted on ATM connections according to the state of network elements.
4. Usage Network Parameter Control UPCrNPC .. This is a set of actions
taken by the network to monitor and control traffic, in terms of traffic
offered and validity of the ATM connection, at the user access and
network access, respectively. Their main purpose is to protect network
resources from malicious as well as unintentional misbehavior that can
affect the QoS of other already established connections by detecting
violations of negotiated parameters and taking appropriate actions.
5. Priority Control. The user may generate different priority traffic flows
by using the CLP. A congested network element may selectively discard
cells with low priority, if necessary, to protect as far as possible the
network performance for cells with higher priority.
Figure 16.30 is a reference configuration for traffic and congestion control.
16.11.2
Events, Actions, Time Scales, and Response Times
Figure 16.31 shows the time scales over which various traffic control and
congestion control functions can operate. The response time defines how
quickly the controls react. For example, cell discarding can react on the order
Figure 16.30. Reference configuration for traffic control and congestion control. Note: NPC
may apply as well at some intranetwork NNIs, and the arrows indicate the direction of the cell
flow. UPC, usage parameter control; CAC, connection admission control; PC, priority control;
NPC, network parameter control; RM, resource management; Others, for further study. (From
Figure 1rI.371, page 7, ITU-T Rec. I.371, Ref. 23.)
652
Figure 16.31. Control response times. (From Figure 2rI.371, page 8, ITU-T Rec. I.371,
Ref. 23.)
of the insertion time of a cell. Similarly, feedback controls can react on the
time scale of round-trip propagation times. Because traffic control and
resource management functions are needed at different time scales, no single
function is likely to be sufficient.
16.11.3 Quality of Service, Network Performance,
and Cell Loss Priority
QoS at the ATM layer is defined by a set of parameters such as cell delay,
cell delay variation sensitivity, cell loss ratio, and so forth.
A user requests a specific ATM layer QoS from the QoS classes which a
network provides. This is part of the traffic contract at connection establishment. It is a commitment for the network to meet the requested QoS as long
as the user complies with the traffic contract. If the user violates the traffic
contract, the network need not respect the agreed-upon QoS.
A user may request at most two QoS classes for a single ATM connection,
which differ with respect to the cell loss ratio objectives. The CLP bit in the
ATM header allows for two cell loss ratio objectives for a given ATM
connection.
Network performance objectives at the ATM SAP are intended to capture
the network ability to meet the requested ATM layer QoS. It is the role of
653
upper layers, including the AAL, to translate this ATM layer QoS to any
specific application requested QoS.
16.11.4
Traffic Descriptors and Parameters
Traffic parameters describe traffic characteristics of an ATM connection.

Traffic parameters are grouped into source traffic descriptors for exchanging
information between the user and the network. Connection admission control
procedures use source traffic descriptors to allocate resources and derive
parameters for the operation of the user parameter controlrnetwork parameter control UPCrNPC..
We now define the terms traffic parameters and connection traffic
descriptors.
Traffic Parameters. A traffic parameter is a specification of a particular
traffic aspect. It may be qualitative or quantitative. Traffic parameters, for
example, may describe peak cell rate, average cell rate, burstiness, peak
duration, and source type such as telephone, videophone.. Some of these
traffic parameters are interdependent, such as burstiness with average and
peak cell rate.
Traffic Descriptors. The ATM traffic descriptor is the generic list of traffic
parameters which can be used to capture the intrinsic traffic characteristics of an ATM connection. A source traffic descriptor is the set of traffic
parameters belonging to the ATM traffic used during the connection setup
to capture the intrinsic traffic characteristics of the connection requested
by the source.
Connection Traffic Descriptor. This specifies the traffic characteristics of the
ATM connection at the public or private UNI. The connection traffic
descriptor is the set of traffic parameters in the source traffic descriptor,
CDV tolerance, and the conformance definition that is used to unambiguously specify the conforming cells of an ATM connection. Connection
admission control CAC. procedures will use the connection traffic descriptor to allocate resources and to derive parameter values for the
operation of the UPC. The connection traffic descriptor contains the
necessary information for conformance testing of cells of the ATM connection at the UNI.
Any traffic parameter and the CDV tolerance in a connection traffic
descriptor should fulfill the following requirements:
1. They should be understandable by the user or terminal equipment, and
conformance testing should be possible as stated in the traffic contract.
2. They should be useful in resource allocation schemes, meeting network
performance requirements as described in the traffic contract.
3. They should be enforceable by the UPC.
654
16.11.5
UserNetwork Traffic Contract
16.11.5.1 Operable Conditions. CAC and UPCrNPC procedures require

the knowledge of certain parameters to operate efficiently. For example, they
should take into account the source traffic descriptor, the requested QoS,
and the CDV tolerance defined below. in order to decide whether the
requested connection can be accepted.
The source traffic descriptor, the requested QoS for any given ATM
connection, and the maximum CDV tolerance allocated to the CEQ customer
equipment. define the traffic contract at the TB reference point see Figure
16.30.. Source traffic descriptors and QoS are declared by the user at
connection setup by means of signaling or subscription. Whether the maximum allowable CDV tolerance is also negotiated on a subscription or on a
per connection basis is for further study by the ITU-T organization.
The CAC and UPCrNPC procedures are operator specific. Once the
connection has been accepted, the value of the CAC and UPCrNPC parameters are set by the network on the basis of the network operators policy.
ITU-T Rec. I.371 Ref. 23. notes that all ATM connections handled by
network connection related functions CRF. have to be declared and enforced by the UPCrNPC. ATM layer QoS can only be assured for compliant
ATM connections. As an example, individual VCCs inside user end-to-end
VPC are neither declared nor enforced at the UPC and hence no ATM layer
QoS can be assured for them.
16.11.5.2 Source Traffic Descriptor, Quality of Service, and Cell Loss
Priority. If a user requests two levels of priority for an ATM connection, as
indicated by the CLP bit value, the intrinsic traffic characteristics of both cell
flow components have to be characterized in the source traffic descriptor.
This is by means of a set of traffic parameters associated with the CLP s 0
component and a set of traffic parameters associated with the CLP s 0 q 1
component.
16.11.5.3 Impact of Cell Delay Variation on UPCr
r NPC Resource Allocation. ATM layer functions such as cell multiplexing may alter the traffic
characteristics of ATM connections by introducing cell delay variation, as
shown in Figure 16.32. When cells from two or more ATM connections are
multiplexed, cells of a given ATM connection may be delayed while cells of
another ATM connection are being inserted at the output of the multiplexer.
Similarly, some cells may be delayed while physical layer overhead or OAM
cells are inserted. Therefore, some randomness affects the time interval
between reception of ATM-cell data requests at the endpoint of an ATM
connection and the time that an ATM-cell data indication is received at the
UPCrNPC. Besides, AAL multiplexing may cause CDV.
The UPCrNPC mechanism should not discard or tag cells in an ATM
connection if the source conforms to the source traffic descriptor negotiated
655
656
at connection establishment. However, if the CDV is not bounded at a point

where the UPCrNPC function is performed, it is not possible to design a
suitable UPCrNPC mechanism and to allocate resources properly. Therefore, it is required that a maximum allowable value of CDV be standardized
edge-to-edge e.g., between the ATM connection endpoint and TB , between
TB and an internetwork interface, and between internetwork interfaces; see
Figure 16.30..
UPCrNPC should accommodate the effect of the maximum CDV allowed
on ATM connections within the limit resulting from the accumulated CDV
allocated to upstream subnetworks including customer equipment CEQ..
Traffic shaping partially compensates for the effects of CDV on the peak cell
rate of the ATM connection. Examples of traffic-shaping mechanisms are
re-spacing cells of individual ATM connections according to their peak cell
rate or suitable queue service schemes.
Values of the CDV are network performance issues. The definition of a
source traffic descriptor and the standardization of a maximum allowable
CDV may not be sufficient for a network to allocate resources properly.
When allocating resources, the network should take into account the worstcase traffic passing through the UPCrNPC in order to avoid impairments to
other ATM connections. The worst-case traffic depends on the specific
implementation of the UPCrNPC. The tradeoffs between UPCrNPC complexity, worst-case traffic, and optimization of network resources are made at
the discretion of network operators. The quantity of available network
resources and the network performance to be provided for meeting QoS
requirements can influence these tradeoffs.
16.11.5.4 Cell Conformance and Connection Compliance. Conformance applies to the cells as they pass the UNI and are in principle tested
according to some combination of generic cell rate algorithms GCRAs.. The
first cell of the connection initializes the algorithm, and from then on each
cell is either conforming or not conforming. Because in all likelihood even
with the best intentions a cell or two may be nonconforming, thus it is
inappropriate for the network operator to only commit to the QoS objectives
for connections whose cells are conforming. Thus the term compliant, which
is not precisely defined, is used for connection in which some of the cells may
be nonconforming.
The precise definition of a compliant connection is left to the network
operator. For any definition of a compliant connection, a connection for
which all cells are conforming are identified as compliant.
Based on action of the UPC function, the network may decide whether a
connection is compliant or not. The commitment by the network operator is
to support the QoS for all connections that are compliant.
For compliant connections at the public UNI, the agreed QoS is supported
for at least the number of cells equal to the conforming cells according to the
conformance definition. For noncompliant connections, the network need
657
not respect the agreed QoS class. The conformance definition that defines
conformity at the public UNI of the cells of the ATM connection uses a
GCRA configuration in multiple instances to apply to particular combinations of the CLP s 0 and CLP s 1 q 0 cell streams with regard to the peak
cell rate and to particular combinations of CLP s 0, CLP s 1, and CLP s 0
q 1 cell streams with regard to the sustainable cell rate and burst tolerance.
For example, the conformance definition may use the GCRA twice, once for
peak cell rate of the aggregate CLP s 0 q 1. cell stream and once for the
sustainable cell rate of the CLP s 0 cell stream. The network operator may
offer a limited set of alternative conformance definitions all based on
GCRA. from which the user may choose for a given ATM connection.
16.11.5.5 Generic Cell Rate Algorithm (GCRA). The GCRA is a virtual
scheduling algorithm or a continuous-state leaky bucket algorithm as defined
by the flow chart in Figure 16.33. The GCRA is used to operationally define
the relationship between peak cell rate PCR. and the CDV tolerance and
the relationship between sustained cell rate SCR. and the burst tolerance.
In addition, for the cell flow of an ATM connection, the GCRA is used to
specify the conformance at the public or private UNI to declared values of
the above two tolerances, as well as declared values of the traffic parameters
PCR and SCR and burst tolerance.
For each cell arrival, the GCRA determines whether the cell is conforming
with the traffic contract of the connection, and thus the GCRA is used to
provide the formal definition of traffic conformance to the traffic contract.
Although traffic conformance is defined in terms of the GCRA, the network
provider is not obligated to use this algorithm or this algorithm with the
same parameter values. for the UPC. Rather, the network provider may use
any UPC as long as the operation of the UPC does not violate the QoS
objectives of a compliant connection.
The GCRA depends only on two parameters: the increment I and the
limit L. These parameters have been denoted by T and , respectively, and
in ITU-T Rec. I.371 Ref. 23., but have been given more generic labels by
the ATM Forum. herein because the GCRA is used in multiple instances.
We will now use the ATM Forum notation GCRA I, L., which means
the GCRA with the value of the increment parameter set equal to I and the
value of the limit parameter set equal to L.
The GCRA is formally defined in Figure 16.33, which the ATM Forum
took as a generic version of Figure 1 in Annex A of I.371. The two algorithms
in Figure 16.33 are equivalent in the sense that for any sequence of cell
arrival times w t a k ., k G 1x, the two algorithms determine the same cells to be
conforming and thus the same cells to be nonconforming. The two algorithms
are easily compared if one notices that at each arrival epoch, t a k ., and after
the algorithms have been executed, TAT s X q LCT see Figure 16.33..
The virtual scheduling algorithm updates a theoretical arrival time TAT.,
which is the nominal arrival time of the cell assuming equally spaced cells
658
Figure 16.33. Equivalent versions of the generic cell rate algorithm (GCRA). (From Figure
A-1rI.371, page 62, ITU-T Rec. I.371, Ref. 23.)
when the source is active. If the actual arrival time of a cell is not too early
relative to the TAT, in particular if the actual arrival time is after TAT-L,
then the cell is conforming; otherwise the cell is nonconforming.
The continuous-state leaky bucket algorithm can be viewed as a finitecapacity bucket whose real-valued content drains out at a continuous rate of
1 unit of content per time unit and whose content is increased by the
increment I for each conforming cell. Equivalently, it can be viewed as the
work load in a finite-capacity queue or as a real-valued counter. If at a cell
arrival the content of the bucket is less than or equal to the limit value, L,
then the cell is conforming; otherwise the cell is nonconforming. The capacity
of the bucket i.e., the upper bound on the counter. is L q I.
TRANSPORTING ATM CELLS
659
16.11.5.6 Traffic Contract Parameter Specification. Peak cell rate for

CLP s 0 q 1 is a mandatory traffic parameter to be explicitly or implicitly
declared in any source traffic descriptor. In addition to the peak cell rate of
an ATM connection, it is mandatory for the user to declare either explicitly
or implicitly the cell delay variation tolerance within the relevant traffic
contract.
Peak Cell Rate (PCR). The following definition applies to ATM connections
supporting both CBR and VBR services Ref. 23.:
The peak cell rate in the source traffic descriptor specifies an upper bound on
the traffic that can be submitted on an ATM connection. Enforcement of this
bound by the UPCrNPC allows the network operator to allocate sufficient
resources to ensure that the performance objectives e.g., for cell loss ratio. can
be achieved.
For switched ATM connections, the PCR for CLP s 0 q 1 and the QoS
class must be explicitly specified for each direction in the connection-establishment SETUP message.
The CDV tolerance must be either explicitly specified at subscription time
or implicitly specified.
The SCR and burst tolerance comprise an optional traffic parameter set in
the source traffic descriptor. If either SCR or burst tolerance is specified,
then the other must be specified within the relevant traffic contract.
16.12
16.12.1

In the DS3 Frame
One of the most popular higher-speed digital transmission systems in North

America is DS3 operating at a nominal transmission rate of 45 Mbps. It is
also being widely implemented for transport of SMDS. The system used to
map ATM cells into the DS3 format is the same as used for SMDS.
DS3 uses the physical layer convergence protocol PLCP. to map ATM
cells into its bit stream. A DS3 PLCP frame is shown in Figure 16.34. There
are 12 cells in a frame. Each cell is preceded by a 2-octet framing pattern
A1, A2., to enable the receiver to synchronize to cells. After the framing
pattern there is an indicator consisting of one of 12 fixed bit patterns used to
identify the cell location within the frame POI.. This is followed by an octet
of overhead information used for path management. The entire frame is then
padded with either 13 or 14 nibbles a nibble s 4 bits. of trailer to bring the
transmission rate up to the exact DS3 bit rate. The DS3 frame has 125- s
duration.
660
Figure 16.34. Format of DS3 PLCP frame. (From Ref. 25. Courtesy of Hewlett-Packard
Company.
DS3 has to contend with network slips addedrdropped frames to accommodate synchronization alignment.. Thus PLCP is padded with a variable
number of stuff justification . bits to accommodate possible timing slips. The
C1 overhead octet indicates the length of padding. The bit interleaved parity
BIP. checks the payload and overhead functions for errors and performance
degradation. This performance information is transmitted in the overhead.
Figure 16.35 illustrates a 680-bit DS3 frame transporting ATM cells. The
figure also shows 84 bits being grouped into 21 nibbles. Remember a nibble is
half an octet or byte; it is a group of 4 bits.
ATM cells can also be mapped into a DS3 frame by what is called
HEC-based mapping. In this case the ATM cells are mapped into the
payload with the octet structure of the cells aligned with the nibble structure
of the multiframe. The multiframe is organized such that 84 bits of payload
follow every overhead bit. The 84 bits can be assumed to be organized into 21
consecutive nibbles. The ATM cell is placed such that the start of cell always
coincides with the start of nibble. As in most ATM cell mapping situations,
the cells may cross multiframe boundaries. See Figure 16.35.
16.12.2
DS1 Mapping
One approach to mapping ATM cells into a DS1 frame is to use a similar
procedure as used on DS3 with PLCP. In this case only 10 cells are bundled
into a frame, and two of the Z overheads are removed. The padding in the
frame is set at 6 octets. The entire frame takes 3 ms to transmit and spans
661
662
Figure 16.36. DS1 mapping with PLCP. (From Ref. 25. Courtesy of Hewlett-Packard
Company.)
Figure 16.37. Frame structure for 1.544 Mbps (DS1) used to transport ATM cells. (From
Figure 2-1rG.804, page 8, ITU-T Rec. G.804, Ref. 26.)
many DS1 ESF extended superframe . frames. The mapping is shown in

Figure 16.36. One must consider the arithmetic. There are 24 time slots
octets. in a DS1 frame, which breaks down to 192 bits. There are 424
8 = 53. bits in an ATM cell. This, of course, leads to the second method of
carrying ATM cells in DS1 by directly mapping ATM cells octet for octet
time slots.. This is done in groups of 53 octets 1 cell. and would, by
necessity, cross DS1 frame boundaries to accommodate an ATM cell. See
Figure 16.37.
16.12.3
E1 Mapping
E1 PCM has a 2.048-Mbps transmission rate. An E1 frame has 256 bits

representing 32 channels or time slots, 30 of which carry traffic, and which
are available to carry ATM cells. Thus, a frame has 32 time slots TS. where
663
664
TS0 and TS16 are reserved. TS0 is used for synchronization of the frame and
TS16 is used for signaling. The E1 frame is shown in Figure 16.38. Bit
positions 9128 and 137256 may be used for ATM cell mapping. ATM cells
can also be directly mapped into special E3 and E4 frames. The first has 530
octets available for cells thus it fits 10 cells. and the second has 2160 octets
not evenly divisible..
Figure 16.39. Mapping ATM cells into the 155.520 Mbps frame structure of SDH-based UNI.
(From Figure 8rI.432.1, page 13, ITU-T Rec. I.432.1, Ref. 5.)
16.12.4
665
Mapping ATM Cells into SDH
16.12.4.1 At STM-1 (155.520 Mbps). SDH is described in Chapter 15.

Figure 16.39 shows the mapping procedure. The ATM cell stream is first
mapped into the C-4 and then mapped into the VC-4 container along with
the VC-4 path overhead. The ATM cell boundaries are aligned with the
STM-1 octet boundaries. Because the C-4 capacity 2340 octets. is not an
integer multiple of the cell length 53 octets., a cell may cross a C-4
boundary. The AU-4 pointer octets H1 and H2 in the SOH. is used for
finding the first octet of the VC-4.
16.12.4.2 At STM-4 (622.080 Mbps). As shown in Figure 16.40, the ATM
cell stream is first mapped into C-4-4c and then packed into the VC-4-4c
container along with the VC-4-4c path overhead. The cell boundaries are
Figure 16.40. Mapping ATM cells into the 622.080-Mbps frame structure for SDH-based UNI.
(From Figure 8rI.432.1, page 13, ITU-T Rec. I.432.1, Ref. 5.)
666
Figure 16.41. Mapping ATM cells into a SONET STM-1 frame at the UNI.
aligned with STM-4 octet boundaries. Because the C-4-4c capacity 9360
octets. is not an integer multiple of the cell length 53 octets., a cell may
cross a C-4-4c boundary. The AU pointers are used for finding the first octet
of the VC-4-4c.
16.12.5
Mapping ATM Cells into SONET
ATM cells are mapped directly into the SONET payload 49.54 Mbps.. As
with SDH, the payload in octets is not an integer multiple of the cell length,
and thus a cell may cross an STS frame boundary. This mapping concept is
shown in Figure 16.41. The H4 pointer indicates where the cells begin inside
an STS frame. Another approach is to identify cell headers and, thus, identify
the first cell in the frame.
16.12.6
Cell Rates for Various Typical Connections
Table 16.6 shows connection cell rates for various familiar user information
bit rate capacities. By choosing appropriate cell rate values for the traffic
parameters associated with a connection, a user defines the bit rate capacity
associated with an ATM connection. The cell rates listed in the table include
an allowance for the AAL header and operations, administration, and
maintenance OAM. cell overhead.
These rates are used to assign the peak cell rate PCR. and sustainable
cell rate SCR. traffic parameters defined in the ATM Forum UNI Specification Version 3.1.
ATM OPERATION, ADMINISTRATION, AND MAINTENANCE (OAM)
TABLE 16.6
667
Connection Cell Rates for Various Familiar Bit Rate Capacities
User Information
Bandwidth
N = 64 Kbps (DS0)
N = 384 Kbps
N = 1.544 Mbps (DS1)
N = 44.736 Mbps (DS3)
N = 1 Mbps
DS3 PLCP
DS3 direct mapping
Full STS-3c
Full STS-12c
Cell Rate
(Cells r Second)
N = 173
N = 1035
N = 4140
N = 119,910
N = 2689
96,000
104,268
353,207
1,412,828
Comments
N s 1 to 23 (i.e., sub-DS1 rates)
N s 1 to 31 (i.e., up to 12 Mbps)
N s 1, 2 (i.e., 1 or 2 = DS3 rate)
DS3 line rate, PLCP option
DS3 line rate, direct mapping (HEC)
STS-3c line rate
STS-12c line rate
Source: Table 9-1, page 9-1, Bellcore Special Report, SR-3330, Ref. 27.
16.13 ATM OPERATION, ADMINISTRATION,

AND MAINTENANCE (OAM)
16.13.1
OAM Levels in the B-ISDN
OAM functions in the network are performed on five OAM hierarchical

levels associated with the ATM and physical layers of the protocol reference
model. See Figures 16.2 and 16.3.. The functions result in corresponding
bidirectional information flows F1, F2, F3, F4, and F5 referred to as OAM
flows. Not all of these need to be present. The OAM functions of a missing
level are performed at the next higher level. The levels are as given in Table
16.7
TABLE 16.7
Level
OAM Levels in the B-ISDN (in Decreasing Order)

Description
F5 Virtual channel level: Extends between network elements performing virtual channel
connection termination functions and can extend through one or more path
connections.
F4 Virtual path level: Extends between network elements performing virtual path connection termination functions and may extend through one or more transmission paths.
F3 Transmission path level: Extends between network elements assemblingrdisassembling the payload of a transmission system and associating it with its OAM
functions. Cell delineation and header error control (HEC) functions are required at
the endpoints of each transmission path. The transmission path is connected
through one or more digital sections.
F2 Digital section level: Extends between section endpoints and comprises a maintenance entity according to the definition in ITU-T Rec. M.20, Section 3.
F1 Regenerator section level: A regenerator section is a portion of a digital section and as
such is a maintenance subentity.
668
OAM functions related to OAM levels are independent from the OAM
functions of other layers and have to be provided at each layer.
Each layer, where OAM functions are required, is able to carry out its own
processing to obtain quality and status information. OAM functions are
performed by the layer management. These results may be provided to the
plane management or to the adjacent higher layer. Higher-layer functions are
not necessary to support the OAM of the lower layer.
16.13.2
Mechanisms to Provide OAM Flows
16.13.2.1 Physical Layer Mechanisms. The physical layer contains the

three lowest OAM levels as outlined in Table 16.7. The allocation of the
OAM flows is as follows:
F1:
F2:
F3:
regenerator section level

digital section level
transmission path level
The mechanisms to provide OAM functions and to generate OAM flows F1,
F2, and F3 depend on the format of the transmission system as well as on the
supervision functions contained in B-NT1 and B-NT2 for the section crossing
the TB reference point. There are three types of transmission for customer
access: SDH-based, cell-based, and PDH-based.
SDH-Based Transmission Systems. Flows F1 and F2 are carried on bytes in
the section overhead SOH.; flow F3 is carried in the path overhead POH.
of the transmission frame.
Cell-Based Transmission Systems. Such transmission systems may use an
interface structure as specified in Section 4.2 of ITU-T Rec. I.432.1. OAM
flows F1 and F3 are carried through maintenance cells for the physical layer
using a specific pattern in the header for F1 and F3. F2 flows are not
provided, but the associated functions are supported by F3 flows. These cells
are not passed to the ATM layer. The occurrence of a PLOAM* cell is
determined by the requirements of the supported OAM functions. For each
type F1 and F3. of PL-OAM cell, maximum spacing is applied. If maximum
spacing is exceeded, loss of OAM flow LOM. will occur.
PDH -Based Transmission Systems. These systems may only be used on the
network side of the B-NT1. Specific means to monitor the section performance e.g., violation code counting CRC, etc.. are specified for these
systems. The capability to carry OAM information other than bit-oriented
messages is very limited.
*PLOAM or PL-OAM stands for physical layer OAM.
PDH stands for plesiochronous digital hierarchy e.g., E1, DS1, DS3, etc...
16.13.3
669
ATM Layer Mechanism: F4 and F5 Flows
The ATM layer contains the two highest OAM levels shown in Table 16.7.
These are:
F4:
F5:
virtual path level

virtual channel level
These OAM flows are provided by cells dedicated to ATM layer OAM
functions for both virtual channel connections VCC. and virtual path connections VPC.. In addition, such cells are usable for communication within
the same layers of the management plane.
16.13.3.1 F4 Flow Mechanism. The F4 flow is bidirectional. OAM cells
for the F4 flow have the same VPI value as the user cells of the VPC and are
identified by one or more preassigned VCI values. The same preassigned
VCI value shall be used for both directions of the F4 flow. The OAM cells
for both directions of the F4 flow must follow the same physical route so that
any connecting points supporting that connection can correlate the fault and
performance information from both directions.
There are two kinds of F4 flows, which can simultaneously exist in a VPC.
These are as follows:
End-to-End F4 Flow. This flow, identified by a standardized VCI see

Rec. I.361., is used for end-to-end VPC operations communications.
Segment F4 Flow. This flow, identified by a standardized VCI see Rec.
I.361., is used for communicating operations information within the
bounds of one VPC link or multiple interconnected VPC links, where
all of the links are under the control of one administration or organization. Such a concatenation of VPC links is called a VPC segment. A
VPC segment can be extended beyond the control of one administration
by mutual agreement.
F4 flows must be terminated only at the endpoints of a VPC or at the

connecting points terminating a VPC segment. Intermediate points i.e.,
connecting points. along the VPC or along the VPC segment may monitor
OAM cells passing through them and insert new OAM cells, but they cannot
terminate the OAM flow. The F4 flow will be initiated at or after connection
setup.
The administrationrorganization that controls the insertion of OAM cells
for operations and maintenance of a VPC segment must ensure that such
OAM cells are extracted before they leave the span of control of that
administrationrorganization Ref. 28..
670
16.13.3.2 F5 Flow Mechanism. The F5 flow is bidirectional. OAM cells

for the F5 flow have the same VCIrVPI values as the user cells of the VCC
and are identified by the payload type identifier PTI.. The same PTI value
shall be used for both directions of the F5 flow. The OAM cells for both
directions of the F5 flow must follow the same physical route so that any
connecting points supporting that connection can correlate the fault and
performance information from both directions.
There are two kinds of F5 flows, which can simultaneously exist in a VCC.
These are as follows:
End-to-End F5 Flow. This flow, identified by a standardized PTI see

Rec. I.361., is used for end-to-end VCC operations communications.
Segment F5 Flow. This flow, identified by a standardized PTI see Rec.
I.361., is used for communicating operations information within the
bounds of one VCC link or multiple interconnected VCC links, where
all of the links are under the control of one administration or organization. Such a concatenation of VCC links is called a VCC segment. A
VCC segment can be extended beyond the control of one administration
by mutual agreement.
F5 flows must be terminated only at the endpoints of a VCC or at the

connecting points terminating a VCC segment. Intermediate points i.e.,
connecting points. along the VCC or along the VCC segment may monitor
OAM cells passing through them and insert new OAM cells, but they cannot
terminate the OAM flow. The F5 flow will be initiated at or after connection
setup.
The administrationrorganization that controls the insertion of OAM cells
for operations and maintenance of a VCC segment must ensure that such
OAM cells are extracted before they leave the span of control of that
administrationrorganization.
16.13.4
OAM Functions of the Physical Layer
16.13.4.1 OAM Functions. Two types of OAM functions need to be

distinguished:
1. OAM functions supported solely by the flows F1, F2, and F3.
Dedicated to detection and indication of unavailability state
Requiring real time failure information transport toward the affected endpoints for system protection
2. OAM functions with regard to the system management
Dedicated to performance monitoring and reporting, or for localization of failed equipment
May be supported by the flows F1 to F3 or by other means
671
OAM Functions Supported Solely by the Flows F1 to F3. Table 16.8 gives an
overview of the OAM functions and the related OAM flows. It also lists the
different failures to be detected together with the failure indications for
the SDH-based physical layer. Table 16.9 illustrates the same information
for the cell-based physical layer Ref. 28..
16.13.5
OAM Functions of the ATM Layer
The F4 flow relates to the irtual path and the F5 flow relates to the irtual
channel. In both cases the fault management functions consist of monitoring
of the pathrchannel for availability and other, overall performance monitoring. A path or channel is either not available or has degraded performance.
16.13.5.1 VP-AIS and VP-FERF Alarms. The VP-AIS virtual pathalarm
indication signal. and the VP-FERF virtual pathfar-end reporting failure.
alarms are used for identifying and reporting VPC virtual path connection.
failures.
VP-AIS. VP-AIS cells are generated and sent downstream to all affected
VPCs from the VPC connecting point e.g., ATM cross-connect. which
detects the VPC failure. VP-AIS results from failure indications from the
physical layer as shown in Tables 16.8 and 16.9.
VP-AIS Cell Generation Condition. VP-AIS cells are generated and transmitted as soon as possible after failure indication, and transmitted periodically during the failure condition in order to indicate VPC unavailability. The
generation frequency of VP-AIS cells is nominally one cell per second and is
the same for each VPC concerned. VP-AIS cell generation is stopped as soon
as the failure indications are removed.
VP-AIS Cell Detection Condition. VP-AIS cells are detected at the VPC
endpoint and VP-AIS status is declared after the reception of one VP-AIS
cell. VP connecting points may monitor the VP-AIS cells.
VP-AIS Release Condition. The VP-AIS state is removed under either of
the following conditions:
Absence of VP-AIS cell for nominally 3 s

Receipt of one valid cell user cell or continuity check cell.
VP-FERF. VP-FERF is sent to the far-end from a VPC endpoint as soon as it

has declared a VP-AIS state or detected VPC failure.
672
673
674
675
676
VP-FERF Cell Generation Condition. VP-FERF cells are generated and

transmitted periodically during the failure condition in order to indicate VPC
unavailability. Generation frequency of VP-FERF cells is nominally one cell
per second and shall be the same for all VPCs concerned.
VP-FERF cell generation shall be stopped as soon as the failure indications are removed.
VP-FERF Cell Detection Condition. VP-FERF cells are detected at the
VPC endpoint and VP-FERF state is declared after the reception of one
VP-FERF cell. VPC connecting points may monitor the VP-FERF cells.
VP-FERF Release Condition. The VP-FERF state is removed when no
VP-FERF cell is received during a nominally 3-s period.
VPC Continuity Check. The continuity check cell is sent downstream by a
VPC endpoint when no user cell has been sent for a period of t, where
Ts - t - 2Ts and no VPC failure is indicated. If the VPC endpoint does not
receive any cell within a time interval Tr Tr ) 2Ts., it will send VP-FERF to
the far end.
This mechanism can also be applied to test continuity across a VPC
segment. The need for supporting this mechanism for all VPCs simultaneously is for further study.
VP Performance Management Functions. Performance monitoring of a VPC
or VPC segment is performed by inserting monitoring cells at the ends of the
VPC or VPC segment, respectively. In the procedure supporting this function, forward error detection information e.g., the error detection code. is
communicated by the endpoints using the forward outgoing. F4 flow. The
performance monitoring results, on the other hand, are received on the
reverse incoming. F4 flow. Note that when monitoring VPCs that are
entirely within one span of control or when monitoring VPC segments, the
monitoring result may be reported using the reverse F4 flow or via some
other means we.g., TMS telecommunications management network.x Ref. 28..
Performance monitoring is done by monitoring blocks of user cells.
A performance monitoring cell insertion request is initiated after every N
user cells. The monitoring cell is inserted at the first free cell location after
the request.
The block size N may have the values 128, 256, 512, and 1024. These are
nominal block size values, and the actual size of the monitored cell block may
vary. The cell block size may vary up to a maximum margin of 50% of the
value of N for end-to-end performance monitoring. However, for end-to-end
performance monitoring, the monitoring cell must be inserted into the user
cell stream no more than Nr2 user cells after an insertion request has been
initiated. The actual monitoring block size averages out to approximately N
cells.
TABLE 16.10
677
OAM Type Identifiers
OAM Type
Fault management
Performance management
Activationrdeactivation
4-bit
Function Type
4-bit
0001
0001
0001
0010
0010
0010
1000
AIS
FERF
Continuity check
Forward monitoring
Backward reporting
Monitoringrreporting
Performance monitoring
Continuity check
0000
0001
0100
0000
0001
0010
0000
0001
Source: Table 4rI.610, page 17, Ref. 28.
To eliminate forced insertions when monitoring VPC segment performance, the actual monitoring block size may be extended until a free cell is
available after the insertion request. However, in this case, the actual
monitoring block size may not average out to N cells. Forced insertion at the
segment level remains as an option.
16.13.5.2 OAM Functions for the VCC (F5 Flows). VCC F5 functions are
similar to the VPC F4 functions.
16.13.6
ATM Layer OAM Cell Format
The ATM layer OAM cells contain fields common to all types of OAM cells,
as well as specific fields for each specific type of OAM cell. See Table 16.10.
The coding for unused common and specific fields is as follows:
Unused OAM cell information field octets are coded 0110 1010 6AH.
Unused cell information field bits incomplete octets. are coded all
zeros
16.13.6.1 Common OAM Cell Fields. All ATM layer OAM cells have the
following five common fields: see Figure 16.42..
Figure 16.42. Common OAM cell format. (From Figure 8rI.610, page 18, Ref. 28.)
678
1. Header. Details of this field are in Rec. I.361. For F4 flow identification, two preassigned VCIs are used to distinguish OAM cells for VPCs
and VPC segments. These two values are defined in Rec. I.361. For F5
flow identification, two PTI values are used to distinguish OAM cells
for VCCs and VCC segments. These two values are defined in Rec.
I.361.
2. OAM Cell Type (4 bits). This field indicates the type of management
function performed by this cell, e.g., fault management, performance
management, and activationrdeactivation.
3. OAM Function Type (4 bits). This field indicates the actual function
performed by this cell within the management type indicated by the
OAM Cell Type field.
4. Resered Field for Future Use (6 bits). Default value coded all zero.
5. Error Detection Code (EDC) (10 bits). This field carries a CRC-10 error
detection code computed over the information field of the OAM cell.
The CRC-10 generating polynomial is
G x . s 1 q x q x 4 q x 5 q x 9 q x 10
REFERENCES
1. John Atkins and Mark Norris, Total Area Networking, John Wiley & Sons,
Chichester, UK, 1995.
2. Broadband Aspects of ISDN, CCITT Rec. I.121, CCITT Geneva, 1991.
3. ATM UserNetwork Interface Specification, Version 3.1, The ATM Forum,
Prentice-Hall, Englewood Cliffs, NJ, 1996.
4. B-ISDN ATM Layer Specification, ITU-T Rec. I.361, ITU Geneva, November
1995.
5. B-ISDN UserNetwork InterfacePhysical Layer Specification: General Characteristics, ITU-T Rec. I.432.1, ITU Geneva, August 1996.
6. B-ISDN Protocol Reference Model and Its Application, CCITT Rec. I.321, ITU
Geneva, April 1991.
7. B-ISDN ATM Adaptation Layer Specification: Type 1 AAL, ITU-T Rec. I.363.1,
ITU Geneva, August 1996.
8. B-ISDN ATM Adaptation Layer AAL. Specification, CCITT Rec. I.363, ITU
Geneva, 1991.
9. B-ISDN ATM Adaptation Layer Specification: Type 2 AAL, ITU I.363 para 3.
Geneva, March 1993.
10. Radia Perlman, Interconnections, 2nd edition, Addison-Wesley, Reading, MA,
2000.
12. David E. McDysan and Darren L. Spohn, ATM: Theory and Application, McGraw-Hill, New York, 1995.
REFERENCES
679
14. Support of the Broadband Connectionless Data Bearer Serice by the B-ISDN, ITU-T
Rec. I.364, ITU Geneva, November 1995.
15. The International Public Telecommunication Numbering Plan, ITU-T Rec. E.164,
ITU Geneva, May 1997.
16. B-ISDN General Network Aspects, ITU-T Rec. I.311, ITU Geneva, August 1996.
17. Digital Subscriber Signaling System No. 2 DSS2 . UserNetwork Interface UNI .
Layer 3 Specification for Basic Call r Connection Control, ITU-T Rec. Q.2931, ITU
Geneva, February 1995.
18. Interim Inter-Switch Signaling Protocol IISP . Specification, Version 1.0, ATM
Forum, Mt. View, CA, December 1994.
19. Functional Description of the B-ISDN User Part B-ISUP . of Signaling System No.
7, ITU-T Rec. Q.2761, ITU Geneva, February 1995.
20. Signaling System No. 7 B-ISDN User Part B-ISUP . Basic Call Procedures, ITU-T
Rec. Q.2764, ITU Geneva, February 1995.
21. B-ISDN Inter-Carrier Interface B-ICI . Specification, Version 2.0, ATM Forum, Mt.
View, CA, December 1995.
22. ATM UserNetwork Interface UNI . Signaling Specification, Version 4.0, ATM
Forum, Mt. View, CA, July 1996.
23. Traffic Control and Congestion Control in B-ISDN, ITU-T Rec. I.371, ITU Geneva,
August 1996.
24. B-ISDN ATM Layer Cell Transfer Performance, ITU-T Rec. I.356, October 1996.
25. Broadband Testing Technologies, an H-P seminar, Hewlett-Packard Company,
Burlington, MA, October 1993.
26. ATM Cell Mapping into Plesiochronous Digital Hierarchy PDH ., ITU-T Rec.
G.804, ITU Geneva, February 1998.
27. Cell Relay Serice Core Features, Bellcore Special Report SR-3330, Issue 2,
Bellcore, Red Bank, NJ, 1996.
28. B-ISDN Operation and Maintenance Principles and Functions, ITU-T Rec. I.610,
17
LAST-MILE DATA
DISTRIBUTION SYSTEMS
17.1
NEW APPROACHES TO OUTSIDE PLANT DISTRIBUTION
Outside plant is a term used to describe all the equipment involved in taking a
telephone signal from the local serving exchange local switch . and delivering
it to the subscriber. In a traditional sense, this equipment mostly encompassed wire-pair distribution systems. Outside plant represented as much as
one-third of the total investment in plant and equipment of telephone
companies. Historically and through the present, it is wire-pair systems that
served the last mile to the subscriber.
Twisted wire pair is an excellent transmission medium when employed in
some kind of controlled environment. In this last-mile application it transported a simple, band-limited analog signal that interfaced with a telephone
handset on the customer premises. On most of its lay from the local switch to
its termination in a telephone subset, it was encased in a multipair cable
often carrying many hundreds of such pairs. There was access to the cable
along its lay route in distribution boxes and at a pedestal outside the
customer location or at a small distribution frame in an office building.
In nearly all cases, the telephone company could not afford to pamper or
doctor the cable: Here it is, take it as you find it. We find that some pairs
did a better job than others; and as time went by, the cable tended to
deteriorate due to simple aging, splicing, and the effects of the environment
e.g., freezing, flooding, and drying out. Ref. 1..
We take these same pairs from a cable and put them in a laboratory, and
we can demonstrate the ability of carrying megabits of digital data. Still and
yet, it can easily be shown that bit rate on a wire pair varies with
Length of the pair

Diameter of the wire
Capacitance between the pair members
681
682
LAST-MILE DATA DISTRIBUTION SYSTEMS
Place that pair back in its cable sheath, bury it and leave it some years and it
just will not work as well as it did in the laboratory.
Now in this day and age of the Internet, we would like to deliver some
megabitsrsecond to the customer premise on that wire pair downstream..
However, because of their condition, some of the pairs force us to reduce the
bit rate to the 128-kbps range or even less. Of course, the longer the
subscriber line, the lower the bit rate that is achievable.
The category of system we describe here is the digital subscriber line DSL..
Early on we had ISDN BRI Chapter 13. which provided 64-kbps full-duplex
access to the serving exchange from a subscriber on one logical line.* Placing
the two available logical lines together in the BRI, 128 kbps was achievable
with only some constraints. Then HDSL high-speed digital subscriber line.
appeared, providing 784 kbps or a full DS1 rate 1.544 Mbps.. To achieve this
transmission rate requires two pair, each carrying 784 kbps in each direction.
The specified reach of HDSL was 12,000 ft 3700 m..
Up to this point we have described symmetrical operation. All this means is
that there is the same bit rate operating in each direction. Then asymmetrical
operation came along. This is where the bit rate in one direction is different
than in the other direction. Consider the Internet. Where wed like the high
bit rate is from the ISP internet service provider. toward the customer. The
difficult part of this connectivity is from the local serving switch to the
customer. The connection provided from the ISP to the local serving switch
often is at the DS1 rate 1.544 Mbps. or greater. The connectivity from the
customer to the ISP need not have a bit rate any greater than 56 kbps or even
28 kbps. Thus we have asymmetrical service: megabits downstream from
the switch to the customer. and kilobits of bit rate in the other direction
i.e., toward the ISP or upstream..
17.2
INTRODUCING LAST-MILE DISTRIBUTION SYSTEMS
The last-mile distribution, of course, involves the connectivity from some

node or local serving switch to the customer premise. Whether the distance is
exactly one mile is immaterial. In some cases it is much less than a mile,
while in other cases it is more than a mile. Certainly in the rural situation it
may be many miles. In this case the connectivity should be handled in some
other manner.
Currently, there are four ways to deliver a megabit signal to customer
premises:
1. By DSL digital subscriber line.
2. By modified cable television CATV. plant
*Remember that in North America, ISDN BRI is a 2-wire, full-duplex system that carried two
B channels and a D channel. Its bit rate in either direction was 160 kbps, of which 16 kbps was
overhead.
683
3. By LMDS local multipoint distribution system.

4. By other wideband radio wireless . systems
We will briefly discuss DSL. In many cases we believe its weaknesses
outweigh its benefits. The three basic weaknesses were described in Section
17.1. Its single benefit is economy, and the wire pairs are now in place to
provide POTS plain old telephone service.. The proprietary or semiproprietary schemes of comparatively high bit rates on wire pair usually take on an
acronym ending in SL. We described the generic term DSL and the HDSL.
There is ADSL asymmetrical digital subscriber line. supported by ANSI on
one hand and a proprietary scheme on the other. They do not interface and
are quite different, one from the other. The ANSI ADSL supports some 6
Mbps downstream distance limited. and 640 kbps upstream. As the subscriber line gets longer, the supported bit rates drop. Of course, if the
distance is limited to about 1 mile, some 6000 ft 1829 m., the full data rate
should be achieved if the subscriber loop is in good condition and without
resorting to specialized line coding techniques.
XDSL is a generalized acronym for any one of the several semiproprietary
or proprietary schemes available.
As we have discussed, the shorter the wire-pair loop between the DSL
terminal and the user installation, the greater the bit rate that can be
supported. One way to achieve this goal is to bundle a group of wire pairs
and replace the bundle with a fiber-optic pair of strands. This fiber-optic pair
has one strand for downstream and one for upstream. It carries a multiplexed
configuration to a remote terminal near the subscribers to be served. The
remote terminal unit RTU. demultiplexes the downstream and distributes
the signals on to the appropriate pairs. In the other direction, upstream, the
RTU takes the electrical signals from each pair and multiplexes them for
transport to the local serving exchange. Here we can say that the last mile
segment is from the RTU to user sites. With good planning of the system,
these derived copper pair loops are shortened to 6000 ft 1829 m. or less,
permitting the highest bit rate the DSL design allows.
As we have covered in several instances, the principal constraint of data
rate is loop attenuation. When planning a DSL installation, one must
consider that loop attenuation increases with frequency. Considering the
bandwidth of a DSL system and allowing 1 bit per hertz, a DS1 configuration
would have a bandwidth of 1.544 MHz equivalent to 1.544 Mbps.. If,
somehow, we can reduce the signal bandwidth of a DSL system, we can
effectively stretch the reach of the system. In North America, a 2B1Q
waveform has been adopted for ISDN BRI. For every electrical pulse
transmitted with 2B1Q, 2 bits of data are sent. Thus, effectively, the bandwidth is cut in half. This will more than double the reach of a typical DSL
system such as DS1. The 2B1Q line code is discussed in Section 13.6.2.6.
There is also a proprietary line code developed by Paradyne Ref. 2. called
CAP. CAP stands for carrierless amplitude and phase modulation. This is a
684
Figure 17.1. A sketch of an idealized spectra of three DSL waveforms. (From Paradyne,
Figure 6, page 14, Ref. 2.)
line coding technique that allows multiple bits per hertz of bandwidth, up to
9 bits per hertz. CAP further reduces bandwidth requirements on a wire pair
providing considerably greater range or bit rate. than a 2B1Q system
provides.
Figure 17.1 is an idealized sketch of the occupied spectrum on a wire-pair
subscriber line of three transmission techniques: 2B1Q HDSL, CAP T1
HDSL, and conventional T1 AMI.* This latter waveform is described in
Section 7.2.4.
Crosstalk is a dominant factor in the performance of many DSL systems.
Crosstalk is the coupling of signals from adjacent wire pairs in a cable bundle
into the signal bit stream of the individual desired pair under consideration.
Crosstalk can seriously degrade bit error rate.
In a fully configured, traditional DSL installation, data traffic bypasses the
local serving switch. The data traffic derived from the DSL network is
concentrated and handed off to an interswitch digital cross-connect DACS..
A better way of handling this data aggregate is to use a digital subscriber
line access multiplexer DSLAM.. This device groups data channels before
handoff. It uses packet and cell multiplexing technology in addition to
conventional T1DS1.rE1 introduced into the DSLAM, which results in a
more efficient use of bit stream flow resources.
In the local serving switch which houses the DSLrDSLAM equipment we
will expect to find transport system resources such as T1rE1, T3rE3, OC-1,
OC-3, STS-1, and STS-3 facilities being used to support the switch. These
*AMI stands for alternate mark inversion. This is the specialized waveform used in the
transmission on wire pair of T1 DS1. PCM.
685
resources interface the switch to produce interswitch trunks. They will also
interface the DSLAM facilities.
The local access network utilizes the local carrier interswitch network as a
foundation for services. Additional equipment may be required in order to
provide connectivity between multiple service providers and multiple service
users. This equipment may involve frame relay switches, ATM switches
andror routers and digital data cross-connects.
We consider the DSLAM the cornerstone of the DSL solution. Functionally, the DSLAM concentrates data traffic from multiple DSL loops onto the
backbone network for connection to the rest of the network. The DSLAM
provides backhaul services for packet, cell, andror circuit-based applications
through concentration of DSL lines onto T1rE1, T3rE3, or ATM outputs.
The DSLAM facility may be called upon to carry out routing functions such
as supporting dynamic IP address assignment using dynamic host control
protocol see Section 11.4.3..
Figure 17.2 shows a DSL system reference model for full service deployment. ADSL terminology is used throughout. On the left side of the drawing
are the network service providers NSPs.. These could represent frame relay
Figure 17.2. A fully provisioned DSL model. (After Paradyne Figure 32, page 35, Ref. 2.)
686
service provider, ISP internet service provider., ATM service provider, and
so forth. The edge device may be a switch.
ATU stands for ADSL transceiver unit. One is called an ATU-C and is
usually housed in the DSLAM facility which is located in the local service
switch complex with access to its MDF main distribution frame.. Of course
C stands for central office, a distinctly North American term for a telephone switching center. The ATU-C has a companion remote unit called an
ATU-R, which is located on the user premises. The DSL connects the two.
In the center of the drawing is the network access provider NAP.. The
NAP may belong to the local telephone company or LEC or to the CLEC or,
possibly, to an independent operator that provides a DSL service. Often the
DSLAM houses the ATU-C devices. It will have at least one port for N64
services to its associated local switch. Otherwise, it bypasses the local switch
accessing the appropriate networks for service such as a. a SONET network
or b. a PSTN PDH digital network.
There is another aspect to last-mile connectivity. This is a North
American phenomenon. Historically, the outside plant distribution system,
better defined as last-mile delivery, was a monopoly, held by the local
telephone company. That has now changed in the United States. The local
telephone company is officially known as a local exchange carrier or LEC. The
monopoly has been broken and now appearing on the scene is the CLEC or
competitive local exchange carrier. A CLEC may share the outside plant,
probably originally installed and maintained by the LEC; install their own
wire-pair distribution system, or use some other means to provide an electrical signal to the subscriber.
Cable Television CATV.* is a last-mile delivery system of entertainment,
principally television. It is a broadband system that is already in place. We
reiterate those important points: CATV is a broadband system, and it is
already in place in the United States and other nations. With some modifications, it has the potential to be an excellent broadband delivery system for
voice POTS., datarinternet, and conventional television entertainment. It
can provide 10-Mbps or greater data rates for downstream internet for many
users simultaneously as well as be a transporter of ATM, frame relay at
E1rDS1 rates. It can create VPNs and set up IP connectivity. It is much
more versatile than DSL and suffers few of its shortcomings. CATV is
discussed in Section 17.3.
A second very viable other means is to transport and distribute wideband signals via LMDS, which stands for Local Multipoint Distribution
System. This is a wireless or radio system that operates in the 28 to 31-GHz
band and provides better than 1 GHz of bandwidth. LMDS is described in
Section 17.4. It provides straight, head-on competition with CATV.
*CATV stands for community antenna television, the more formal name for cable television.
INTRODUCTION TO CATV
17.3
687
The original concept of CATV was to deliver quality TV program service to

the home and office. TV in itself is a broadband signal 6 MHz.. It is
broadband when compared to the familiar voice channel which has a bandwidth of some 4 kHz. Whereas a wire pair in the conventional outside plant
only delivers just one nominal 4-kHz signal to a customer, CATV delivers a
choice of many TV channels to its customers. In essence, CATV is much
more complex than its wire-pair counterpart. It also has much greater service
potential.
The concept of CATV is fairly simple. In its earlier days ca. 19601970., a
high location was selected near or outside a city which had excellent line-ofsight conditions with the citys various TV emitting antennas. Their signals
were received at this location, which we choose to call a head-end, then
amplified and coupled into a low-loss coaxial cable system. The coaxial cable
distributed this very broadband signal to customers along a preplanned route.
At various locations along the route the composite TV signal aggregate was
amplified. Each broadband amplifier along the route added noise to the
signal, placing a limit on the number of amplifiers in tandem that may be
employed. It was a neat concept because a regular TV set could couple into
the coaxial cable with a tap and tune to any one of some fair number of
channels. The set needed no modification. In fact the set could be simply
switched from a rabbit-ear or other. antenna to the CATV tap and back
again if a user so desired.
Figure 17.3. An early CATV distribution system.
688
Figure 17.4. The radio-frequency spectrum as seen by a TV set, whether off-the-air or off the
cable. Note: The familiar channel numbers are in parentheses. Frequencies are in megahertz.
Figure 17.3 shows a simplified block diagram of the original CATV

concept ca. 1965.. That concept had changed little up to several years ago,
when optical fiber entered the picture. Note that the head-end HE. is where
off-the-air signals are received captured .; added to these may be other TV
signals brought to the head-end by LOS microwave. Later, of course, many of
the channels derived from satellite transport, probably the bulk of the TV
channels.
Figure 17.4 shows the radio-frequency RF. spectrum. Only TV channels
are shown. This is the same spectrum that appears on the coaxial cable, and
only the TV channels are operable.
CATV system designers soon learned not to waste bandwidth. For example, the off-the-air spectrum, which is assigned for usage by the FCC, is not a
continuous spectrum from 54 MHz upwards. Other services such as aeronautical mobile, public safety, marine, and others are assigned frequencies
i.e., from 88 to 174 MHz.. There is the familiar FM broadcast band from 88
to 106 MHz as well. The coaxial cable is well-shielded from unwanted
off-the-air emissions, so then why not use that vacant spectrum from a
CATV operator perspective .?
What the operator did was simply take TV channels from other sources
such as those brought in by microwave or by satellite, and it translated them
in frequency to the desired vacant 6-MHz-wide frequency slots. Typically,
we might find channel 25 off-the-air appearing on our cable in the channel
14 slot.
17.3.1
The Essentials of TV Signal Distribution on a CATV System
The CATV system designer deals with a broadband system based on coaxial
cable for signal transport. The frequency response of coaxial cable is nonlinear, more approaching the quadratic. On the other hand, from our perspective, fiber-optic cable presents a flat response to the user. Figure 17.5 shows a
typical frequency response curve of coaxial cable. For viable TV signal
distribution we would like as flat a response as possible. A hypothetical
frequency response curve for a section of fiber-optic cable is shown in Figure
17.6. Note that the curve is flat. We can flatten the frequency response
689
Figure 17.5. Attenuationfrequency response of typical coaxial cable.
Figure 17.6. A length of hypothetical fiber-optic cable showing flat frequency response.
characteristics of coaxial cable by the use of equalizers. An equalizer, in this

case, adds loss in a mirror image of the response curve in Figure 17.5, making
the response appear flat. Of course, the insertion loss of an equalizer must be
taken into account. The more we extend the useful upper frequency of an
equalizer, the greater the insertion loss. Thus there is a practical limit to how
far we can extend the useful upper frequency. Right now the frequency lies
690
between 800 and 1000 MHz. This gives us a transmission medium nearly
1 GHz wide.
To distribute CATV TV signals to customer premises, we need to flatten
the medium with equalizers and present a signal of sufficient level to the TV
receiver input to achieve acceptable picture quality. The TV picture must
also have a signal-to-noise ratio SrN . of at least 40 dB, preferably ) 46 dB.
Now remember that we end up with a system with some number X of
wideband amplifiers in tandem. Each amplifier adds noise to the system. The
number X therefore must be limited or we will not achieve our SrN
objective. This puts a definite limit on the CATV coverage area of customer
premises.
17.3.2
Extending CATV Coverage Area
CATV system designers began to consider the application of fiber-optic cable

transmission to extend coverage area. One practical approach is to transmit
the entire RF spectrum from 54 MHz to some maximum frequency, say 750
MHz for arguments sake, along with the artificially inserted channels along
the fiber. Remember that optical fiber is a very broadband medium. This
frequency configuration would originate at the head-end and would be
transmitted out on sets of fiber-optic supertrunks to nodes somewhere near
groups of residences or office buildings. Each node converts the signal from a
light signal to an electrical RF signal on a conventional CATV coaxial cable
system. As we have configured our system, there is a maximum capacity of
116 NTSC* TV channels w750 y 54.r6x.
Such systems described here are called HFC hybrid fiber-coax. systems.
They offer many advantages to the CATV operator among them are improved carrier-to-noise ratio CrN., and they result in a better SrN value at
the premises TV set, less self-interference, improved reliability because there
are fewer net active components, and the potential to greatly increase a
serving area. Another extremely important advantage is to provide a means
of implementing a very viable two-way digital service.
One important service in high demand is access to the Internet. The
configuration is asymmetrical in that the downstream bit rate will be considerably greater than the companion upstream data rate. We probably can
allow the upstream data rate to be on the order of 56 or 64 kbps per active
user, along with the companion downstream rate on the order of 16 Mbps.
Lets now take the RF spectrum shown in Figure 17.4, and lets reconfigure it for two-way digital operation. The spectrum layout to meet the
requirements outlined above is shown in Figure 17.7. Note the asymmetry
where the bandwidth assigned for downstream voice and data services is
*NTSC stands for National Television System Committee. This is an analog, color TV configuration where each channel is assigned a 6-MHz bandwidth. NTSC is the North American analog
color TV transmission standard.
691
Figure 17.7. CATV spectrum assigned for upstream and downstream operation. The frequency assignment is somewhat arbitrary but probably will be encountered in practice. The 5to 42-MHz segment for upstream operation is specified by DOCSIS. (See Section 17.3.3.)
Table 17.1 shows DOCSIS downstream operation is in 6-MHz segments from 88 MHz
upwards.
contained in one or two 6-MHz segments way up at the top of the band, in
what are called roll-off channels. This is where filter roll-off occurs. The
channel response will not support conventional TV, but will support downstream data. If more bandwidth is required, the CATV operator can assign
one or several of the top 6-MHz TV segments to provide additional downstream service. Here we have 40 MHz available upstream and 6, 12, and 18
MHz or more. for downstream datarvoice.
Curing the Ills of Bandwidth Scarcity. If we were to implement the CATV
plant as we describe, where the entire plant is served by one coaxial cable
configuration, we would never be able to serve the demand for Internet
service. This is so, leaving aside other desirable data services. Probably the
first shortage would show up on the downstream side. If we were to allow 5
Mbps per user for internet downstream, including guardband, each simultaneous user would take up about 1 TV channel i.e., 6 MHz. of bandwidth.
Either we cut back the bandwidth per user in periods of high demand, thus
cutting bit rate, or we may look to another method of signal distribution.
One method being adopted widely is to bring the serving node closer to
the customer. CATV users are grouped in units of 50 to 200. The group is
served by a node. The node takes the downstream frequency segment from
an FDM configuration on the fiber of some 1000 MHz in bandwidth say in
the band 7 to 8 GHzand translates the signals to the 54- to 750-MHz band.
Note that some additional bandwidth is used for guard band. A nearby nodes
downstream configuration is in the band 6 to 7 GHz, and so forth, for each
serving node. Note that the same fiber strand is used for each node. A
passive optical signal splitter delivers the downstream signal for grooming
and downconversion at each node.
The nodal processing for the upstream direction is shown in Figure 17.8.
Here the frequency band of 5 to 30 MHz is used for upstream traffic, a
25-MHz bandwidth. Three active groupings of 50 to 200 residences are shown
in the figure. In this direction we call the device residing in the node a
coax-to-fiber translator. Because of the richness of bandwidth availability on
692
Figure 17.8. Upstream concept breaking up a user community into groups of from 50 to 200
end-user stations. This figure relates to Figure 17.7, where frequency band assignments are
somewhat arbitrary. The upstream band will carry POTS, data, and internet carriers. The last
mile of operation will be on coaxial cable. Downstream internet, POTS, and data service are
provided on separate frequency segments as well. The downstream frequency assignment is
in 6-MHz segments from about 50 MHz upward to 750 MHz, and in some plants to 860 MHz.
Expect to find the data, internet, and POTS channels at the top of the band for downstream.
the fiber, only every other 25-MHz segment carries upstream traffic where
the other segments are guardbands. Of course these fiber strands derive
from the head-end via supertrunks.
For internet application, upstream operation will be the direction from
end-user to the ISP. In the case of POTS, operation will be symmetrical;
assume 64 kbps per voice channel. The CATV operator will probably also
offer data network connectivity, possibly IP, frame relay, or ATM at differing
bit rates. One common approach is to turn to DOCSIS Data-over-Cable
Service Interface Specifications . developed by Cable Labs and SCTE.
Modems designed to meet DOCSIS requirements are available from the
principal CATV equipment vendors such as GI Jerrold and Motorola.
17.3.3
The DOCSIS Specification
Figure 17.9 shows the reference architecture for the data-over-cable specification. The following acronyms and abbreviations are defined to assist the
reader in understanding DOCSIS operation.
MCNS. Multimedia Cable Network System consortium..
HFC. Hybrid fiber-coax described above..
IP. Internet Protocolsee Section 11.4.
693
694
CMTS. Cable Modem Terminator System. DOCSIS modem located at

CATV head-end. See Figure 17.9.
CM. Cable modem. The DOCSIS modem located at the customer premises
installation. See Figure 17.9.
It should be noted from Figure 17.9 that the upstream spectrum occupies
the band from 5 to 42 MHz, and the downstream spectrum occupies the band
from 50 to 860 MHz.
The key functional characteristics assumed in the DOCSIS specification
are as follows:
Two-way transmission.
A maximum opticalrelectrical spacing between the CMTS and the most
distant customer terminal of 100 miles, although typical maximum
separation may be 1015 miles 1624 km..
A maximum differential opticalrelectrical spacing between CMTS and
the closest and most distant modems of 100 miles 161 km., although
this would typically be limited to 15 miles 24 km..
17.3.3.1 RF Channel Assumptions. The Data-over-Cable system, configured with at least one set of defined physical-layer parameters e.g.,
modulation, forward error correction, symbol rate, etc.. from the range of
configuration settings described in that specification, must be capable of
operating with a 1500-byte packet where the packet loss rate will be less than
1% while forwarding at least 100 packets per second on cable networks
having characteristics defined in this section.
Transmission Downstream. The RF channel transmission characteristics of
the cable network in the downstream direction assumed for the purposes of
minimal operating capability are described in Table 17.1. This assumes
nominal analog video carrier level peak envelope power. in a 6-MHz
channel bandwidth. All conditions are present concurrently.
Transmission Upstream. The RF channel transmission characteristics of the
cable network in the upstream direction assumed for the purposes of minimal
operating capability are described in Table 17.2. All conditions are present
concurrently.
Transmission Levels. The nominal power level of the downstream CMTS
64-QAM* signals. within a 6-MHz channel is targeted to be in the range of
y10 dBc to y6 dBc relative to analog video carrier level. The nominal
power level of the upstream CM signals. will be as low as possible to achieve
*Note that 64-QAM will give us theoretically 6 bitsrHz of bit packing 2 6 s 64..
695
696
the required margin above noise and interference. Uniform power loading
per unit bandwidth is commonly followed in setting upstream signal levels,
with specific levels established by the cable network operator to achieve the
required carrier-to-noise and carrier-to-interference ratios.
17.3.3.2 Communication Protocols. The CM and CMTS operate as forwarding agents and also as end-systems hosts.. The protocol stacks used in
these modes differ as described below.
The principal function of the cable modem system is to transmit Internet
Protocol IP. packets transparently between the head-end and the subscriber
location. Several management functions also ride on IP, so that the protocol
stack on the cable network is as shown in Figure 17.10. These management
functions include, for example, supporting spectrum management functions
and the downloading of software.
697
Figure 17.10. The protocol stack on the DOCSIS RF interface. (From Figure 3-1, page 11,
DOCSIS Specification, Ref. 3.)
The CMs and CMTSs will operate as IP and LLC hosts in terms of IEEE
Standard 802 for communication over the cable network. The protocol stack
at the CM and CMTS RF interfaces is shown in Figure 17.10.
The CM and CMTS function as IP hosts. As such, the CM and CMTS
support IP and ARP over DIX link-layer framing. The CM and CMTS may
also support IP and ARP over SNAP framing.
CM and CMTS as Hosts. The CM and CMTS also function as LLC hosts. As
such, the CM and CMTS respond appropriately to TEST and XID requests
per ISO 8802-2 LLC..
Data Forwarding Through CM and CMTS
General. Data forwarding through the CMTS may be transparent bridging, or may employ network-layer forwarding routing, IP switching. as shown
in Figure 17.11.
698
Figure 17.11. Data forwarding through CM and CMTS. (From Figure 3-2, page 12, DOCSIS
Specification, Ref. 3.)
Data forwarding through the CM is link-layer transparent bridging, as

shown in Figure 17.11. Forwarding rules are similar to ISOrIEC 10038 with
modifications described in CMTS and CM Forwarding Rules below. This
allows support of multiple network layers.
Forwarding of IP traffic is supported. However, support of other networklayer protocols is optional. The ability to restrict the network layer to a single
protocol is required.
Support for the IEEE 802.1d spanning tree protocol of ISOrIEC-10038
with the modifications described in CM Forwarding Rules is optional for CMs
intended for residential use. CMs intended for commercial use and bridging
CMTSs support this version of the spanning tree. CMs and CMTSs include
the ability to filter and disregard. 802.1d BPDUs. The DOCSIS specification
assumes that CMs intended for residential use will not be connected in a
configuration which would create network loops such as that shown in Figure
17.12.
The MAC Forwarder. The MAC forwarder is a MAC sublayer that resides on
the CMTS just below the MAC service access point MSAP. interface, as
shown in Figure 17.13. It is responsible for delivering upstream frames to:
One or more downstream channels

The MSAP interface
In Figure 17.13, the LLC sublayer and link security sublayers of the upstream
and downstream channels on the cable network terminate at the MAC
Forwarder.
699
Figure 17.12. Example condition for network loops.
Figure 17.13. MAC Forwarder. (From Figure 3-4, page 15, DOCSIS Specification, Ref. 3.)
The MSAP interface user may be the NSI-RFI Forwarding process or the
CMTSs host protocol stack.
Delivery of frames may be based on data-link-layer bridging. semantics,
network-layer routing. semantics, or some combination thereof. Higher-layer
semantics may also be employed e.g., filters on UDP port numbers.. The
CMTS should provide IP connectivity between hosts attached to cable
modems, and it must do so in a way that meets the expectations of Ethernetattached customer equipment. For example, the CMTS must either forward
ARP packets or must facilitate a proxy ARP service. The CMTS MAC
Forwarder may provide service for non-IP protocols.
Note that there is no requirement that all upstream and downstream
channels be aggregated under one MSAP as shown above. The vendor could
just as well choose to implement multiple MSAPs, each with a single
upstream and downstream channel.
700
Network Layer. As stated at the outset, the purpose of the DOCSIS system
is to transport IP traffic transparently through the system. The network layer
protocol is the Internet Protocol IP. version 4, as defined in RFC-791, and
migrating to IP version 6. The DOCSIS specification imposes no requirements for reassembly of IP packets.
Above the Network Layer. Users of DOCSIS will be able to use the transparent IP capability as a bearer of higher-layer services. Use of these services
will be transparent to the CM.
In addition to the transport of user data, there are several network
management capabilities that depend upon the network layer. These include:
SNMP Simple Network Management Protocol. based on RFC-1157.

See Chapter 18 for a general discussion of network management and of
SNMP in particular.
TFTP Trivial File Transfer Protocol., RFC-1350, a file transfer protocol, for downloading software and configuration information.
DHCP Dynamic Host Configuration Protocol, RFC-1541., a framework
for passing configuration information to hosts on a TCPrIP network.
Data-Link Layer. The data-link layer is divided into sublayers in accordance

with IEEE 802 family of standards see Chapters 9 and 10.. The sublayers
from the top are:
Logical link control LLC. sublayer Class 1 only.

Link-layer security sublayer
Media Access Control MAC. sublayer
17.3.3.3 Logical Link Control. The LLC sublayer is provided in accordance with Chapter 9 IEEE Std. 802.2.. Address resolution is used as
defined in Chapter 11 and RFC-826. The MAC-to-LLC service definition is
specified in ISOrIEC-10039.
17.3.3.4 Link-Layer Security Sublayer. Link-layer security is provided in
accordance with specifications MCNS2 DOCSIS. and MCN8.
17.3.3.5 MAC Sublayer. The MAC sublayer defines a single transmitter
for each downstream channelthe CMTS cable modem termination system.. The CMs cable modems. listen to all frames transmitted on the
downstream channel upon which they are registered and accept those where
the destinations match the CM itself or CPEs reached via the CMCI cable
modem-to-customer interface . port. CMs can communicate with other CMs
only through the CMTS. See Figure 17.9..
701
The upstream channel is characterized by many transmitters CMs. and

one receiver the CMTS.. Time in the upstream channel is slotted, providing
for TDMA at regulated time ticks. The CMTS provides the time reference
and controls the allowed usage for each interval. Intervals may be granted for
transmissions by particular CMs, or for contention by all CMs. CMs may
contend to request transmission time. To a limited extent, CMs may also
control to transmit actual data. In both cases, collisions can occur and retries
are used.
MAC Oeriew. Some of the MAC protocol highlights include:
Bandwidth allocation controlled by CMTS

A stream of mini-slots in the upstream
Dynamic mix of contention- and reservation-based upstream transmit
opportunities
Bandwidth efficiency through support of variable-length packets
Extensions provided for future support of ATM or other Data PDU
Class-of-service support
Extensions provided for security at the datalink layer
Support for a wide range of data rates
Physical Layer. The physical PHY. layer comprises two sublayers:
Transmission Convergence sublayer present in the downstream direction only.

Physical media-dependent PMD. sublayer
17.3.3.6 Downstream Transmission Convergence Sublayer. The

Downstream Transmission Convergence sublayer exists in the downstream
direction only. It provides an opportunity for additional services over the
physical-layer bitstream. These additional services, for example, might include digital video.
The sublayer is defined as a continuous series of 188-byte MPEG see
ITU-T Rec. H.222.0. packets, each consisting of a 4-byte header followed by
184 bytes of payload. The header identifies the payload as belonging to the
data-over-cable MAC. Other values of the header may indicate other payloads. The mixture of payloads is arbitrary and controlled by the CMTS.
17.3.3.7 PMD Sublayer Overview. The PMD sublayer involves digitally
modulated RF carriers on the analog cable network. In the downstream
direction, the PMD sublayer is based on ITU-T Rec. Q.83-B, with the
exception of the Scalable Interleaving to Support Low Latency Capability.
702
The PMD includes the following features:
64- and 256-QAM modulation formats*

6-MHz occupied spectrum coexists with all other signals on the planet
Concatenation of ReedSolomon block code and Trellis code supports
operation in a higher percentage of North American cable networks
Variable-depth interleaver supports both latency sensitive and insensitive data.
The features in the upstream direction are as follows:
Flexible and programmable CM under control of the CMTS

Frequency agility
Time division multiple access
QPSK and 16 QAM modulation formats
Support of both fixed-frame and variable-length PDU formats
Multiple symbol rates
Programmable ReedSolomon block coding
Programmable preambles
17.3.3.8 Interface Points. Three RF interface points are defined at the

PMD sublayer:
a. Downstream output on the CMTS
b. Upstream input on the CMTS
c. Cable inrout at the cable modem
Separate downstream output and upstream input interfaces on the CMTS
are required for compatibility with typical downstream and upstream combining and splitting arrangements in head-ends.
17.3.3.9 Physical-Media-Dependent Sublayer Specification. The
DOCSIS standard defines the electrical characteristics and protocol for a
cable modem CM. and cable modem termination system CMTS.. Some of
the high points of the PMD are covered in the following text.
Upstream
Oeriew. The upstream PMD sublayer uses FDMArTDMA burst modulation format, which provides five symbol rates and two modulation formats
QPSK and 16-QAM.. The modulation format includes pulse shaping for
*QAM modulation, where QAM stands for quadrature amplitude modulation, an efficient
method of getting more bits per hertz. 16-QAM achieves 4 bits per Hz bit packing; 64-QAM,
6 bits per Hz; and 256-QAM achieves 8-bitsrHz bit packing. For example, with 256-QAM,
6-MHz bandwidth can accommodate 6 = 8 or 48-Mbps bit rate in theory; in practice the actual
bit rate is considerably less.
703
spectral efficiency, is carrier-frequency agile, and has selectable output power

level. The PMD format also includes a variable-length modulated burst with
precise timing beginning at boundaries spaced at integer multiples of 6.25 s
apart which is 16 symbols at the highest data rate.. Each burst supports a
flexible modulation, symbol rate, preamble, randomization of the payload,
and programmable FEC encoding.
At the upstream transmission, parameters associated with burst transmission outputs from the CM are configurable by the CMTS via MAC messaging. Many of the parameters are programmable on a burst-by-burst basis.
The PDM sublayer can support a near-continuous mode of transmission,
wherein ramp-down of one burst may overlap the ramp-up of the following
burst, so that the transmitted envelope is never zero. The system timing of
the TDMA transmission from the various CMs will provide that the center
of the last symbol of one burst and the center of the first symbol of the
preamble of an immediately following burst are separated by at least the
duration of five symbols. The guard time will be greater than or equal to
the duration of five symbols plus the maximum timing error. Timing error is
contributed by both the CM and CMTS. Maximum timing error and guard
time vary with CMTSs from different vendors.
The upstream modulator is part of the cable modem that interfaces with
the cable network. The modulator contains the actual electrical-level modulation function and the digital signal processing function. The latter provides
the FEC, preamble prepend, symbol mapping, and processing steps. The
specification is written with the idea of buffering the bursts in the signal
processing portion, along with the signal processing portion 1. accepting the
information stream a burst at a time, 2. processing this stream into a
complete burst of symbols for the modulator, and 3. feeding the properly
timed bursted symbol stream to a memoryless modulator at the exact burst
time. The memoryless portion of the modulator only performs pulse shaping
and quadrature upconversion.
At the demodulator, similar to the modulator, there are two basic functional components: the demodulation function and the signal processing
function. Unlike the modulator, the demodulator resides in the CMTS and
the specification is written with the concept that there will be one demodulation function not necessarily an actual physical demodulator. for each
carrier frequency in use. The demodulation function would receive all bursts
on a given frequency.
Note: The unit design approach should be cognizant of the multiple-channel nature of the demodulation and signal processing to be carried out at the
head-end, and partitionrshare functionality appropriately to optimally leverage the multichannel application. A demodulator design supporting multiple
channels in a demodulator unit may be appropriate.
The demodulation function of the demodulator accepts a varyinglevel signal centered around a commanded power level and performs
symbol timing and carrier recovery and tracking, burst acquisition, and
demodulation.
704
Additionally, the demodulation function provides an estimate of burst

timing relative to a reference edge, an estimate of received signal power, and
an estimate of signal-to-noise ratio and may engage adaptive equalization to
mitigate the effects of a. echoes in the cable plant, b. narrowband ingress,
and c. group delay. The signal-processing function of the demodulator
performs the inverse processing of the signal-processing function of the
modulator. This includes accepting the demodulated burst data stream and
decoding, and so on, and possibly multiplexing the data from multiple
channels into a single output stream. The signal-processing function also
provides the edge-timing reference and gating-enable signal to the demodulators to activate the burst acquisition for each assigned burst slot. The
signal-processing function may also provide an indication of successful decoding, decoding error, or fail-to-decode for each codeword and the number of
corrected ReedSolomon symbols in each codeword. For every upstream
burst, the CMTS has a priori knowledge of the exact burst length in symbols.
17.3.3.10 Modulation Formats and Rates. The upstream modulatorr
demodulator providesrsupports both QPSK and 16-QAM modulation formats. The upstream modulator using QPSK provides 160, 320, 640, 1280, and
2560 kilo-symbols per second, and at 16-QAM it provides 160, 320, 1280, and
2560 kilo-symbols per second. The variety of modulation rates and flexibility
in setting upstream carrier frequencies permits operators to position carriers
in gaps in the pattern of narrowband ingress. This is discussed in Appendix F
of the reference publication DOCSIS, Ref. 3..
17.3.3.11 Symbol Mapping. The modulation mode, whether QPSK or
16-QAM, is programmable. The symbols transmitted in each mode and the
mapping of the input bits to the I and Q constellation is defined in Table
17.3. In the table, I 1 is the MSB of the symbol map, Q 1 is the LSB for QPSK,
and Q 0 is the LSB for 16-QAM. Q 1 and I 0 have intermediate bit positions in
16-QAM. The MSB is the first bit in the serial data into the symbol mapper.
The upstream QPSK symbol mapping is shown in Figure 17.14. The 16-QAM
noninverted Gray-coded. symbol mapping is shown in Figure 17.15. The
16-QAM differential symbol mapping is shown in Figure 17.16.
If differential quadrant encoding is enabled, the currently transmitted
symbol quadrant is derived from the previously transmitted symbol quadrant
and the current input bits are derived in accordance with Table 17.4.
TABLE 17.3
Ir
rQ Mapping
QAM Mode
Input Bit Definitions
QPSK
I1 Q 1
16-QAM
I1 Q 1 I 0 Q 0
705
Figure 17.14. QPSK symbol mapping.
Figure 17.15. 16-QAM Gray coded symbol mapping. (From Figure 4-3, page 23, DOCSIS
Specification, Ref. 3.)
Figure 17.16. 16-QAM differential-coded symbol mapping. (From Figure 4-3, page 23,
DOCSIS Specification, Ref. 3.)
706
TABLE 17.4
Derivation of Currently Transmitted Symbol Quadrant
Current Input
Bits I1 Q 1
Quadrant Phase
Change
00
MSBs of Previously
Transmitted Symbol
MSBs for Currently

Transmitted Symbol
11
11
00
01
01
00
00
00
00
10
10
01
90
11
01
01
90
01
00
01
90
00
10
01
90
10
11
11
180
11
00
11
180
01
10
11
180
00
11
11
180
10
01
10
270
11
10
10
270
01
11
10
270
00
01
10
270
10
00
Spectral Shaping. The upstream PMD layer shall support 25% Nyquist
square root cosine shaping. The occupied spectrum shall not exceed the
channel widths shown in Table 17.5. Note that the DOCSIS specification
defines the channel width as the y30 dB bandwidth. .
Upstream Frequency Agility and Range. The upstream PMD sublayer supports operation over the frequency range of 542 MHz edge to edge. Offset
frequency resolution should be supported with a range of "32 kHz increment s 1 Hz; implement within "10 Hz..
TABLE 17.5
Maximum Channel Width
Symbol Rate (ksym r s)
Channel Width (kHz)
160
200
320
400
640
800
1280
1600
2560
3200
707
Spectrum Format. The upstream modulator must provide operation with the
format s t . s I 9t .* cos t . y Q t .* sin t ., where t denotes time and
denotes angular frequency.
FEC Encode
FEC Encode Modes. The upstream modulator must be able to provide the
following selections: ReadSolomon codes over GF256. with T s 1 to 10 or
no FEC coding.
The following ReedSolomon generator polynomial must be supported:
g x . s x q 0 . x q 1 . x q 2Ty1 .
where the primitive element alpha is 0 = 02 hex.
The following ReedSolomon primitive polynomial must be supported:
p x . s x8 q x 4 q x 3 q x 2 q 1
The upstream modulator must provide codewords from a minimum size of 18
bytes 16 information bytes w k x plus two parity bytes for T s 1 error correction. to a maximum size of 255 bytes k-bytes plus parity-bytes .. The uncoded
word size can have a minimum of one byte.
In Shortened Last Codeword mode, the CM must provide the last codeword of a burst shortened from the assigned length of k data bytes per
codeword as described in Section 4.2.10.1.2 of the reference document.
The value of T must be configured in response to the Upstream Channel
Descriptor from the CMTS.
FEC Bit-to-Symbol Ordering. The input to the ReedSolomon Encoder is
logically a serial bit stream from the MAC layer of the CM, and the first bit
of the stream is mapped into the MSB of the first ReedSolomon symbol
into the encoder. The MSB of the first symbol out of the encoder is mapped
into the first bit of the serial bit stream fed to the Scrambler.
Note: The MAC byte-to-serial upstream convention calls for the byte LSB
to be mapped into the first bit of the serial bit stream per Section 6.2.1.3 of
the reference document.
Scrambler. The upstream modulator will implement a scrambler shown in
Figure 17.17. where the 15-bit seed value should be arbitrarily programmable.
At the beginning of each burst, the register is cleared and the seed value is
loaded. The seed value should be used to calculate the scrambler bit that is
combined in an XOR with the first bit to data of each burst which is the
MSB of the first symbol following the last symbol of the preamble..
708
Figure 17.17. Scrambler block diagram. (From Figure 4-4, page 26, DOCSIS Specification,
Ref. 3.)
The scrambler seed value is configured in response to the Upstream

Channel Descriptor from the CMTS.
The polynomial is x 15 q x 14 q 1.
Preamble Prepend. The upstream PMD sublayer supports a variable-length
preamble field that is prepended to the data after they have been randomized and ReedSolomon encoded. The first bit of the Preamble Pattern is
the first bit into the symbol mapper see Figure 17.18. and is I1 in the first
symbol of the burst. The first bit of the Preamble Pattern is designated by the
Preamble Value Offset as described in Table 6-15 and Section 6.3.2.2 of the
reference document.
The value of the preamble that is prepended is programmable and the
length is 0, 2, 4 . . . , or 1024 bits for QPSK and 0, 4, 8, . . . , or 1024 for
16-QAM. Thus, the maximum length of the preamble is 512 QPSK symbols
or 256 QAM symbols.
Figure 17.18 is a block diagram of the signal processing sequence.
Figure 17.18. Functional block diagram showing signal processing sequence. (From Figure
4-8, page 35, DOCSIS Specification, Ref. 3.)
709
17.3.3.12 Burst Profiles. The transmission characteristics are separated

into three portions:
a. Channel parameters
b. Burst profile attributes
c. User unique parameters
The channel parameters include: 1. the symbol rate there are rates from
160 ksymrs to 2.56 Msymrs in octave steps., 2. center frequency Hz., and
3. the 1024-bit preamble superstring. The channel parameter characteristics
are shared by all users on a given channel. The burst profile attributes are
listed in Table 17.6. These parameters are shared attributes corresponding to
a burst type. The user-unique parameters may vary for each user even when
using the same burst type on the same channel as another user for example,
power level. and are listed in Table 17.7.
The CM generates each burst at the appropriate time as conveyed in the
mini-slot grants provided by the CMTS MAPs MAP stands for multiaccess
protocol.. The CM supports all burst profiles that are commanded by the
CMTS via burst descriptors in the UCD upstream channel descriptor . and
that are subsequently assigned for transmission in a MAP.
The CM implements the offset frequency to within "10 Hz.
The CM is capable of switching burst profiles with no reconfiguration time
required between bursts except for changes in the following parameters: 1.
output power, 2. modulation, 3. symbol rate, 4. offset frequency, 5.
channel frequency, and 6. ranging offset.
TABLE 17.6
Burst Profile Attributes
Burst Profile Attributes
Configuration Settings
Modulation
QPSK, 16 QAM
Diff Enc
OnrOff
Preamble length
0 to 1024 bits (Note Section 4.2.5 of Ref. 3)
Preamble value offset
0 to 1022
FEC error correction (T bytes)
0 to 10 (0 implies FEC s off)
FEC codeword information bytes (k)
Fixed: 16 to 253 (assuming FEC on)

Shortened: 16 to 253 (assuming FEC on)
Scrambler seed
15 bits
Maximum burst length (minislots) a
0 to 255
Guard time
5 to 255 symbols
Last codeword length
Fixed, shortened
Scrambler onroff
On r Off
A burst length of 0 mini-slots in the channel profile means that the burst length is variable on that
channel for that burst type. The burst length, while not fixed, is granted explicitly by the CMTS to the
CM in the MAP.
710
For symbol rate, offset frequency, and ranging offset, the CM transmits
consecutive bursts as long as the CMTS allocates at least 96 symbols in
between the last symbol center of one burst and the first symbol center of the
following burst. The maximum reconfiguration time of 96 symbols should
compensate for the rampdown time of one burst and the rampup time of the
next burst as well as the overall transmitter delay time including the pipeline
delay and optional pre-equalizer delay. For modulation-type changes, the
CM transmits consecutive bursts as long as the CMTS allocates at least 96
symbols in between the last symbol center of one burst and the first symbol
center of the following burst. Output power, symbol rate, offset frequency,
channel frequency, and ranging offset should not be changed until the CM is
provided sufficient time between bursts by the CMTS. Transmitted output
power, symbol rate, offset frequency, channel frequency, and ranging offset
are not changed while more than y30 dB of any symbols energy of the
previous burst remains to be transmitted, or more than y30 dB of any
symbols energy of the next burst has been transmitted, excluding the effect
of the transmit equalizer if present in the CM.. Negative ranging offset
adjustments will cause the 96 symbol guard to be violated. The CMTS assures
711
that this does not happen by allowing extra guard time between bursts that is
at least equal to the amount of negative ranging.
If the channel frequency is to be changed, then the CM changes between
bursts as long as the CMTS allocates at least 96 symbols plus 100 ms between
the last symbol center of one burst and the first symbol of the following burst.
The output transmit power is maintained constant within a TDMA burst to
within less than 0.1 dB excluding the amount theoretically present due to
pulse shaping, and amplitude modulation in the case of 16-QAM..
Burst Timing Convention. Figure 17.19 illustrates the nominal burst timing.
Figure 17.20 shows worst-case burst timing. In this example, burst N
arrives 1.5 symbols late, and burst N q 1 arrives 1.5 symbols early, but
separation of 5 symbols is maintained; 8-symbol guardband is shown.
At a symbol rate of Rs, symbols occur at a rate of one each Ts s 1rRs
seconds. Rampup and rampdown are the spread of a symbol in the time
domain beyond Ts duration owing to the symbol-shaping filter. If only one
symbol were transmitted, its duration would be longer than Ts due to the
shaping filter impulse response being longer than Ts. The spread of the first
and last symbols of a burst transmission effectively extends the duration of
the burst to longer than N = Ts, where N is the number of symbols in the
burst.
Frame Structure. Figure 17.21 shows two examples of frame structure: one
where the packet length equals the number of information bytes in a
codeword, and another where the packet length is longer than the number of
information bytes in one codeword, but less than two codewords. Example 1
illustrates the fixed codeword-length mode, and Example 2 illustrates the
shortened last codeword mode. These modes are defined below.
Codeword Length. The CM operates in a fixed-length codeword mode or
with the shortened codeword capability enabled. The minimum number of
information bytes in a codeword, for fixed or shortened mode, is 16 bytes.
Shortened codeword capability is available with k G 16 bytes, where k is the
number of information bytes in a codeword. With k - 16, shortened codeword capability is not available.
Upstream Demodulator Input Power Requirements. The maximum total
input power to the upstream demodulator should not exceed 35 dBmV in the
5- to 42-MHz frequency range of operation. The intended receive power in
each carrier should be within the values shown in Table 17.8. The demodulator operates within its defined performance specifications with received
bursts within "6 dB of the nominal commanded received power.
Upstream Electrical Output of the CM. The CM operates with an RF output
modulated signal with the characteristics given in Table 17.9.
712
713
Figure 17.20. Worst-case burst timing. (From Figure 4-6, page 30, DOCSIS Specification,
Ref. 3.)
Figure 17.21. Example frames structure with flexible burst length mode. (From Figure 4-7,
page 34, DOCSIS Specification, Ref. 3.)
714
TABLE 17.8 Maximum Range of Commanded Nominal Receive

Power in Each Carrier
Symbol Rate
(ksymrs)
Maximum Range (dBmV)
160
y16 to q14
320
y13 to q17
640
y10 to q20
1280
y7 to q23
2560
y4 to q26
TABLE 17.9
Electrical Output From CM
Parameter
Value
Frequency
5 to 42 MHz edge to edge
Level range (one channel)
q8 to q55 dBmV (16-QAM)

q8 to q58 dBmV (QPSK)
Modulation type
QPSK and 16-QAM
Symbol rate (nominal)
160, 320, 640, 1280, and 2560 ksymrs
Bandwidth
200, 400, 800, 1600, and 3200 kHz
Output impedance
75 ohms
Output return loss
) 6 dB (542 MHz)
Connector
F connector per [IPS-SP-401] (common with the input)
Downstream. The downstream PMD sublayer conforms to ITU-T Rec.

J.83, Annex B, for low-delay video applications with the exception of the
scalable interleaving discussed below.
Scalable Interleaving to Support Low Latency. The downstream PMD sublayer supports a variable-depth interleaver with the characteristics defined in
Table 4-11 of the DOCSIS reference publication. It should be noted that the
table contains a subset of the interleaver modes found in ITU-T Rec. J.83-B.
Downstream Frequency Plan. The downstream frequency plan should comply with harmonic related carrier HRC., incremental related carrier IRC.,
or standard STD. North American plans in accordance with EIA-542.
However, operation below a center frequency of 91 MHz is not required.
CMTS OutputElectrical. The CMTS has an output that is an RF modulated signal with the characteristics given in Table 17.10.
715
716
TABLE 17.11
CM BER Performance
Adjacent
Channel
Item
BER
Image Reject
64-QAM
Implementation loss
such that CM achieves
a post-FEC BER better
than or equal to 10 y 8
when E b rN0 s 23.5
dB or better.
Meeting BER performance requirements,

with analog or digital
signal at q10 dBc in
any portion of the band
other than adjacent
channels.
Meeting BER performance requirements

with digital signal at
0 dBc in adjacent
channels. With analog
signal in adjacent
channels at q10 dBc.
With additional 0.2 dB
allowance with "10
dBc in adjacent chan nels.
256-QAM
Implementation loss of
CM is such that CM
achieves a post-FEC
BER of equal or better
than 10 y 8 with an
E s rN0 of 30 dB or
more.
Meeting BER performance requirements,

signal at q10 dBc in
any portion of RF band
other than adjacent
channels.

signal at 0 dBc in adjacent channels.
with an additional 0.5
dB allowance, shall be
met with analog signal
at q10 dBc in adjacent
channels.
with an additional 1.0
dB allowance, shall be
met with digital signal
at q10 dBc in adjacent
channels
Source: Derived from material in Section 4.3.6, Ref. 3.
Downstream Electrical Input to the CM. The CM accepts an RF-modulated

signal with the characteristics shown in Table 17.10.
CM BER Performance. The bit error rate performance of a CM is described in Table 17.11. The requirements apply to the I s 128, J s 1 mode of
interleaving.
17.4
LOCAL MULTIPOINT DISTRIBUTION SYSTEM (LMDS)
LMDS as we know it today evolved out of a one-way radio wireless .

broadband cellular system that delivered conventional 6-MHz analog TV
channels to residential and business customers. This system was developed by
717
Bernard Bernie. Bossard whose company became Cellular Vision USA. Its
beta tests were carried out in a test bed in Brooklyn, New York in the period
1989 to 1995. The system carried 49 TV channels in an FDM-type configuration where the RF modulation was FM. By digitizing and then applying
MPEG compression techniques, it was found that the number of operational
TV channels was tripled over the original system for a given RF bandwidth
Ref. 4..
The operational frequency band was the 28- to 30-GHz band. In this band,
LMDS competes for usage with conventional LOS microwave and communication satellite feeder uplinks. The U.S. FCC broke up the band so all parties
would be satisfied. Keep in mind that LMDS was really a wireless CATV
technique. With implementations being fielded today it serves as a two-way
configuration. Here, of course, downstream means from an LMDS hub to
end-user, and upstream operation is from the end-user to the LMDS hub.
The specific LMDS frequency bands in the United States are 27.5 to 28.35
GHz 850-MHz bandwidth ., 29.1 to 29.25 GHz 150-MHz bandwidth ., and
31.075 to 31.225 GHz 150-MHz bandwidth .. The range from an LMDS hub
to the furthermost end-user depends on rainfall analysis. It can vary from
1 mile 1.6 km. to 5 miles 8.0 km. Refs. 9 and 10..
The LMDS concept is shown in Figure 17.22. In the figure a hub is a
facility that houses a radio wireless . transmitter and receivers. which
provide connectivity with end-users. The hub connects through a fiber-optic
cable or LOS microwave radiolink wireless . with a central processing and
routing unit. This unit may be colocated with a local switching center
providing interface with the PSTN.
The antenna at each hub will normally be omnidirectional with q10- to
q15-dBi gain, whereas the antenna at an end-user location will be directional with a q35-dBi gain. Cost and performance dictate that the RF front
end be colocated with the antenna. The antenna unit feeds and accepts IF
from an inside unit which may be likened to a CATV set-top box. The inside
unit accepts the IF signal from the antenna and provides, in the transmit
direction, an IF signal to the antenna unit.
LMDS provides different levels of services. A residence accepts a more
modest service capability, such as 1 Mbps of downstream internet, symmetric
upstream and downstream data in increments of 64 kbps, and from one to
eight POTS plain old telephone service. lines and equivalent CATV
downstream TV service using MPEG-2 frame video. Depending on requirements, a business may require one or more DS1 or E1 symmetric line
equivalents, several 1-Mbps downstream internet connections, or one 5-Mbps
connectivity to an on-premises server as well as several frame relay connectivities andror packet service based on TCPrIP. They may not need or want
the entertainment TV. Many of the configurations in North America have
turned to using a version of DOCSIS Section 17.3. providing POTS, TV,
data, and, of course, internet.
As we see from this discussion, the LMDS frequency band is split into
three segments. We could make the assignment highly asymmetrical with
718
Figure 17.22. Sketch illustrating the LMDS concept. Both residential and business services
are shown.
1000 MHz assigned for downstream and 150 MHz for upstream. For efficient
operation, frequency reuse is mandatory.
17.4.1
Frequency Reuse
The original LMDS attempted frequency reuse, with some success. It depended on free space loss, polarization isolation, and some frequency offset
of television radio carriers. Although the value of polarization isolation was
719
Figure 17.23. Cell B hub transmitter antenna faces the back of closest cell A end-user
antenna. Cell radius is 5 km. Gain at the back of receiving antenna is y10 dBi. (Note:
y10 dBi based on ITU-R model.)
claimed to be some 36 dB, it was more on the order of 30 dB. Polarization

isolation works on the premise that adjacent cells will be on opposite
polarizations. Cell A, for example, would transmit its signal with horizontal
polarization, and adjacent cell B would be on vertical polarization. If we
measured the level of cell As signal at cell Bs receiver, it would be down
attenuated . at least 30 dB due to the polarization alone. Actually it would be
down much more because of the free space loss between As location and Bs.
Remember, in this model cell A and cell B transmit on the same frequency.
In this first case the isolation at the receiver at cell A from the transmit
signal of cell B, which is on the same frequency but different polarization,
depends on see model in Figure 17.23.*:
Free space loss at 30 GHz assuming that cell radius of 5 km is

y135.9dB
Polarization isolation
y30
dB
.
Receive antenna discrimination back of antenna
y 10
dB
Total path losses
y175.9 dB
Transmit power is 1 watt for a 10-Mbps bitstream; transmit antenna gain
is q14 dBi, thus the EIRP is q14 dBW.
The isotropic receive level IRL. is the difference between two values
i.e., the EIRP and the FSL.: y161.9 dBW.
Add the receive antenna gain i.e., y10 dBi.. This is the RSL receive
signal level.: y171.9 dBW.
The receiver has a 6-dB noise figure. Its thermal noise threshold is
y198 dBWrHz N0 ..
*For a tutorial on the design of wireless radio. systems, consult Roger L. Freeman, Radio
System Design for Telecommunications, 2nd edition, John Wiley, New York, 1997 Ref. 5..
720
Figure 17.24. Interference level into receiver in cell A where cell Bs transmit antenna main
beam is oriented such that it directly enters cell A receive antenna main beam. Distance is
15 km.
The energy per interference bit s y171.9 dBW y 10 log 10 = 10 6 s

y241.9 dBWrHz s I0 .
The I0rN0 s y241.9 y y198. s y43.9 dB. This value shows that
co-channel interference based on the model presents a very acceptable
value.
For a description of calculation of interference levels, consult Ref. 5,

Chapter 13..
This tells us that same frequency operation is feasible with this geometry
where the offending transmit antenna faces the back of the victim receive
antenna, distance 5 km.
We must examine another model that may not be so forgiving. The model
is illustrated in Figure 17.24. This is where the offending transmitter of cell B
is on the other side of the cell circle and it is looking right down the beam of
the receive antenna in an adjacent cell A circle.* Here the victim receive
antenna is at a 15-km distance from the offending transmitter. It is on the far
side of the 5-km-range circle; and the offending transmitter, of course, is in
the center of its range circle. Thus, the total range is 15 km 5 q 5 q 5.. Now
the offending transmitters potentially interfering signal has a free space loss
at the victim receiver of 145.5 dB. Consult Table 17.12 which is used to
determine the interference potential in this particular case.
The interference signal would essentially reduce the system margin of the
wanted signal by 1.5 dB. It would only seriously affect the system where, by
happenchance, a hub transmit antenna was facing directly into the aperture
of the customer receiver antenna. The chances are very slim that this would
occur. Thus we still hold that same frequency operation is quite feasible.
Note: A tutorial on decibels and their applications may be found in
Appendix C of Ref. 6.
*Commonly, hexagons are used to represent cells because they can be arranged without overlap.
We use circles because they are more realistic for propagation studies.
TABLE 17.12
721
Calculation Table
Parameter
Wanted Signal
)a
Interference Signal
EIRP (effective isotropic radiated power
q14 dBW
y14 dBW
FSL (free-space loss)
y135.97 dB
y145.51
IRL (isotropic rec level)
y123.97 dBW
y131.5 dBW
Polarization loss
0.0 dB
y30 dB
Sum
y123.97 dBW
y161.5 dBW
Antenna gain
q35 dBi
q35 dBi
RSL
y88.97 dBW
y126.5 dBW
(1) E b , (2) I0
y158.97 dBWrbit
y196.5 dBWrHz
N0
y198 dBW
y198 dBW
(1) E b rN0 , (2) E b r( N0 q I0 )
39.03 dB
35.21 dB
Required E b r( N0 q I0 ), 16
QAM, BER s 1 = 10 y 8
Margin
16 dB
23.03 dB
Note that the EIRP is the sum in dB units of the transmit power and the antenna gain. Transmission
line losses have been neglected.
Other designers take a different position and divide the 800-MHz downstream band into four 200-MHz bands. One well call frequency family A, the
next one well call frequency family B, the next C, and the last D. They will
occupy a single group of adjacent cells with the same nomenclature i.e., A,
B, C, and D.. This concept is illustrated in Figure 17.25.
For improved isolation, 90-degree sectorized antennas are used which
would appear more or less as in Figure 17.26. Here we would have a
90-degree pattern group on the line between cells D and C.
Propagation is a most challenging problem in LMDS design. In the models
we used above, it is assumed that the receive antenna is oriented correctly
toward the transmit antenna, and they are in line-of-sight of each other. This
means that the two antennas are visible one-to-the-other, and there is
sufficient clearance of the ray beam to obstacles in its path. This feature
alone is difficult to achieve for all receive antennas being served by this nodal
transmit antenna. Verdure e.g., tree leaves. can block the signal; there will
be diffraction paths and signal reflections. These latter conditions can cause
multipath conditions and will disperse the signal.
Rainfall is yet another matter. In areas of heavy downpour rain such as in
Florida and Louisiana, path lengths can be severely limited, to about some
2-km maximum. In other areas of the country which do not consistently have
such conditions, maximum path lengths can be lengthened to up to 5 or 6 km
3.1 or 3.7 mi.., typically Los Angeles and Phoenix. When designing an
LMDS system, it is mandatory that a rainfall study be carried out. Methodology for these studies can be found in Radio System Design for Telecommunications, 2nd edition Ref. 5..
722
Figure 17.25. Splitting the downstream 800-MHz bandwidth into four 200-MHz bandwidths
for improved frequency reuse and to minimize co-channel interference. A 90-degree sectorized pattern is shown. Four transmitters, each with its own 90-degree sector antenna, sit at
the center of the pattern. (From Figure 5, page 4, H-P Application Note 1296, Ref. 7.)
Figure 17.26. A large grouping of frequency reuse cells using a family of four center
frequencies for downstream: A, B, C, and D. This is identical to a cellular radio pattern where
N s 4. For our model, the cells have a 5-km radius.
723
The only way to improve system time availability is to optimize link

margin. Time aailability may be defined as the percentage of time that a
system is operational meeting its BER performance requirements. Link
margin represents those extra, surplus decibels left over when doing the link
budget. One way to increase link margin is to shorten link length; another
way is to use larger power amplifiers, reduce bit rate, use forward error
correction FEC., and increase antenna aperture at the user site.
Suppose we were to use some of the link parameters from the interference
link budget in Table 17.12 and call it the signal link budget; however, the
maximum distance required for signal transmission is 10 km 2 radii or
2 = 5.. In this case, the transmit and receive antennas are on the same
polarization, so the polarization isolation loss drops out. We have
EIRP
FSL
Receive antenna gain
RSL
q14 dBW
y142 dB 30 GHz, 10 km.
q35 dBi
y93 dBW
Convert to energy spectral density for 10 Mbps:

RSL 0 s y93 y 10 log 10 = 10 6
s y163 dBWrHz s Eb
The receiver noise figure is 6 dB. Then
N0 s y204 dBW q 6 dB
s y198 dBWrHz
EbrN0 s y163 dBWrbit y y198 dBWrHz .
s 35 dB
The required EbrN0 is 11 dB for QPSK modulation and BER s 1 = 10y10 .
The calculated EbrN0 s 35 dB
The required EbrN0 s 11 dB
Remainder subtract . s 24 dB link margin .
The link margin in this case is 24 dB. These are the dBs left over, which we
discussed previously. They are put to very good use. First, we include them in
the calculations for rain margin and they become the rain margin. Simultaneously they help to burn the signal through verdure, to penetrate buildings, to
help in refracting a signal around corners, and for the overall improvement of
coverage. The question arises: Why not increase the margin? We are using a
724
1-watt HPA high power amplifier. now; increase this to 2 watts, and our
margin goes up 3 dB. A 2-watt amplifier costs something in the order of three
times the cost of a 1-watt amplifier. We can see that the price tends to go up
exponentially. There is a point where increasing amplifier power is not
cost-effective any more. If we were to cut the radius of the cell in half, we
would increase the margin about 6 dB.
As the cost of the system increases, the cost per CPE customer premise
equipment. goes up. One goal may be to limit the cost of CPE to $300 or
thereabouts. So every system design change has implications on CPE cost.
Suppose we increase antenna height and HPA power. This we find adds two
thousand more possible customers to the system. How does this system
change affect the cost per CPE?
Business customers most assuredly will be handled differently than residential customers. They will demand more telephone service, more data
service, and more internet ports. The $300 goal must be discarded. Other
cost goals should be established depending on the level of business.
For larger businesses say on the order of ) 200 FX telephone lines, ten
internet ports, a 45-Mbps frame relay port, and two 10-Mbps IP
portsthought must be given to the use of another service type such as LOS
microwave at 23 GHzfor example, with a link to the hub. This separate
radiolink would serve all the datartelephony requirements of the business in
question. It may also be shared with other larger businesses in the same
building, and other radio carriers can be added as required.
17.5
17.5.1
OTHER METHODS TO BREACH THE LAST MILE

Remote Wireless (Radio) Bridges
These systems originally operated in single, comparatively long-range pointto-point connections. Such simple architectures still predominate; however,
point-to-multipoint products are now appearing in the marketplace. ISPs
providing wireless last-mile connections have formed the largest market for
such systems. Some systems provide as much as 100-Mbps connectivities, and
many systems provide up to 30 Mbps. There is essentially no standardization.
Compliance with IEEE 802.11 wireless LAN, Section 10.1.5. is not even
contemplated by equipment manufacturers of these systems.
The 5.8-GHz ISM band is widely used for wireless bridge applications.
FCC Part 18.301 stipulates that operation shall be in the band 5.725 to 5.875
GHz, providing a 150-MHz bandwidth. Even with this wide bandwidth,
frequency reuse sectorized antennas is commonplace.
Point-to-multipoint systems are beginning to appear on scene using some
of the concepts discussed in Section 17.4. There are at least two manufacturers of multipoint-to-multipoint systems. The principal market addressed is
that of ISP connectivities. No base stations are used, and each node is a
REFERENCES
725
router. Network intelligence is distributed throughout, and the network

adapts instantaneously to changing interference and traffic levels. These
nodes are connected in a mesh network where every and all stations are
connected one to the other Ref. 8..
17.5.2
In-Building Wireless Telephone Systems
This is a most pragmatic application of wireless systems where offices have

many moves and rearrangements. The wireless PABX, along with the wireless LAN WLANsee Section 10.1.5., makes moves and rearrangements a
breeze. Of course we assume telephone number portability. This means that
when you are moved to another office in this or a nearby building, you keep
the same telephone number.*
The IP telephone takes on wider applications allowing voice to be used
over IP along with numerous data applications. It consolidates voice and data
into unified network. Wireless IP telephony operates in the 2.4-MHz ISM
band rather than the conventional 1.9-MHz PCS band.
REFERENCES
1. Roger L. Freeman, Telecommunications Transmission Handbook, John Wiley &
Sons, New York, 1998.
2. The DSL Sourcebook, Paradyne Corporation, Lago, FL, 1997.
3. Data-Oer-Cable Serice Interface Specification, Revision 1, Radio Frequency
Interface Specification, SP-RFI-980202, SCTE, Exton, PA, February 1998.
4. Various private communications between Bernard Bossard and Roger Freeman
dealing with early design issues of LMDS as well as formulation of FCC Position
Papers for Cellular Vision New York and USA, February 1988 through December
1995.
5. Roger L. Freeman, Radio System Design for Telecommunications, 2nd edition,
John Wiley, New York, 1997.
6. Roger L. Freeman, Fundamentals of Telecommunications, John Wiley & Sons,
New York, 1999.
7. LMDSThe Wireless Interactie Broadband Access, HP Application Note 1296,
Hewlett-Packard Co., Englewood, CO, 1997.
8. Wireless Integration, Buyers Guide, Pennwell, Nashua, NH, 2000.
9. Peter B. Papazian et al., Study of the Local Multipoint Distribution Serice Radio
Channel, IEEE Transactions on Broadcasting, Vol. 43, No. 2, IEEE, New York,
June 1997.
10. Gene Heftman, LMDS Set to Challenge for Last-Mile Supremacy, Microwaes
and RF, April 1999.
*Telephone number portability can be extended to an entire area code, statewide, countrywide,
and even worldwide. It can take on this connotation.
18
NETWORK MANAGEMENT
FOR ENTERPRISE NETWORKS
18.1
WHAT IS NETWORK MANAGEMENT?
Effective network management optimizes a telecommunication networks

operational capabilities. The key word here is optimizes. How does this
function optimize?
It keeps the network operating at peak performance.

It informs the operator of deteriorating performance.
It provides the tools for pinpointing causes. of performance deterioration or failure.
It provides easy alternative routing and work-arounds when deterioration andror failure take place.
It serves as the front-line command post for network survivability.
There are numerous secondary functions of network management. They are

important but, in our opinion, still secondary. Among these items are:
18.2
It informs in quasi-real time regarding network performance.

It maintains and enforces network security, such as link encryption code
changes, state variables., issuance, and use of passwords.
It gathers and files data on network usage.
It performs a configuration management function.
It also performs an administrative management function.
THE BIGGER PICTURE
Many seem to view network management as a manager of data circuits only.

There is a much bigger world out there. Numerous enterprise and government networks serve for the switching and transport of multimedia communi727
728
NETWORK MANAGEMENT FOR ENTERPRISE NETWORKS
cations. The underlying network will direct switch . and transport voice, data,
and image traffic. Each will have a traffic profile notably differing from the
other. Nevertheless, they should be managed as an entity. It is more cost
effective to treat the whole than to treat the parts.
There is a tendency in the enterprise scene to separate voice telephony
calling it telecommunications. and data communications. This is unfortunate
and a major error on the part of management. Perhaps that is why network
management seems to often operate on two separate planes. One is data and
very sophisticated, and the other is voice, which may have no management
facilities at all. This section will treat network management as a whole
consisting of its multimedia parts: voice, image, and data, which includes
facsimile, telemetry, and CADrCAM.
18.3
TRADITIONAL BREAKOUT BY TASKS
There are five tasks traditionally involved with network management:
Fault management
Configuration management
Performance management
Security management
Accounting management
18.3.1
Fault Management
This is a facility that provides information on the status of the network and
subnetworks. The information on the status should not only display faults
i.e., failures . and their location, but should also provide information on
deteriorated performance. One cause of deteriorated performance is congestion. Thus, ideally, we would like to isolate the cause of the problem.
Fault management also includes the means to bypass troubled sections of
a networkthat is, the means to patch in new equipment for deteriorated or
failed equipment.
The complexity of modern telecommunication networks is such that as
many network management tasks as possible should be automated. All
displays, readouts, and hard copy records should be referenced to a network
time base down to 0.1 s. This helps in correlating events, an important
troubleshooting tool.
18.3.2
Configuration Management
Configuration management establishes an inventory of the resources to be

managed. It includes resource provisioning timely deployment of resources
to satisfy an expected service demand. and service provisioning assigning
services and features to end users.. It identifies, exercises control over,
SURVIVABILITY WHERE NETWORK MANAGEMENT REALLY PAYS OFF
729
collects data from, and provides data to the network for the purpose of
preparing for, initializing, starting, and providing for the operation and
termination of services. Configuration management deals with equipment
and services, subnetworks, networks, and interfaces. Its functions are closely
tied to fault management, as we have defined it previously.
18.3.3
Performance Management
Performance management is responsible for monitoring network performance to ensure that it is meeting specified performance. Some literature
references Ref. 1. add growth management. They then state that the
objective of performance and growth management is to ensure that sufficient
capacity exists to support end-user communication requirements.
Of course, there is a fine line defining network capacity. If too much
capacity exists, there will probably be few user complaints, but there is excess
capacity. Excess capacity implies wasted resources, thus wasted money.
Excess capacity, of course, can accommodate short-term growth. Therefore
performancergrowth management provides vital information on network
utilization. Such data provide the groundwork for future planning.
18.3.4
Security Management
Security management controls access to and protects both the network and
the network management subsystem against intentional or accidental abuse,
unauthorized access, and communication loss. It involves link encryption,
changes in encryption keys, user authentication, passwords, and unauthorized
usage of telecommunication resources.
18.3.5
Accounting Management
Accounting management processes and records service and utilization

records. It generates customer billing reports for services rendered. It can
identify costs and establish charges for the use of services and resources in
the network. In other words, accounting management measures the usage of
an enterprises one, several, or many networks. It is a repository of data
dealing with enterprise network plant-in-place information, along with usage
of that plant, both internal LANs. and external WANs.. The network usage
data is vital for planning future upgrades and expansion.
18.4 SURVIVABILITYWHERE NETWORK MANAGEMENT

REALLY PAYS OFF
The network management center is the front-line command post for the
battle for network survivability. We can model numerous catastrophic events
affecting a telecommunications network, whether the public switched
730
telecommunication network PSTN. or a privaterenterprise network. Among

such events are fires, earthquakes, floods, hurricanes, terrorism, and public
disorder. Telecommunications has brought revolutionary efficiencies to the
way we do business and is thus very necessary, even for life itself. Loss of
these facilities could destroy a business, even possibly destroy a nation. A
properly designed network management system could mitigate losses, even
save a network almost in its entirety.
Cite the World Trade Center explosion in New York City. The entire
PABX telephone system survived, except for extensions in the immediate
area of the bomb, simply because the PABX was installed on the sixth floor.
All offices in the building had communications within the building and with
the outside world. Thus vital communications were never lost.
This brings in the first rule toward survivability. A network management
center is a place of point failure. Here we mean that if the center is lost,
probably nearly all means of reconfiguring the network are lost. To avoid
such a situation, a second network management center should be installed.
This second center should be geographically separate from the principal
center. It is advisable that the second center share the network management
load and be planned and sized to be able to take the entire load with the loss
of the principal center. There should be a communications orderwire between the two centers. Both centers should be provided with no-break power
and backup diesel generators.
One simple expedient for survivability is to back up circuits with an
arrangement with the local telephone company. This means that there must
be compatibility between the enterprise network and the PSTN as well as
one or more points of interface. A major concern is the network clock. One
easy solution is to have the enterprise network derive its clock from the
digital PSTN.
In the following sections we describe means to enhance survivability still
further.
18.5 AVAILABILITY ENHANCEMENTRAPID

TROUBLESHOOTING
Aailability is defined by the IEEE Ref. 1.: The ability of an itemunder
combined aspects of its reliability, maintainability and maintenance support
to perform its required function at a stated instant or over a stated period
of time. In the availability formula, only two terms may be varied: reliability,
measured as mean time between failures MTBF., and time to repair,
measured as mean time to repair MTTR.. Both are measured in hours.
Availability A . s MTBFr MTBF q MTTR .
Availability values for a network should well exceed 99.9%. The 99.9%
AVAILABILITY ENHANCEMENT RAPID TROUBLESHOOTING
731
figure means that the network will be down i.e., out of operation. 0.1% of
the time or 0.001 = 8760 hours per year or 8.76 hours per year. This value
may be acceptable to a network operator, but for many it is unacceptable.
Suppose we require an availability of 99.999%. In this case the network will
be down 0.001% of the time or 0.0001. Multiply this value by 8760 hours, the
number of hours in a year, and it will be out of service 0.0876 hoursryear or
5.256 minutesryear. To achieve such an excellent availability will be expensive. No way can a fault be recognized, isolated, and repaired in 5 minutes if
we work on yearly increments. To achieve such an availability value, all
circuits must be redundant, all signal paths must also be redundant, and
redundant no-break power would be a requirement.
A well-designed network management system can notably reduce MTTR,
thus improving the availability value. The system can do this in two ways:
1. Warn of deterioration
2. Identify exact location of a fault down to the card level
Telephone companies use two values for MTTR; 2 hours when a fault occurs
in a facility with a craftsperson on duty e.g., at a switch. and 4 hours when a
fault occurs at a remote device such as at a fiber-optic repeater or at a
microwave relay site Ref. 2.. The U.S. Navy often uses 0.3 hours because of
the proximity of servicing personnel and the availability of spare parts.
It should be noted that much of the modern equipment is either equipped
with BITE built-in test equipment. or carries a software function to isolate
and report faults. The BITE information, which is commonly go no-go,
can usually be remoted to a network management facility.
Another approach, which is described later, is the use of software, such as
SNMP simple network management protocol., which carries out
a queryanswer regime automatically regarding equipment status and
performance.
An ideal network management system will advise the operator of one or
more events,* and will reveal where in the network they occurred, as well as
provide handy troubleshooting data. This should include the ability to pinpoint a fault down to the circuit board level. Ideally, the network management system should, in most cases, warn the operator in advance of impending faultrfaults. This is a boon to improving network availability and its
survivability during disaster. In the case of the latter, the network data
reporting system can quickly identify circuits that are still up and functional
and those that are down.
*I define an eent as something out of the ordinary that occurred. In the telecommunication
management arena an event might be a line dropout, switchover to backup power, loss of frame
alignment, and so on.
732
18.5.1
Troubleshooting
Many network faults will be correctly isolated by the network management

system, and in some cases no immediate action may be required by a network
technician. However, other cases may prove to be more difficult. In such
cases the network management system becomes an adjunct to the human
craftsperson assisting in the troubleshooting effort.
We will assume that the troubleshooter will have available certain units of
test equipment to aid in pinpointing the cause of an event.
We consider four steps in the finding fault process:
1.
2.
3.
4.
Observing symptoms
Developing a hypothesis
Testing the hypothesis
Forming conclusions
Observing Symptoms. It should be kept in mind that often many symptoms

will appear at once, usually in a chain reaction. We must be able to spot the
real causal culprit or we may spend hours or even days chasing effects, not
the root cause.
We thus offer four reminders Ref. 3.:
1. Find the range and scope of the symptoms. Does the problem affect all
stations all users? .. Does it affect random users or users in a given
area? We are looking for a pattern here.
2. Are there some temporal conditions to the problem? How often does it
occur per day, per hour, and so on. Is the problem continuous or
intermittent? What is its regularity. Can we set our watches by its
occurrence, or is it random in the temporal context?
3. Have there been recent rearrangements, additions to the network, or
reconfigurations?
4. Software and hardware have different vintages that are called release
dates in the case of software. My computer is using Windows 2000, but
what is its release date? The question one should ask then would be:
Are all items of a certain genre affected the same way? Different
release dates of a workstation operating system may be affected differently, some release dates not at all. The problem may be peculiar to a
certain release date.
Before we can move forward in the troubleshooting analysis, we must have
firmly at hand the troublesome networks baseline performance. What is
meant here is that we must have a clear idea of normal operation of the
network so that we can really qualify and quantify its anomalous operation.
For example, what is the expected BER, and is the value related to a time
AVAILABILITY ENHANCEMENT RAPID TROUBLESHOOTING
733
distribution? Can we express error performance in error-free seconds EFS.,

errored seconds ES., and severely errored seconds SES.? These are defined
in ITU-T Rec. G.821 Ref. 4..
Five network-specific characteristics that the troubleshooter should have
familiarity with or data on are as follows:
1. Network Utilization. What is the average network utilization? How does
it vary through the workday? Characteristics of congestion, if any,
should be known, as well as where and under what circumstances
congestion might be expected.
2. Network Applications. What are the dominant applications on the
network? What version numbers is it running?
3. Network Protocol Software. What protocols are running on the network?
What are the performance characteristics of the software, and are
these characteristics being achieved?
4. Network Hardware. Who manufactured the network interface controllers, media attachment units, servers, hubs, and other connection
hardware? What versions are they? What are their performance characteristics? Expected? Met?
5. Internetworking Equipment. Who manufactured the repeaters, bridges,
routers, and gateways on the network? What versions of software and
firmware are they running? What are the performance characteristics?
What are the characteristics of the interfaced network that are of
interest?
Developing a Hypothesis. In this second step, we make a statement as to
the cause of the problem. We might say that T1 or E1 frame alignment is lost
because of deep fades being experienced on the underlying microwave
transport network. Or we might say that excessive frames being dropped on a
frame relay network is due to congestion being experienced at node B.
Such statements cannot be made without some strong bases to support the
opinion. Here is where the knowledge and experience of the troubleshooter
really pay. Certainly there could be other causes of E1 or T1 loss of frame
alignment; but if underlying microwave is involved, that would be a most
obvious place to look. There could be other reasons for dropping frames in a
frame relay system. Errored frames could be one strong reason.
Testing the Hypothesis. We made a statement, and now we must back it up
with tests. One test I like is correlation. Are the fades on the microwave
correlated with the fade occurrences? That test can be done quite easily. If
they are correlated, we have some very strong indication that the problem is
with the microwave. The frame relay problem may be another matter. First,
we could check the FECN and BECN bits to see if there were a change of
state passing node B. If there is no change, assuming that flow control is
734
implemented, then congestion may not be the problem. Removing the frame
relay from the system and carrying out a bit error rate test BERT. over some
period of time would prove or disprove the noise problem.
A network analyzer is certainly an excellent tool in assisting in the
localization of faults. Some analyzers have preprogrammed tests which can
save the troubleshooter time and effort. Many networks today have some sort
of network monitoring equipment incorporated. This equipment may be used
in lieu of, or in conjunction with, network analyzers. Again we stress the
importance of separating cause and effects. Many times network analyzers or
network management monitors and testers will only show effects. The root
cause may not show at all and must be inferred, or separate tests must be
carried out to pinpoint the cause.
We might digress here to talk about what is often called in Spanish
tonterias. This refers to silly things. Such tonterias are often brought about
by careless installation or careless follow-up repair. Coaxial cable connectors
are some of my favorites. Look for intermittents and cold solder joints. A
good but not necessarily foolproof. tool is a time-domain reflectometer
TDR.. It can spot where a break in a conductor is down to a few feet or less.
It can do the same for an intermittent, when in the fault state. In fact,
intermittents can prove to be a nightmare to locate. An electrically noisy
environment can also be very troublesome.
Forming Conclusions. A conclusion or conclusions are drawn. As we say,
the proof of the pudding is in the eating. The best proof that we were right
in our conclusion is to fix the purported fault. Does it disappear? If so, our
job is done, and the network is returned to its normal baseline . operation.
What conclusions can we draw from this exercise? There are two basic
ingredients to network troubleshooting: 1. expertise built on experience of
the troubleshooter and 2. the availability of basic test equipment. Troubleshooting time can be reduced degraded operation or out-of-service time
reduction. by having on-line network management equipment. With ideal
network management systems, this time can be cut to nearly zero.
18.6
SYSTEM DEPTHA NETWORK MANAGEMENT PROBLEM
An isolated LAN is a fairly simple management problem. There is only a

singular transmission medium, and, under normal operating conditions, only
one user is transmitting information to one or several recipients. It is limited
to only two OSI layers. For troubleshooting, often a protocol analyzer will
suffice, although much more elaborate network management schemes and
equipment are available.
Now connect that LAN to the outside world by means of a bridge or
router and network management becomes an interesting challenge. One
example from experience was a VAX running DECNET which was a station
on a CSMArCD LAN. The LAN was bridged to a frame relay box which fed
SYSTEM DEPTH A NETWORK MANAGEMENT PROBLEM
735
Figure 18.1. A typical multilevel network that employs a network management system. Note
the multiple convergences.
a 384-kbps channel in an E1 hierarchy i.e., 6 E0 channels . via tandemed

microwave links to a large facility at the distant end 550 km. with similar
characteristics. This connectivity is shown diagrammatically in Figure 18.1.
Such is typical of a fairly complex network where we would apply a network
management system. To make it even more difficult, portions of the network
were leased from the local common carrier but were soon cut over to its own
ownership.
18.6.1
Aids in Network Management Provisioning
Modern E1 and DS1 digital systems are provided with a means of operational monitoring of performance. The monitoring is done in quasi-real time
and while operational i.e., in-traffic ..
736
In Chapter 7 we discussed PCM systems and the digital network including

T1DS1. and E1 hierarchies. Let us quickly review their in-service performance monitoring capabilities.
DS1 or T1 has a frame rate of 8000 frames per second. Each frame is
delineated with a framing bit, or F bit. With modern alignment and synchronization algorithms, it is excessive and unnecessary to maintain framing
repetition of the F bit 8000 times per second. Advantage was taken of F-bit
redundancy by the development of the extended superframe ESF.. The ESF
consists of 24 consecutive DS1 frames. With 24 frames we expect to have 24
framing bits. Of these, only 6 bits need to be used for framing, 6 are used for
a cyclic redundancy check CRC-6. on the frame, and the remaining 12 bits
form a 4-kbps data link for network control and maintenance. It is this
channel that can serve as transport for network management information. It
can also serve as an ad hoc test link.
Our present concern is in-service monitoring. For instance, we can get
real-time error performance with the CRC-6, giving us a measure of errored
seconds and severely errored seconds in accordance with ITU-T Rec. G.821
Ref. 4.. Such monitoring can be done with test equipment such as the HP
37702A or HP 37741A.*
Permanent monitoring can be carried out using Newbridge network monitoring equipment, which can monitor an entire T1 or E1 network for frame
alignment loss, errored seconds, and severely errored seconds. One can
look at each link and examine its performance in 15-minute windows for a
24-hour period. The Newbridge equipment can also monitor selected other
data services such as frame relay, which, in our preceeding example, rides on
E1 aggregates. Such equipment can be a most important element in the
network management suite.
The E1 digital network hierarchy also provides capability of in-service
monitoring and test. We remember from Chapter 7 that E1 has 32 channels
or time slots: 30 are used for the payload, and 2 channels or time slots serve
as support channels. The first of these is channel or time slot. 0, and the
second is channel time slot. 16. This latter is used for signaling. Time slot
channel . 0 is used for synchronization and framing. Figure 18.2 shows E1
multiframe structure. It also shows the sequence of bits in the frame
alignment TS0. signal of successive frames. In frames not containing the
frame alignment signal, the first bit is used to transmit the CRC multiframe
signal 001011. which defines the start of the submultiframe SMF.. Alternate frames contain the frame alignment word 0011011. preceded by one of
the CRC-4 bits. The CRC-4 remainder is calculated on all the 2048 bits of
the previous SMF, and the 4-bit word sent as C1, C2, C3, C4 of the current
SMF. Note that the CRC-4 bits of the previous SMF are set to zero before
the calculation is made.
*HP stands for Hewlett-Packard Company, a well-known manufacturer of electronics test
equipment.
SYSTEM DEPTH A NETWORK MANAGEMENT PROBLEM
737
Figure 18.2. E1 multiframe, CRC-4 structure, Notes: E s CRC-4 error indication bits. S a 4 to
S a 8 s space bits. May be used for maintenance (network management) data link. C 1 to
C 4 s cyclic redundancy check-4 (CRC-4) bits. A s remote alarm indication. (From Table
4brG.704, page 81, CCITT Rec. G.704, Ref. 5.)
At the receive end, the CRC remainder is recalculated for each SMF, and
the result is compared with the CRC-4 bits received in the next SMF. If they
differ, then the checked SMF is in error. What this is telling us is that a block
of 2048 bits had one or more errors. One thousand CRC-4 block error checks
are made every second. It should be noted that this in-service error detection
scheme does not indicate BER unless one assumes a certain error distribution random or burst errors. to predict the average errors per block. Rather
it provides a block error measurement.
This is very useful for estimating percentage of errored seconds %ES.,
which is usually considered the best indication of quality for data transmissionitself a block or frame transmission process. CRC-4 error checking is
fairly reliable, with the ability of detecting 94% of errored blocks even under
poor BER conditions see Ref. 6..
Another powerful feature of E1 channel 0 when equipped. is the provision of local indication of alarms and errors detected at the far end. When an
errored SMF is detected at the far end, one of the E bits see Figure 18.2. is
changed from a 1 to a 0 in the return path multiframe TS0.. The local end,
therefore, has exactly the same block error information as the far-end CRC-4
checker. Counting E-bit changes is equivalent to counting CRC-4 block
errors. Thus the local end can monitor the performance of both the go and
return paths. This can be carried out by the network equipment itself, or by a
test set such as the HP 37722A monitoring the E1 2.048-Mbps data bit
738
Figure 18.3. (a) Loopback of a DS0 channel with BERT test in place (intrusive or on a spare
DS0). (b) Loopback of 8-kbps data channel derived from ESF on DS1 or from channel 0 of E1
(nonintrusive testing).
stream. In the same way, the A bits return alarm signals for loss of frame or
loss of signal from the remote end.
Loopback testing is a fine old workhorse in our toolbox of digital data
troubleshooting aids. There are two approaches for DS1 T1. and E1 systems: intrusive and nonintrusive. Intrusive, of course, means that we interrupt traffic by taking one DS0 or E0 channel out of service, or the entire
aggregate. We replace the channel with a pseudorandom binary sequence
PRBS. or other sequence specifically designed to stress the system. Commonly we use conventional bit error rate test BERT. techniques when
looping back. The concept of loopback is illustrated in Figure 18.3. ESF and
Channel 0 data channel testing is nonintrusive. It does not interfere with
customer traffic. Trouble can also be isolated, whether in the go or
return channel of the loopback. Both intrusive and nonintrusive testing is
commonly automated in the network management suite.
Many frame relay equipment also have forms of in-service monitoring as
well as a system of fault alarms. In fact, most complex telecommunication
equipment has built-in monitoring and test features. The problem often is
that these features are proprietary, whereas our discussion of DS1rE1
systems reveals that they have been standardized by ITU-T and Telcordia
recommendations and publications. This is probably the stickiest problem
facing network management systemsthat is, handling, centralizing, and
controlling network management features in a multivendor environment.
AN INTRODUCTION TO NETWORK MANAGEMENT PROTOCOLS
739
18.6.2 Communications Channels for the Network

Management System
A network management facility is usually centrally located. It must monitor
and control distant communications equipment. It must have some means of
communicating with this equipment, which may be widely dispersed geographically. We have seen where DS1 and E1 systems provide a data channel
for OAM operations and maintenance. Higher levels of the DS1 and E1
hierarchies have ready communication channel s. for OAM. So do SONET
and SDH.
The solution for a LAN is comparatively straightforward. The network
management facilityrLAN protocol analyzer becomes just another active
station on the LAN. Network management traffic remains as any other
revenue-bearing traffic on the LAN. Of course, the network management
traffic should not overpower the LAN with message unit quantity which we
might call network management overhead.
WANs vary in their capacity to provide some form of communicating
network management information. X.25 provides certain types of frames or
messages dedicated for network control and management. However, these
framesrmessages are specific to X.25 and do not give data on, say, error rate
at a particular point in the network. Frame relay provides none with the
exception of flow control and the CLLM, which are specific to frame relay.
For a true network management system, a separate network management
communication channel may have to be provided. It would have to be
sandwiched into the physical layer. However, SNMP described below. was
developed to typically use the transport services of TCPrIP. See Section
11.4 for a discussion of TCPrIP and related protocols.. It is additional
overhead, and care must be taken of the percentage of such overhead traffic
compared to the percentage of revenue-bearing traffic. Some use the term
in-band when network management traffic is carried on separate frames on
the same medium, and the term out-of-band is used when a separate
channel or time slot is used such as with E1rT1.
18.7 AN INTRODUCTION TO NETWORK MANAGEMENT

PROTOCOLS
18.7.1
Two Network Management Protocols
Two separate communities have been developing network management protocols:
The TCPrIP community: SNMP

The ISOrOSI community: CMIP
740
The most mature and most implemented is the Simple Network Management
Protocol SNMP.. Certain weak points arose in the protocol, and a version
called SNMPv.2 has been developed and fielded.
The Common Management Information Protocol CMIP. has been developed for the OSI environment. It is more versatile but requires about five
times the memory of SNMP.
18.7.2
An Overview of SNMP
SNMP is the dominant method for devices on a network to relay network

management information to centralized management consoles that are designed to provide a comprehensive operational view of the network. Having
come on-line in about 1990, literally thousands of SNMP systems have been
deployed.
There are three components of the SNMP protocol:
The management protocol itself

The management information base MIB.
The structure of management information SMI.
Figure 18.4 shows the classic clientserver model. The client runs at the
managing system. It makes requests and is typically called the network
management system NMS. or network operations center NOC.. The server is
in the managed system. It executes requests and is called the agent.
Structure of Management Information (SMI). This defines the general framework within which an MIB can be defined. In other words, SMI is the set of
rules which define MIB objects, including generic types used to describe
management information. The SNMP SMI uses a subset of Abstract Syntax
Notation One ASN.1. Ref. 7. specification language that the ISO International Standards Organization. developed for communications above the OSI
presentation layer. Layer 7, for example, may use ASN.1 standards and
ITU-T Recs. X.400 and X.500. It was designed this way so that SNMP could
Figure 18.4. SNMP basic management architecture. SMI, structure of management information; NOC, network operations center; MIB, management information base.
741
be aligned with the OSI environment. The SMI organizes MIB objects into
an upside-down tree for naming purposes.
Management Information Base (MIB). This is the set of managed objects or
variables that can be managed. Each data element, such as a node table, is
modeled as an object and given a unique name and identifier for management purposes. The complete definition of a managed object includes its
naming, syntax, definitions, and access method such as read-only or readwrite. which can be used to protect sensitive data and status. By allowing a
status of required or optional, the SNMP formulating committee allows
for the possibility that some vendors may not wish to support optional
variables. Products are obligated to support required objects if they wish to
be compliant with the SNMP standard Ref. 8..
Figure 18.4 also shows an agent in every equipment to be managed, and
Figure 18.5 illustrates how a management console manages agents.
SNMP utilizes an architecture that depends heavily upon communication
between one or a small number of managers and a large number of remote
agents scattered throughout the network. Agents use the MIB to provide a
view of the local data that are available for manipulation by the network
management station. In order for a variable, such as the CPU utilization of a
remote Sun workstation, to be monitored by the network management
station, it must be represented as an MIB object.
Yemini Ref. 9. introduces the MIB in an interesting manner; and as many
other authors in the field, he portrays it as a tree, as shown in Figure 18.6. It
is located in the expanded agents box. The MIB is organized as a directory
tree of managed data located at the leaves, and it is shown as shaded cells.
The internal nodes of the tree are used to group managed variables by
related categories. Dr. Yemini shows that all information associated with a
given protocol entity we.g., Internet Protocol IP., TCP, User Datagram
Figure 18.5. The network management console manages agents.
742
Figure 18.6. SNMP pictorial overview. Note the tree formation. (From page 29, Ref. 9,
reprinted with permission of the IEEE Press.)
Protocol UDP.x is recorded in a given subtree. The tree structure provides a

unique path identifier index. for each leaf. This identifier is formed by
concatenating the numerical labels of its path from the root. The managed
information stored at the leaves can be accessed by passing the respective
identifier to the agent.
The management station sends get and set requests to remote agents.
They provide a query mechanism for platform programs managers. to
retrieve or modify MIB data. GET retrieves data, SET changes the data, and
GET-NEXT provides a means to traverse the MIB tree and retrieve data in
order.
When an unexpected event takes place, these agents initiate traps to the
management station. In such a configuration, most of the burden for retrieving and analyzing data rests on the management application. Unless data are
743
Figure 18.7. An SNMP messageembedded in a local network frame.
requested in a proactive way, little information will be displayed at the

management station. This poll-based approach increases network traffic,
especially on the backbone where some users report a 510% increase in
traffic due to SNMP packets or messages.. SNMP is a connectionless
protocol that initially was designed to run over a UDPrIP see Sections 11.5
and 11.4, respectively . stack. Because of its design, it is traditionally high in
overhead with the ratio of overhead to usable data running about 10 octets to
one. A typical SNMP message PDU. embedded in a local network frame
Inside the frame in Figure 18.7 we find an internet protocol IPsee
Section 11.4. datagram that has a header. The header has an IP destination
address that directs the datagram to the intended recipient. Following the IP
header, there is the UDP datagram that identifies the higher-layer protocol
process. This is the SNMP message, shown in Figure 18.7 as embedded in the
UDP. The application here is typically for LANs. If the IP fragment is too
long for one frame, it is fragmented segmented. into one or more additional
frames.
Figure 18.8 shows a typical SNMP PDU structure that is valid for all
messages but the trap format. This structure is embedded as the message
field in Figure 18.7. The SNMP message itself is divided into two sections: 1.
a version identifier plus community name and 2. a PDU. The version
identifier and community name are sometimes referred to as the SNMP
authentication header.
The ersion field assures that all parties in the management transaction
are using the same version of the SNMP protocol. We must remember that
the origins of SNMP evolved from TCPrIP, and they were originally designed for the exclusive use on TCPrIP networks. In Chapter 11, we became
familiar with the version field.
Each SNMP message contains a community name, which is one of the only
security mechanisms in SNMP. The agent examines the community name to
ensure that it matches one of the authorized community strings loaded in its
configuration fields or novolatile memory. Each SNMP PDU is one of five
744
Figure 18.8. An SNMP PDU structure for GetRequest, GetNextRequest, GetResponse, and
SetResponse. [From Managing Technology Overview, Chapt. 52, CISCO, June 1999 and
from RFC 1067 (J. Case, Univ. Tenn., Aug. 1988) and RFC 1157 (J. Case, Univ. Tenn., May
1990).]
Figure 18.9. SNMP V.1 trap PDU format. [From Internetworking Technology Overview,
Chapt. 52, CISCO, June 1999 and from RFC 1067 (J. Case, Univ. Tenn., Aug. 1988) and RFC
1157 (J. Case, Univ. Tenn., May 1990).]
types sometimes called erbs.: GetRequest, GetNextRequest, SetRequest,

GetResponse, and Trap. The trap PDU is shown in Figure 18.9.
The PDU shown in Figure 18.8 has five initial fields. The first field is the
PDU type. There are five types of PDU as we discussed previously. These
types are shown in Table 18.1.
The Request ID is the second field of the PDU field. It is an INTEGERtype field that correlates the managers request with the agents response.
INTEGER-type is a primitive type used in ASN.1.
TABLE 18.1
PDU-Type Field Values
GetRequest
GetNextRequest
GetResponse
SetRequest
Trap
TABLE 18.2
Error Type
noError
tooBig
noSuchName
badValue
readOnly
genErr
745
0
1
2
3
4
SNMP Error Codes

Value
Description
0
1
2
3
4
5
Success
Response too large to fit in single datagram
Requested object unknown r unavailable
Object cannot be set to specified value
Object cannot be set
Some other error occurred
The Error Status field is also an ASN.1 primitive type. It indicates normal
operation noError. or one of five error conditions as shown in Table 18.2.
When an error occurs, the Error Index field identifies the entry within the
variable bindings list that caused the error. If, for example, a readOnly error
occurred, the error index returned would be 4.
A Variable Binding pairs a variable name with its value. A VarBindList is a
list of such pairings. Note that within the Variable Bindings field of the
SNMP PDU, the word Object identifies the variable name OID encoding of
object type plus the instance . for which a value is being communicated.
A trap is an unsolicited packet sent from an agent to a manager after
sensing a prespecified condition such as a cold start, link down, authentication failure, or other such event. Agents always receive SNMP requests on
UDP port 161, and network management consoles always receive traps on
UDP port 162. This requirement means that multiple applications on the
same management station that wish to receive traps must usually pass control
of this port to an intermediate software process. This process receives traps
and routes them to the appropriate application. Figure 18.9 illustrates the
SNMP trap PDU structure. We must appreciate that the trap PDU structure
differs from the structure of the other four PDUs shown in Figure 18.8. Like
those other PDU structures, the first field, PDU type, will be in this case
PDU Type s 4. See Table 18.1..
The next field is called the Enterprise field and identifies the management
enterprise under whose registration authority the trap is defined. As an
example, the OID prefix 1.3.6.1.4.1234 would identify Newbridge Networks
Corporation as the address sending the trap. Further identification is provided in the Agent address field, which contains the IP address of the agent.
746
TABLE 18.3
SNMP Trap Codes
Trap Type
coldStart
warmStart
linkDown
linkUp
authenticationFailure
egpNeighborLoss
enterpriseSpecific
Value
0
1
2
3
4
5
6
In some circumstances a non-IP transport protocol is used. In this case the

value 0.0.0.0 is returned.
After the Agent Address field in Figure 18.9, we find the Generic Trap
Type that provides more specific information on the event being reported.
There are seven defined values enumerated INTEGER types. for this field
as shown in Table 18.3.
The next field is the Timestamp field, which contains the value of the
sysUpTime Object. This represents the amount of time elapsed between the
last re-initialization of the agent in question and the generation of the trap.
As illustrated in Figure 18.8, the last field contains the Variable Bindings.
A trap is generated by an agent to alert the manager that a predefined
event has occurred. To generate a trap, the agent assigned PDU Type s 4
and has to fill in Enterprise, Agent Address, Generic Trap, Specific Trap
Type, Timestamp fields, and Variable Bindings list.
18.7.3
SNMP Version 2
SNMPv.2 design used the field experience gained by SNMP to sharpen and
simplify the mappings to different transports. The management protocol has
been separated from the transport environment, encouraging its use over
practically any protocol stack.
One of the weaknesses of SNMP is in the unreliable manner of handling
trap messages. Of course, management communications are most critical at
times when the network is least reliable. A managers communications with
agents are vital. The use of UDP by SNMP means potentially unreliable
transport. SNMP leaves the function of recovery from loss to the manager
application. The GET-RESPONSE frame confirms respective GET, GETNEXT, and SET. A network manager can detect the loss of a request when a
response does not return. For instance, it can repeat the request. Traps are
another matter. Trap messages are generated by the agent and are not
confirmed. If a trap is lost, the agent applications would not be aware that
there is a problem, nor would the manager for that matter. Because
747
trap messages signal information which is often of great significance, securing

their reliable delivery is very important.
Accordingly, SNMPv.2 has pursued an improved mechanism to handle
event notifications, namely traps. For example, the trap primitive has been
eliminated. It is replaced by an unsolicited GET-RESPONSE frame which is
generated by an agent and directed to the trap manager UDP port 162.. Now
event notifications can be unified as responses to virtual requests by event
managers. In SNMPv.2, a special trap MIB has been added to unify event
handling, subscription by managers to receive events, and repetitions to
improve reliable delivery.
In the case of SNMP over TCPrIP, we are looking at what one might call
in-band communications. In other words, we are using the same communications channel for network management as we use for revenue-bearing
communications. This tends to defeat the purpose of network management.
The problem can be somewhat alleviated if the access path to agents is
somewhat protected from the entities that they manage.
Unlike SNMP, SNMPv.2 delivers an array of messaging options that
enables agents to communicate more efficiently with management stations.
Furthermore, SNMPv.2s bulk retrieval mechanism lets management stations
obtain reports from agents about a range of variables without issuing repeated requests i.e., GetBulkRequest.. This feature should cut down the
level of packet activity dedicated to network management and yet improve
network management efficiency. Another advantage of SNMPv.2 over its
predecessor is manager-to-manager communications, which allows a station
to act as either manager or agent. This allows SNMPv.2 systems to offer
hierarchical management using midlevel managers to offload tasks from the
central network management console.
SNMPv.2 does not have to rely on TCPrIP. It can run over a variety of
protocol stacks including OSI and Internet packet exchange IPX.. It also has
notably enhanced security features over its predecessor.
Figure 18.10 shows a generic SNMP-2 managed configuration.
18.7.4
Common Management Information Protocol (CMIP)
SNMP was developed rapidly with an objective to serve as a network

management communications standard until CMIP was issued and ready for
implementation. SNMP met such success that it now has a life of its own.
CMIP acceptance and implementation have shown slow progress. It has not
reached the popularity of SNMP.
CMIP is an ISO development Ref. 12., and it is designed to operate in
the OSI environment. It is considerably more complex than its SNMP
counterpart. Figure 18.11 illustrates the OSI management architecture, which
uses CMIP to access managed information. In Figure 18.11 this managed
information is provided by an agent in the LAN hub.
748
Figure 18.10. A generic SNMPv.2 managed configuration. (From Chapter 3, Ref. 11, reprinted
with permission.)
CMIP uses different terminology from SNMP. An agent maintains a

management information tree MIT. as a database; it models platforms and
devices using managed objects MOs.. These may represent LANs, ports, and
interfaces. CMIP is used by a platform to change, create, retrieve, or delete
MOs in the MIT. It can invoke actions or receive event notifications.
CMIP i.e., OSI management communications. communications are very
different from those found in SNMP. These communications are embedded
in the OSI application environment, and they rely on conventional OSI peer
layers for support. They use connection-oriented transport where SNMP uses
the datagram connectionless .. In most cases these communications are
acknowledged Refs. 13 and 14..
TELECOMMUNICATIONS MANAGEMENT NETWORK (TMN)
749
Figure 18.11. A typical overall architecture of an OSI network management system. (From
Figure 2-12, page 48, Ref. 9.)
18.8
18.8.1

Objective and Scope
TMN provides a framework for telecommunications management. By introducing the concept of generic network models for management, it is then
possible to perform general management of diverse equipment using generic
information models and standard interfaces.
TMN was initially defined by the ITU in 1988. It is designed to facilitate
supplier independence for the customer regarding telecommunication management systems. The objective of TMNs architecture is to allow interoperable interfaces between managing and managed systems and between network
management systems themselves. TMN is a set of standards whose goal is to
provide unambiguous definitions for the information in support of management functions and how they are exchanged across an interface between
communicating parties. TMN is still in a stage of evolvement Ref. 13..
Figure 18.12 shows the relationship between TMN and the telecommunication network that it manages.
TMN uses concepts from OSI Systems Management architecture and
applies them in the context of telecommunications network management.
750
751
With the OSI system management regime we find the familiar manageragent
role that is to some degree at odds with the OSI peer-to-peer basic relationship, thus we have peer-to-peer versus master-slave.
There are five independent functional blocks comprising the TMN:
1.
2.
3.
4.
5.
OSFoperations system function

MFmediation function
WSFworkstation function
NEFnetwork element function
QAFQ adapter function
The processes associated with the management of the telecommunication

network are represented by the operations system function OSF.. This
includes gathering alarm information, processing received management information, and directing the managed entities to take appropriate action.
The mediation function MF. facilitates the information exchange between
two other functional blocks, typically between the OSF and the NEF, the
network element function. This function may include the capabilities to store,
filter and adapt data to meet the requirements of the two sides being
connected.
The WSF functional block takes machine type data and makes it meaningful to a human user; it facilitates the man-machine interface. The QAF
bridges the non-TMN system to the TMN Refs. 15, 16, and 18..
18.8.2 Network Management Functions Carried Out in Upper
OSI Layers
Crucial network management activity is accomplished in the application
layer. This layer is further structured using application service elements
ASEs.. There are two types of activities that concern the application layer:
information processing and the communication of this information to a
remote system.
Our concern here is the application entity, which is composed of one or
more application service elements ASEs.. An ASE is a building block that
meets a specific functional need of the applications. There are four distinct
ASEs that make up an application entity. These are Association Control
Service Element ACSE., Remote Operations Service Element ROSE.,
Common Management Information Service Element CMISE., and Systems
Management Application Service Element SMASE.. Figure 18.13 shows
how these four ASEs tie together in the TMN environment.
Service definitions are used in the OSI Reference Model framework. In
the TMN context, the application service elements also contain service
definitions. Here the services offered by the requestreply application service
element are used by the CMISE, and in turn the services of the CMISE are
used by SMASE.
752
Figure 18.13. Network management application entity. (From Figure 3.4, page 67, Ref. 15.)
In Figure 18.13, the coordination functions are used in order to provide

for the flexibility of introducing different ASEs and allowing for negotiation
of the features on an association. It will be noted that there is a dotted line
around the ASEs used for data transfer. The management message is a result
of combining protocol definitions for these ASEs. This is briefly covered
below.
Association Control Service Element (ACSE). This is used by any connection-oriented application to set up and release an application. This application element is used to set up and release the association. The association is
set up by specifying a context for the association. There are four service
primitives beside authentication, used to exchange security information.
These four are:
A-Associate Request and Confirmation

A-Release Request and Confirmation
A-Abort Request and Indication
A-P-Abort Indication
753
The A-Abort is issued by the user of ACSE, whereas the Provider Abort
A-P-Abort. is issued by the protocol device because the event received is
invalid for the current state.
The ACSE only sets up and releases an association. The actual data
transfer is performed by the three ASEs described below.
Remote Operations Service Element (ROSE). ROSE is a simple protocol
that defines the structure for invoking operations remotely and responding to
the invokation. The services support correlation of responses to requests.
When a response is received, it can be correlated with the corresponding
request. ROSE is further augmented with CMISE and SMASE to achieve
the management data transfer.
Common Management Information Service Element (CMISE). CMISE supports ROSE in that it refines the structure offered by the requestrreply
framework of ROSE. Its specialization is in the context of operations
common to all management functions. Irrespective of the functional areas
and the resources managed, a basic set of operations is defined by CMISE.
CMISE offers services and features grouped together into functional units.
The requirements for both Q3 and X interfaces specified in ITU-T Rec.
Q.812 are specified in terms of profiles for network management.
Systems Management Application Service Element (SMASE). SMASE covers all management functions in a generic sense. This makes it different from
other service elements in its structure, and there is no singular document
providing protocol definition. SMASE represents a collection of one or more
System Management Functions SMFs. and, based on which functions are
used on an association, the corresponding services to be used. These SMFs
further refine the generic framework set by CMISE. Here, in this case,
CMISE is not complete until further refined by functions that define specific
event types. The generic event report service then is augmented by service
definitions and protocol specification. Thus each specific function has service
definitions and protocol specifications included.
TMN Management Message Structure Description. L. G. Raman, in
Ref. 15, describes the structure of a management message to appreciate the
basic differences between the TMN approach and existing message-based
approaches e.g., SNMP. in a high-level view.
In many current approaches, messages are specified in a readable format
using character strings. Formulating the message using the TMN interface is
different. The message itself is not completely laid out as a string of
characters; instead, using level abstractions, different parts of messages are
supplied by different application service elements.
Figure 18.14 shows how the flow of information occurs for this case. For
understanding the application level messages, assume that the lower layer
754
Figure 18.14. System management message components. (From Figure 3.5, page 70,
Ref. 15. Reprinted with permission.)
functions are available. The first step is establishing a virtual connection

between the two application entities using ACSE as shown in the figure.
Assume that a fabric in the network element has a failure and as a result
an alarm is released. The message is assembled by using the structure
defined in ROSE. The figure indicates that the filled fields are defined
within that protocol according to the shades. The filled boxes are shown to
illustrate the point that the message is formed by using the three service
elements mentioned above. The Remote Operations Protocol defines the
field for performing requestrreply correlation and identifying the operations.
The unfilled box indicates that ROSE leaves the field to be populated by its
user. CMISE defines fields that are common to all management functions. A
System Management function such as alarm reporting completes the other
fields. In addition, the appropriate values for these fields are determined by
the information model for the management information. This is shown by a
tree of circles on both ends. It illustrates that there exists a shared management knowledge between the two systems on the resources and their characteristics so that the exchanged information can be interpreted appropriately
Refs. 15 and 17..
Figure 18.15 defines the completed structure for the message by combining definitions from the application service elements and the model for
REFERENCES
755
Figure 18.15. Typical system management message. (From Figure 3.6, page 71, Ref. 15.
Reprinted with permission.)
management information. The message for management application is then

encoded by the presentation layer functions into the set of octets transferred
over the established transport and network level connections using standard
OSI structure.
REFERENCES
1. The IEEE Standard Dictionary of Electrical and Electronics Terms, 6th edition,
IEEE Std 100-1996, IEEE, New York, 1996.
2. Transport Systems Generic Requirements: Common Requirements, GR-499-CORE,
Issue 2, Bellcore, Piscataway, NJ, December 1998.
3. Steven M. Dauber, Finding Fault, BYTE Magazine, McGraw-Hill, New York,
March 1991.
4. Error Performance of an International Digital Connection Forming Part of an
Integrated Serices Digital Network, CCITT Rec. G.821, Fascicle III.5, IXth Plenary
5. Synchronous Frame Structures Used at Primary and Secondary Hierarchical Leels,
CCITT Rec. G.704, Fascicle III.4, IXth Plenary Assembly, Melbourne, 1988.
6. Frame Alignment and Cyclic Redundancy Check CRC . Procedures Relating to Basic
Frame Structures Defined in Recommendation G.704, CCITT Rec. G.706, CCITT
Geneva, 1991.
7. Information Processing Systems: Open Systems InterconnectionAbstract Syntax
Notation One ASN.1., ISO 8824, Geneva, 1987.
8. A Simple Network Management Protocol, RFC 1157, DDN Network Information
Center, STI International, Menlo Park, CA, 1990.
9. A Critical Survey of Network Management Protocol Standards, Yechiam
Yemini, Chapter 2, Telecommunications Network Management into the 21st Century, Edited by Salah Aidarous and Thomas Plevyak, IEEE Press, New York,
1994.
10. Mark A. Miller, Managing Internetworks with SNMP, M & T Books, New York,
1993.
11. William Stallings, Network Management, IEEE Computer Society Press, Los
Alamitos, CA, 1993.
756
12. Information TechnologyOpen Systems InterconnectionCommon Management

Information Protocol Specification, ISOrIEC 9596, 1st edition, ISO Geneva, 1990.
13. Information TechnologyOpen Systems InterconnectionSystems Management
Oeriew, ISOrIEC 10040-1998, ISO Geneva, 1998.
14. Information TechnologyOpen Systems InterconnectionCommon Management
Information Serice, ISOrIEC 9595, ISO Geneva, 1998.
15. Lakshmi G. Raman, Fundamentals of Telecommunications Network Management,
IEEE Press, New York, 1999.
16. Principles for Telecommunications Management Network, ITU-T Rec. M.3010, ITU
Geneva, May 1996.
17. TMN Interface Specification Methodology, ITU-T Rec. M.3020, ITU Geneva, 1995.
18. TMN Management Serices and Telecommunications Managed Areas: Oeriew,
ITU-T Rec. M.3200, ITU Geneva, April 1997.
APPENDIX
ADDRESSING CONVENTIONS
I.1 TRANSFORMING DECIMAL NUMBERS TO BINARY
NUMBERS AND TRANSFORMING BINARY NUMBERS
TO DECIMAL NUMBERS
When working with data transmission, more often than not we work with
octets or bytes. A byte in this book contains 8 bits. Unless we tell you
otherwise, we work from right to left. Then the rightmost digit is the zero
digit and will be the decimal number raised to the zero power. A decimal
number is often used as shorthand for a binary number..
Consider the decimal number 457,623. In these exercises we must take
into account the positional location of each digit starting with the 0
location. Thus we have a zero location, a one location, a two location, and so
on. Again, right-to-left, the first digit will be 10 to the zero power, the second
digit will be 10 to the one power, and the third digit will be 10 to the two
power i.e., squared., as follows, but write the number conventionally. To
assist in understanding what we are doing, the positional location is written
above each digit grouping in the line of digit calculations:
Positional
location:
5
4 = 10
4
5.
q 5 = 10
3
4.
q 7 = 10
2
3.
q 6 = 10
1
2.
q 2 = 10
0
1.
q 3 = 10 0 .
Lets do the addition the long way to see if we derive 457,623:

400,000
q50,000
q 7,000
q 600
q
20
q
3
457,623
757
758
APPENDIX I
Binary-to Decimal. Given a binary sequence of

01100111
convert to its binary equivalent. There are eight binary digits, which means
that my first decimal number will be 2 to the 7th power, the second will be 2
to the sixth power, and so on. Here we assume that transmission is from left
to rightthe 0 will be the first digit out of the pipe. Of course, in some
situations the first digit out of the pipe is the rightmost digit. In the
translation process, we skip each positional location where a binary zero
appears and only include those positional locations where a binary one
appears. Lets translate:
01100111 would be
0 = 2 7 . q 1 = 2 6 . q 1 = 2 5 . q 0 = 2 4 . q 0 = 2 3 . q 1 = 2 2 . q 1 = 2 1 . q1 = 2 0 .
Convert each digit to its decimal equivalent, as follows and sum:

0 q 64 q 32 q 0 q 0 q 4 q 2 q 1 s 103
In some protocols you are told that the least significant bit LSB. is transmitted first. Lets use the same binary sequence, namely 01100111. The bit order
now must be reversed, or
11100110
The binary 1 on the left end is going out the pipe first. It is the least
significant bit, and has the positional location of 0. By placing the positional
location above each binary digit, we get
0
1 = 2 q 1 = 2 q 1 = 2 q 0 = 2 q 0 = 2 q 1 = 2 q 1 = 2 q 0 = 2 7 .
0.
1.
2.
3.
4.
5.
32
6.
64
0 s 103
In most situations the least significant bit is transmitted first. For example,
according to ITU-T Rec. X.25, serial bit stream derived from a data frame
based on LAPB, addresses, commands, responses, and sequence numbers
shall be transmitted with the low-order bit first, for example, the first bit of
the sequence number that is transmitted shall have the weight of 2 0 .. Be
careful, in some texts and protocol descriptions, the rightmost bit goes out
the pipe first, rather than the left.
To further add to this discussion, the Sixth edition of the IEEE Dictionary
defines most significant bit MSB . as the bit having the greatest effect on the
value of a binary numeral, usually the leftmost bit.Logic values expressed
DECIMAL DIGIT REPRESENTATION IN IP ADDRESSES
759
in binary form are shown with their most significant bit on the left. In our
example above, namely 11100110, with the positional location as given, the
rightmost bit, the binary 0, is the most significant digit, not the leftmost.
We follow the IEEE convention in the remainder of this appendix with the
leftmost digit being the MSB.
Decimal Number to Binary Sequence. This is the reverse procedure as
above. In this case we are given a decimal number and we are to find the
binary sequence that matches that number.
For example, find the binary sequence that matches the decimal number
245. It would be helpful to remember 2 8 s 255, not 256, because our first
digit is 0.. We find the largest value of n so that 2 n F 245. It is 7. That is,
2 7 s 128. Thus, in the 7th bit position the 8th bit. there will be a 1. Subtract
128 from 245 and we have 117. Now again we find the largest value of n such
that 2 n F 117. The value of n will be 6, because 2 6 s 64. Thus, the 6th
position will be a 1. Subtracting 64 from 117 leaves 53. Still again we find the
largest value of 2 n such that 2 n F 53. Since 2 5 s 32, in the 5th bit position
we have a binary 1. Then, 53 y 32 s 21. Once again find the largest value of
2 n such that 2 n F 21. Since 2 4 s 16, in the 4th bit position, we have a 1,
then, 21 y 16 s 5. Still again find the value of 2 n such that 2 n F 5. Since
2 2 s 4, then, 5 y 4 s 1. 2 0 s 1.
By inspection we see that this is an 8-bit binary number. We know this
because 245 - 255 but is more than 127. If it were less than 127, it would be
a 7-bit binary number. Using a guide for positional notation to define our
equivalent binary number:
7
1
6
1
5
1
4
1
3
0
2
1
1
0
0
1
Wherever there is a valid power of two, insert a 1; when there is a blank in

the series 76543210, insert a 0. We put in binary 0s because there was no 2 3
value nor a 2 1 value. For all the rest, we inserted binary 1s.
I.2
DECIMAL DIGIT REPRESENTATION IN IP ADDRESSES
IP4 addresses have 32 bits or 4 octets. We might see an IP address something

like the following:
241.14.6.88
These decimal digits represent the first, second, third, and fourth octet of an
IP address. How do we convert these digits to the more familiar binary
sequences?
760
APPENDIX I
Each digit group is taken by itself and treated as any other decimal
number to be converted to binary. In other words, what 8-bit binary sequence
derives from decimal 241? We use the same methodology as in Section I.1.
What if 2 n is F 241? Then 2 7 s 128. This tells us that the first binary digit
of the 8-bit sequence is a 1. Subtract 241 y 128 s 113. If 2 n F 113, then
2 6 s 64. This tells us that the second place from the left is a binary 1.
Subtract again: 113 y 64 s 49. If 2 n F 49. Then, 2 5 s 32. This tells us
that in the third place from the left there is a binary 1.
Subtract still again. 49 y 32 s 17. If, 2 n F 17. Then, 2 4 s 16. This tells us
that the fourth place from the right contains a binary 1.
Where are we so far?: 1111xxxx.
Subtract still again. 17 y 16 s 1. If, 2 n F 1, then, 2 0 s 1.
This tells us that the 0 position from the right or the 8th position from
the left is a binary 1. Because we have no valid 2 n values for positions three,
two, and one, we then insert binary 0s in those positions. Thus we now have
11110001
Lets check it to see if we are right.
7 6 5 4 3 2 1 0 position number from the right
1 1 1 1 0 0 0 1
Add
2 7 s 128
2 6 s 64
2 5 s 32
2 4 s 16
2 3 s no value present
20 s 1
241 our original decimal value.
The largest decimal value we can have for an 8-bit binary sequence, of
course, is 11111111 or 255 decimal. The smallest value is 00000000 or decimal
0.
This addressing technique, so commonly found in IP, is called dotted
decimal notation system.
BINARY, OCTAL, AND HEXADECIMAL NUMBERS
TABLE I.1
761
Decimal, Binary Octal, and Hexadecimal Numbers
Decimal
Binary
Octal
Hexadecimal
4-Bit String
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0
1
10
11
100
101
110
111
1000
1001
1010
1011
1100
1101
1110
1111
0
1
2
3
4
5
6
7
10
11
12
13
14
15
16
17
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
Convert the second address octet 14 to binary: If, 2 n F 14 then, 2 3 s 8.

Subtract 14 y 8 s 6. If, 2 n F 6 then, 2 2 s 4. Subtract 6 y 4 s 2. Then
2 1 s 2. In the 8-digit binary number there will be 1s in the 3rd, 2nd, and 1st
places from the right.. Insert 0s in the remainder of the positions as follows:
Positional location:
76543210
00001110
I.3
BINARY, OCTAL, AND HEXADECIMAL NUMBERS
Most computer software that we deal with uses either the octal number
system, which has a base of 8, or the hexadecimal number system, which uses a
number base of 16. Table I.1 shows the decimal numbers 0 through 1 and
their binary, octal, and hexadecimal equivalents.
To convert a binary sequence to an octal sequence, starting from the left,
separate the binary digits into groups of three; to convert to a hexadecimal
sequence, starting from the left, separate the binary sequence into groups of
four digits. For example:
101011000110 s 101 011 000 110 2 s 5306 8
s 1010 1100 0110 2 s AC6 16
II
ACRONYMS AND ABBREVIATIONS
2B1Q
4B5B
8B6T
8B10T
A baseband waveform, 2 binary 1 quaternary

Baseband waveform coding 4 bits to 5 bits.
Baseband waveform coding 8 binary to 6 ternary.
Baseband waveform coding 8 binary to 10 ternary.
A
AAL
ABM
AC
ACEG
ACF
ACK
ACSE
ADM
ADCCP
ADSL
AFI
AIS
ALBO
AM
AMI
AM-VSB
AMP
ANSI
AP
APD
AR
ARPA
ARPANET
ARM
ARP
ATM adaptation layer

Asynchronous balanced mode
Access control
AC equipment ground
Access control field
Acknowledgement.
Association control service element
Adddrop multiplexer; asynchronous disconnect mode
Advanced data communications control procedures
Asymmetric digital subscriber line
Authority and format identifier
Alarm indication signal
Automatic line build out
Amplitude modulation
Alternate mark inversion
Amplitude modulationvestigial sideband
Auxiliary multiplexing pattern
American National Standards Institute
Access point
Avalanche photodiode
Access rate
Advanced Research Projects Agency
ARPA network
Asynchronous response mode
Address resolution protocol
763
764
APPENDIX II
ARQ
ASCII
ASE
ASK
ASN.1
ASTM
AT & T
AU
AUG
AUI
AutoVoN
AUU
AWG
B
B3ZS, B6ZS, B8ZS
BASE
BAsize
BBE, BBER
BCC
BCD
BECN
Bellcore
BEOFD
BER
BERT
BEtag
BICI
BIP
B-ISDN
BISUP
BITE
BIU
BLF
B-NT1, B-NT2
BNZS
BOM
BPDU
bsp
BPSK
BR
BRI
BROAD
Automatic repeat request

American Standard Code for Information Interchange
Application service elements.
Amplitude shift keying
Abstract syntax notation1
American Society of Testing and Materials
American Telephone and Telegraph Company.
Access unit
Administrative unit group
Attachment unit interface
Automatic voice network DoD.
ATM user-to-ATM user
American wire gauge
Binary 3, 6, 8 zeros substitution

Baseband
Buffer allocation size
Background block error, ratio
Block check count
Binary coded decimal
Backward explicit congestion notification
Bell Communications Research formerly BTL., now
Telcordia
Broadband end-of-frame delimiter
Bit error rate
Bit error rate test
Beginning-end tag
Broadbandintercarrier interface
Bit interleaved parity
Broadband-ISDN
Broadband ISDN User Part SS7.
Built-in test equipment
Baseband interface unit
Basic low layer functions
Broadband reference interface . NT1 and NT2 NT s
network termination.
Binary N-zeros substitution
Beginning of message
Bridge protocol data unit
Bits per second
Binary phase shift keying
Bit rate
Basic rate interface
Broadband, indicates broadband
APPENDIX II
BSA
BSS
BSSID
BT
C
CAC
CADrCAM
CAP
CAT
CATV
CBR
CBDS
CC
CCITT
CD
CDMA
CDV
CE
CEQ
CF
CFP
CI
CIB
CIR
CLEC
CLLM
CLNAP
CLNP
CLP
CLSF
CM
CMCI
CMI
CMIC
CMIP
CMISE
CMTS
COM
CONUS
CP
CPI
CPCS
765
Basic service area

Basic service set
Basic service set identification
British Telecom
Connection admission control

Computer-aided designrcomputer-aided
manufacturing
Carrierless amplitude and phase modulation.
Category
Community antenna television
Constant bit rate
Connectionless broadband data service
Country code
International Consultive Committee on Telephone and
Telegraph Now ITU-T.
Compact disk
Code division multiple access
Cell delay variation
Connection element; collision enforcement
Customer equipment
Coordination function
Contention free period
Control-in
CRC indicator bit
Committed information rate
Competitive local exchange carrier
Consolidated link-layer management
Connectionless network access protocol
Connectionless network protocol
Cell loss priority
Connectionless service functions
Cable modem
Cable modem customer interface
Coded mark inversion
Concentrator medium interface connector
Common management information protocol
Common management information service element
Cable modem termination system
Continuation of message
Contiguous United States
Contention period
Common part indicator
Common part convergence sublayer
766
APPENDIX II
CPE
CPI
CPU
CrR
CRC
CRF
CRS
CS
CSI
CSMA
CSMArCA
CSMArCD
CSU
CTS
CW
D
DA
DACS
DAMA
DARPA
dB
dB, dBm, dBW,
dBi, dBc, dBmV
DBPSK
DC
DCC
DCE
DCF
DDS
DE
DECnet
DIFS
DISC
DLCI
DLL
DM
DMPDU
DNHR
DNIC
DoD
DPTE
DQDB
DQPSK
Customer premise equipment

Common part indicator
Central processing unit
Commandrresponse
Cyclic redundancy check
Connection-related functions
Carrier sense; configuration report server
Convergence sublayer
Convergence sublayer indicator
Carrier sense multiple access
Carrier sense multiple accessrcollision avoidance
Carrier sense multiple access with collision detection
Channel service unit
Clear to send
Carrier wave sometimes, affectionately, refers to
Morse operation.
Destination address
Digital automatic cross-connect system
Demand assignment multiple access
Defense Advanced Research Projects Agency
Decibel
Decibel-related terms
Differential binary PSK
Direct current
Data country code
Data communication equipment
Distributed coordination function
Digital data system
Discard eligibility
Digital Equipment Corporation network
DCF interframe space
Disconnect
Data-link connection identifier
Data-link layer
Disconnect mode
Derived MAC protocol data unit
Direct nonhierarchical routing
Data network identification code
Department of Defense, U.S.
Data processing terminal equipment s DTE.
Distributed queue dual bus
Differential quadrature phase shift keying
APPENDIX II
DS
DS0, DS1
DSAP
DSE
DSL
DSLAM
DSM
DSP
DSS
DSS1, DSS2
DSSS
DSU
DTE
DWDM
E
E0, E1
EA
EB
EBCDIC
ED
EDAC
EDC
EDD
EDFA
EFS
EIA
EIFS
EF
EIRP
EMC
EMI
eoc
EPO
ER
ES
ESA
ESF
ESI
ESR
ESS
ETD
ETSI
EUTP
767
Direct sequence spread spectrum.

Referring to multiplex levels of U.S. digital network
Destination service access point
Data switching equipment
Digital subscriber line
Digital subscriber line access multiplexer
Distribution system medium
Domain-specific part
Distribution system services.
Digital signaling system No. 1, No. 2
Direct sequence spread spectrum
Data service unit
Data terminal equipment
Dense wave-division multiplex
Multiplex levels of European digital network

Address extension bit
Errored block
Extended Binary Coded Decimal Interchange Code
End delimiter
Error detection and correction
Error detection code
Envelope delay distortion
Erbium-doped fiber amplifier
End-of-frame sequence
Electronic Industries Alliance
Extended interframe space
Entrance facility
Equivalent isotropic radiated power
Electomagnetic compatibility
Electromagnetic interference
Embedded operations channel
Errored performance objective
Equipment room
Errored second
Extended service area
Extended superframe
End system identifier
Errored second ratio
Extended service set
End of transmission delimiter
European Telecommunications Standardization
Institute
Enhanced unshielded twisted pair
768
APPENDIX II
F
FC
FCS
FDDI
FDM
FDMA
FEBE, febe
FEC
FECN
FER
FERF
FH
FHSS
FIFO
FM
FMIF
FR
FRAD
FRF
FRMR
FRS
FS
FSK
FSL
FTP
Frame control
Frame check sequence
Fiber distributed data interface
Frequency division multiplex
Frequency division multiple access
Far-end block error
Forward error correction
Forward explicit congestion notification
Frame error ratio
Far-end receive failure
Frequency hop
Frequency hop spread spectrum
First in, first out
Frequency modulation
Frame mode information field
Frame relay
Frame relay access device
Frame Relay Forum
Frame reject response
Frame relay switch
Frame status
Frequency shift keying
Free-space loss
File transfer protocol
G
Gbps
GBSVC
GCRA
GFC
GFSK
GHz
GMII
GTE
Gigabits per second

Generic broadcast signaling virtual channel
Generic cell rate algorithms.
Generic flow control
Gaussian frequency shift keying
Gigahertz
Gigabit medium-independent interface
General Telephone & Electronics
H
HC
HCS
HDB3
HDLC
HDR
HDSL
HE
HEC
Horizontal cross-connects
Header check sequence
High density binary 3
High-level data-link control
Header
High-speed digital subscriber line
Head-end
Header error control
APPENDIX II
HFC
HIC
HLPI
HOC
HPA
HRC
HRP
HRX
HTC
Hz
I
IA5
IBSS
ICD
ICMP
IDI
IDL, idl
IDU
IEC
IEEE
IFG
IrG
IHL
IISP
ILMI
IM
IMPDU
INFO
IrO
IP
IPX
IrQ
IR
IRC
IRL
IRP
ISDN
ISM
ISP
ISO
ISUP
Hybrid fiber coax

Highest incoming channel
Higher-layer protocol identifier
Highest outgoing channel
High-power amplifier
Harmonic-related carrier; hybrid ring control;
horizontal redundancy check
Hypothetical reference path
Hypothetical digital connection
Highest two-way channel
Hertz
International alphabet no. 5

Independent basic service set
Interface control document; international code
designator
Internet control message protocol
International domain identifier
Idle
Interface data unit
International Electrotechnical Commission
Institute of Electrical and Electronics Engineers
Interframe gap
Individualrgroup
Internet header length
Interim interswitch signaling protocol
Interim local management interface
Intermodulation
Initial MAC protocol data unit
Information field.
inputroutput device.
Internet protocol
Internet packet exchange
In-phase quadrature
Infrared
Incremental related carrier
Isotropic receive level
Internal reference point
Integrated services digital networks.
In-service measurements; industrial, scientific and
medical
Internet service provider
International Standards Organization
ISDN user part SS7.
769
770
APPENDIX II
ITA
ITU
IXC
K
kbps
kHz
km
L
LAN
LAPB
LAPD
LAPF
LCN
LCT
LD
LEC
LED
LEO
LI
LLC
LLCDU
LMDS
LNA
LO
LOC
LOS
LRC
LSAP
LSB
LSDU
LSDV
LSI
LTE
LX-PMD
M
MAC
MAN
MAP
M-ary
International telegraph alphabet

International Telecommunication Union
Interexchange carrier typically ATT, Worldcom, BT,
etc..
Kilobits per second

Kilohertz
Kilometer
Local area network

Link access protocolB-channel
Link access protocolD-channel
Link access protocolframe relay
Logical channel number
Last compliance time
Laser diode
Local exchange carrier
Light-emitting diode
Low earth-orbit satellite .
Length indicator
Logical link control
LLC data unit
Local multipoint distribution system
Low-noise amplifier
Local oscillator
Lowest outgoing channel; loss of cell delineation .
Line-of-sight
Longitudinal redundancy check
Link service access point
Least significant bit
Link-layer service data unit
Link segment delay value
Large-scale integration
Line terminating equipment; link terminating
equipment
Long-wavelength PMD
Medium access control

Metropolitan area network
Multiaccess protocol
Gives modulation level of waveform e.g., 16-QAM,
where m s 16.
APPENDIX II
771
MAU
Mbps
MCNS
MDI
MDL
MF
MIB
MIC
MID
MIT
MLC
MLP
MMPDU
MO
MODEM
MNS.
MPDU
MPEG
ms
MSAP
MSB
MSDU
MSVC
MTP
mV
Medium attachment unit

Megabits per second
Multimedia cable network system
Medium-dependent interface
Management data link
Mediation function
Management information base
Medium interface connector
Message identification
Management information tree
Multilink control field.
Multilink procedure
MAC management PDU
Managed object
Modulator-demodulator
Multilink NS.
MAC protocol data unit
Motion picture experts group
Milliseconds.
MAC service access point
Most significant bit
MAC service data unit
Metasignaling virtual channel
Message transfer part
Millivolt
N
NA
NACK
NANP
NAV
NDC
NDM
NE
NEF
NEXT
NF
NHDR
nm
NMS
NMT
NNI
NOC
NPC
NPI
Numerical aperture
Negative acknowledgment
North American Numbering Plan
Net allocation vector
National destination code
Normal disconnect mode
Network element
Network element function
Near-end crosstalk
Noise figure
Network header
Nanometers.
Network management system
Network management
Networknetwork interface or networknode interface
Network operations center
Network parameter control
Numbering plan identification
772
APPENDIX II
N R.
NRM
NRZ
NRZI
N S.
ns
NSAP
NSDU
NSN
NSP
NT e.g., NT1, NT2.
NTSC
Receive sequence number

Normal response mode
Non-return to zero
NRZ inverted ones
Send sequence number
Nanoseconds.
Network service access point
Network service data unit
National significant number
Network service provider
Network termination
National Television System Committee
O
OAM, OA & M
OC
OSF
OSI
OSIE
OSP
OSPF
Operation, administration, and maintenance

Optical carrier used with SONET.
Operations system function
Open systems interconnection
OSI environment
Outside plant
Open shortest path first
P
PABX
PAD
PAM
PAR
PC
PCF
PCI
PCM
PCR
PCS
PDH
PDU
PDV
PrF bit.
PH
PHY
PIFS
PIN
PLCP
PLL
Private automatic branch exchange

Meaning pad or padding, extra false bits added; packet
assembler-disassembler
Pulse amplitude modulation
Positive acknowledgment with retransmission
Personal computer; point coordinator; priority control
Point coordination function
Protocol control information
Pulse code modulation
Peak cell rate
Physical coding sublayer; personal communication
services
Plesiochronous digital network
Protocol data unit
Path delay value
Pollrfinal
Packet handling handler .
Physical layer.
PCF interframe space
Positive-intrinsic-negative
Physical layer convergence protocol
Phase lock loop
APPENDIX II
773
PLOAM
PLS
PM
PMA
PMC
PMD
P-NNI
POH
POI
POTS
PPM, ppm
P R.
PRBS
PRI
P S.
PSC
PSDN
PSF
PSH
PSK
PSPDN
PSTN
PTI
PTR
PTT
PVC
Physical layer OA & M

Physical layer signaling
Phase modulation; physical medium
Physical medium attachment
Physical medium components
Physical medium dependent
Private networkNNI
Path overhead
Path overhead indicator
Plain old telephone service
Parts per million
Receive sequence number
Pseudo-random binary sequence
Primary rate interface
Send sequence number
Physical signaling components
Public switched data network
Packet switching facilities
Push function
Phase shift keying
Packet-switched public data network
Public switched telecommunications network
Payload type identifier
Pointer
Post-Telegraph-Telephone
Permanent virtual circuit; polyvinylchloirde
Q
QAF
QAM
QoS
QPSK
Q adapter function
Quadrature amplitude modulation
Quality of service
Quadrature phase shift keying
R
RARP
REJ
REM
RF
RFC
RFI
RI
RIM
RIP
RM
rms
Reverse address resolution protocol

Reject
Ring error monitor
Radio frequency
Request for comment
Radio frequency interference; request for information
Routing information
Request initialization
Routing information protocol
Resource management
Root mean square
774
APPENDIX II
RNR
ROSE
RPS
RR
RSL
RSU
RTS
RTU
RX
RX BITS
RX CLK
RXD
RZ
Receive not ready

Remote operations service element
Ring parameter server
Receive ready
Receive signal level
Remote switching unit
Request to send
Remote terminal unit
Receive
Receive bits
Receive clock
Receive data
Return-to-zero
S
SA
SABME
SAP
SAPI
SAR
SARM, SARME
SBS
SCR
SCSI
SCTE
SD
SDLC
SDH
SDP
SDU
SEAL
SES
SESR
SFD
SFS
SIFS
SIM
SIR
SLA
SLP
SMASE
SMDS
SMF
SMI
SMT
Source address
Set asynchronous balanced mode extended .
Service access point
Service access point identifier
Segmentation and reassembly
Set asynchronous response mode, extended
Selective broadcast signaling
Sustained cell rate
Small computer system interface
Society of Cable Telecommunication Engineers
Starting delimiter
Synchronous data-link control
Synchronous digital hierarchy
Severely disturbed period
Service data unit
Simple and. efficient AAL layer
Severely errored seconds.
Severely errored second ratio
Start frame delimiter
Start frame sequence
Short interframe space
Set initialization mode
Sustained information rate
Service level agreement
Single-link procedure
Systems management application service element
Switched multimegabit data service
Submultiframe; system management information
Structure of management information
Station management
APPENDIX II
SMTP
SN
SrN
SNAP
SNMP
SNP
SNRM
SOH
SOM
SONET
SPE
SREJ
SRL
SRTS
SS7, SSN7
SSAP
SSD
SSM
ST
STA
STE
STL
STM
STP
STS
SVC
SWP
SX-PMD
SYNC, sync
T
TA
TBB
TBBIC
TBI
TC
TCM
TCP
TCU
TDM
TDMA
TDR
TE
Tel, te2
775
Simple mail transfer protocol

Subscriber number
Signal-to-noise ratio
Subnetwork access protocol
Simple network management protocol
Sequence number protection
Set normal response mode
Section overhead; start-of-heading
Start of message
Synchronous optical network
Synchronous payload envelope
Selective reject
Structural return loss
Synchronous residual timestamp
Signaling system No. 7
Source service access point
Start of stream delimiter
Single segment message
Segment type
Station management
Station terminating equipment
Standard Telephone Laboratories
Synchronous transport module
Shielded twisted pair; signal transfer point
Space time space
Switched virtual circuit; signaling virtual channel
Switching pattern
Short-wavelength PMD
Synchronization
Terminal adapter; transmission address

Telecommunication bonding backbone
Telecommunication bonding backbone interconnecting
conductor
Ten-bit interface
Transmission convergence; telecommunications closet
Trellis-coded modulation
Transmission control protocol
Trunk coupling unit
Time division multiplex
Time division multiple access
Time-domain reflectometer
Terminal equipment; terminal endpoint
Terminal equipment 1, 2
776
APPENDIX II
TEF
TEI
Telex
TFTP
TGB
THDR
THT
THz
TIA
TMGB
TMN
TOA
TOS
TPDU
TS
TSI
TSN
TST
TTL
TU
TUG
TWT
TX
TX bits
TX clk
U
UA
Telecommunications entrance facility

Terminal endpoint identifier
A trade name telegraph exchange.
Trivial File Transfer Protocol
Telecommunications grounding busrbar
Transport header
Token holder timer
Terahertz
Telecommunications Industry Association
Telecommunications main grounding bar
Telecommunication management network
Type of address
Type of service
Transport protocol data unit
Time slot
Time slot interchange
Transmit sequence number
Time-space-time
Time to live
Tributary unit
Tributary unit group
Traveling wave tube
Transmit
Transmit bits
Transmit clock
UCD
UDP
UHF
UI
ULP
UMD
UNI
UPC
UPS
s
USAF
UTP
UU
Unnumbered acknowledgement. command.; user

agent
Upstream channel descriptor
User datagram protocol
Ultrahigh frequency
Unit interval; unnumbered information
Upper-layer protocol
Unscrambled mode delimiter
Usernetwork interface
Usage parameter control
Uninterruptible power supply; United Parcel Service
Microsecond
United States Air Force
Unshielded twisted pair
User-to-user
V
VBR
VC
Variable bit rate

Virtual container; virtual connection; virtual channel
APPENDIX II
VCC
VCI
VHF
VHSIC
VLAN
VLSI
VP
VP-AIS
VPC
VPI
VPN
V R.
VRC
V S.
VSAT
VT
VTAM
Virtual channel connection

Virtual channel identifier
Very high frequency
Very high-speed integrated circuit
Virtual local area network
Very large-scale integration
Virtual path
Virtual pathalarm indication signal
Virtual path connection
Virtual path identifier
Virtual private network
Receive state variable
Vertical redundancy check
Send state variable
Very small aperture terminal
Virtual tributary
Virtual telecommunications access method
W
WAN
WEP
WLAN
WDM
WSF
Wide area network

Wired equivalent privacy
Wireless LAN
Wavelength division multiplex
Workstation function
X
XDSL
XID
Generic DSL
Exchange identification command.
777
INDEX
A bold page number indicates extensive coverage of that topic.

An italic page number indicates where a definition of that topic may be found.
A-bit address bit., 382
A bits, 738
A-law logarithmic curve, 130
AAL ATM adaption layer.:
-0, 620
-1, 621
-2, 622
convergence sublayer CS., 623
-3r4, 623624, 633
-5, 610, 628
-PCI bit errors, 622
types, 620
ABM asynchronous balanced mode., 43
abort, 52
sequence, 295296, 511
TCP, 404
above the network layer, DOCSIS, 700
absolute zero, 101
absorption and scattering, 561
ac fault currents, 545
acceptance cone, figure, 557
access:
through B channel, basic rules, 524
connection element model, ISDN, figure,
492
control AC., 296297
field of a frame, 298
through D Channel, basic rules, 524
interference and recovery, 229230
to the Internet, 690
ISDN virtual circuit service, case B, 524525
network, 685
pathway, 59
PSPDN services, case A, 523
PSTN, subscriber network, 122123
rate AR., 455456

accounting management, 729
accumulated CDV, 656
ACEG bus, 544
ACFrheader, 602
ACK:
control bit, 409
and NACK, 64
returned, 407
ACK frame, 328
figure, 323
format, 322323
acknowledged operation, 508, 509
acknowledgement:
number, 409
of receipt of frames, packet, 8
acoustic coupler, 364
acquisition time, 239
ersus bit error probability, figure, 616
ACSE association control service element.,
TMN, 751
active:
bit, 463, 467
monitor, 305
open TCP., 404, 405
state, 52
topology, 344, 349
actual arrival time, 658
aCWmax, 331
aCWmin, 330
ad hoc test link, 736
ADCCP advanced data communications
control procedures ., 41, 44
adding intelligence to a bridge, 336337
address block format, figure, 384
779
780
INDEX
address designation, 319

address fields., 44, 225, 384, 438439
bit EA., 441, 511
extension, 438
format, 511
figure, 226
frame relay, figure, 439
ISDN, 510
WLAN, 319320
address part CLNP., 419
address-recognized A. bits, 301
address resolution protocol ARP., 388
address signaling, repeated, 70
address space, partitioned, 320
addresses determine routing, 390
addressing, 59, 343344
capacity, 371
conventions, 757762
and geographical significance, 356357
in X.25, 382383
administration of access resources, 515
administrative unit AU., 585
group AUG., 587
-n AU-n., 587
phase alignment, 587
pointers, 588589
in STM-1 frame, figure, 590
in STM-N, 589590
advanced data communications control
procedures ADCCP., 41
agent, 740, 741
address field, 745
aging time, 337, 343
agreed-upon QoS, 652
air interface, WLAN, 332335
ALBO automatic line build out., 136
algorithm, for SWP, 186
alignment AL.:
field, 626
ATM cell, 613
all paths explorer, 353
frames, 354
allocation of SAPI values, 512
allowable implementations, parameterized
values, 238239
ALOHA, 428
alpha times consecutively, 615
alternate mark inversion AMI., 134
American National Standards Institute
ANSI., 77
American Standard Code for Information
Interchange ASCII., 7778
AMI alternate mark inversion., 134, 135, 684
amplitude:
distortion, 160
frequency response, 97, 98
equalizer, 163
modulation AM., 96
-phase relationships with V.29 modem.
table, 169
and phase distortion, 159165
response, 596
shift keying ASK., 166
analog modem, 188, 198
ANSI ADSL, 683
ANSI American National Standards
Institute ., 77
CRC16, 63
-defined field formats, 441
frame relay, 436437
relay specifications, 474
T1.602 protocols, 442
antenna height ersus range calculation,
592593
anycast IPV6., 413
APD avalanche photodiode., 566
application entity, TMN, 751
application layer, 1920
apportionment rules, table, 106
APrSTA mapping WLAN., 316
architectural model of ISDN, 491
architectural positioning:
of 100BASE-T, 258
of gigabit Ethernet, 275
ARM asynchronous response mode., 43
ARP address resolution protocol., 386, 398
cache, 398
ARPANET, 386, 396
ARQ automatic repeat request., 47
-delays, 603
-function, 427
-type error correction, 68
artificial frame added, 148
ASCII American Standard Code for
Information Interchange ., 61, 77
character A, 89
ASE application service element., 751
ASK amplitude shift keying., 166
modulation, 96
aSlotTime, 329, 330
ASN.1 Abstract Syntax Notation.1., 740
primitive type, 745
assignment of logical channels, X.25, 385
association control service element ACSE.,
752753
asymmetric carrier frequencies, V.34, 180
INDEX
asymmetrical operation, 682

asynchronous:
advantage, 91
as ATM, 604
balanced mode ABM., 43, 220
cell time slots, 619
character-mode interfacing, 200
nonperiodic, 604
response mode ARM., 43
transfer mode ATM., 69, 601678
transmission, FDDI, 315
versus nonsynchronous, 89
ATM asynchronous transfer mode., 79, 69
adaptation layer AAL., 620632
addressing, 644645
plans, comparison, Figure, 644
-based, 59
cells., 70, 602
data indication, 654
flow, 616
header, 610, 611, 612, 618
information fields, 620
key to operation, 606613
mapping, 660, 664
mapping field, 662
transfer outcome, 647648
connection-oriented, 603
forum, 605, 619, 643, 657
header structure, 613
introduction, 602605
-LAN bridge, 645
layer, 616, 619
functions supported at UNI, table, 618
mechanism, F4 and F5 flows, 669670
OAM cell format, 677
primitives, 631
QoS, 654
SAP, 652
layering and B-ISDN, 615632
links, carry voice, data, image, figure, 603
network, 603
operation, administration , maintenance
OAM., 667678
routing and switching, 637638
SAP, 652
service quality review, 647
signaling protocol, 643
switch, figure, 605
rcross-connect, 633
switching, 633
UNI, 605
-user-to-user indication AUU. parameter,
631
781
attachment unit interface AUI., 235

attenuation distortion, 160, 162
attenuation:
of connecting hardware, for 100-ohm UTP
cable, table, 537
-frequency characteristic, 136
response of typical coaxial cable, figure,
689
of optical fiber, 561
of UTP cable, 534
variation with frequency, figure, 98
ATU ADSL transceiver unit., 686
ATU-R, 686
AU access unit., ISDN, 521, 522, 523
AU-4:
and AU-3, 591
pointer, 665
AUG administrative unit group., 585
AU-n:
frame, 588
pointers., 589, 666
rTU-3 pointer H1, H2, H3. coding, table,
589
authentication header, SNMP, 743
automatic equalization, 165
automatic equalizers, 93
automatic repeat request, 6465
automode procedures, 178
autonegotiation process, 278
autoVON, 156
auxiliary channel:
multiplexing parameters, table, 186
multiplexing pattern, AMP, 185
V.34, 180
availability available time., 105
availability enhancement, 730734
avalanche photodiodes, 566, 567
awaiting process TCP., 404
backbone:
cabling, 530532
hierarchical star topology, figure, 531
optical fiber cable EIArTIA-598., 540
routers, 609
UTP cabling, 535538
consists of, 535
background block error BBE., 105
backoff, 223, 324
time, 328
window, 332
back-up circuits, 730
backward explicit congestion notification
BECN. bit, 442
782
INDEX
balance:
bit to limit build up of DC component, 498
to ground, 122
balanced:
configuration, 42
digital interface circuit EIA-422, figure, 115
mode attenuation, 538
balancing network, 158
bandwidth, 605
allocated, 603
allocation, 595
controlled by CMTS, 701
data, how many bps fit?, 116
defined, 116
-distance factor, 558
500 MHz, 591, 596
parameter limits, table, 162
question, 116118
and redundant signals, 67
ersus bit rate, 116, 436
of voice channel, 95
1BASE5, 240
implementations, table, 238
10BASE5, 239
10BASE-FL, 10BASE-FP, 252
100BASE-FX, 252
100BASE-LX, 240
100BASE-T, 254
100 Mbps at CSMArCD MAC, 261
architectural positioning, figure, 258
full-duplex mode, 276
link segment, 254
MII, repeater and optimal autonegotiation,
264
100BASE-T4, 240
PCS and PMA, figure, 262
signaling, 263
100BASE-TX, 255
PCS, 265
100BASE-X:
block diagram, 266
PCS, 265
and PMA sublayers, 277
1000BASE-X:
block diagram, 266
PCS, PMA sublayers, 277
PHY, 269, 278
diagram, 281
and PMDs relationship, figure, 280
sublayers summary, 4278280
1000BASE-LX, 240
1000BASE-SX, figure, 280
baseband:
100 Mbps CSMArCD networks, 252274
1000 Mbps repeater set, 275

LAN transmission, 223
transmission, 94, 95
baseline performance, 732
basic:
access method, figure, WLAN, 332
protocol functions, 911
rules, ISDN Packet, 523
service BRI. 2B q D, 493
set identification BSSID., 319
signal, SONET, 571
structure of X.25 packets, 373374
structuring technique, OSI model, 1415
user channel, 484
X.25 packet structure, figure, 374
baudot code, 76, 77
BBE ratio BBER., 105
BBSSID field, 320
B-channel, 483
packet calls, 522
beacon management frames, 324
bearer channel bits, mapping into, 436
BECN backward explicit congestion
notification., 437
bit, 442, 601
clear, 454
set, 454
congestion control on a step count S, 454
rFECN bits, 449
indicator, 442
usage, 454455
beginning of scanning sequence, 131
BER bit error rate, bit error ratio.: 87
at digital cross-connects, 107
frame relay, 436
on fiber optic links, 647
of modern network, 433
performance of LANs, 204
requirements, LMDS, 723
at PLS, 244
on PSTN, 122
requirement, 177
token ring, 307
Bernard Bernie. Bossard, 717
BERT bit error rate test., 734
in place, figure, 738
bias distortion, 93
B-ICI signaling interface, 643
bigger picture, network management, 727728
billing, handled frame relay., 445
binary code:
conventions, 8485
for data communication, 7679
-to-decimal, 757758
INDEX
digit, 73
digital regenerator, figure, 82
GFSK, 333
NRZ, figure, 302
words, 66
binder groups, 535
bipolar waveform, 134,
AMI, 136
B-ISDN, 602
rATM functional layering, figure, 617
rATM network, 637640
protocol reference model, figure, 605, 606
reference model, 615
user-network interface UNI., figure, 605
BISUP broadband-ISUP ., 646
bit:
alignment, 616
binary transmission of information, 7375
error detection, 10
error interpreted as collision, 245
error performance, 102
error rate BER., 102, 103
VSAT, 427
integrity in a time slot, 150151
interval, 570
notation, 282
order, 212
packing, 165, 167, 595, 694
-parallel and -serial, 8586
parsing, 192
per Hz, getting more, 165166
period equals 1rbit rate, 161
rate, 6, 87
constant or variable, 620
capacity, 2, 436, 445, 551
on wire-pair varies with, 681
ersus antenna aperture, 428
stream flow resources, 684
stuffing, 138
synchronization, 225
time, 241
times BT. per meter, 287
shorter, 254
timing at 192 kbps, 497
slips, 612
symbols, 241
bitwise AND function, 358
block codes 6566
block diagram:
of fiber-optic repeater, 568
of fiber-optic transmitter, figure, 564
block error:
performance 102
rate, 103
783
block representation of signals, figure, 104

blocking probability, 147
blue noise, 568
B-NT1, B-NT2, 668
BNZS binary N-zeros substitution . codes, 135
BOM, 602, 604
bonding:
to building metal frame, 544545
conductor for telecommunications, 542
and grounding layout for a large commercial
building, figure, 541
and sizing the TBB 543
Bonds to the TGB, 544
boundary flow control, 28
BPDU bridge protocol data unit., 346, 347,
350351
BPSK binary phase shift keying., 166
BRI basic rate interface ., 484
differences, U.S., 499500
bridges., 204
802.1 application, 338
802.1 active topology, figure, 339
address, 344
connects segments, 3
diameter, maximum 349
hello time, 347, 349
hubs and repeaters, figure, 335
identifier, 346, 350
maps service data units, 339
numbers, 353
PDU BPDU., 346, 347
ports., 336, 346
protocol data unit BPDU., 350351
entity, 340
transparent, 336
transit delay not exceeded, 339
bridge forward delay, 350
bridge max age, 350
bridged local area network, 343, 347
topology, 340
bridged taps, 528
bridged topology design, 271272
broad is broad, how?, 551
broad spectral width, 565
broadband:
amplifier, 687
cable system electrical requirements, table,
249
cable systems, figures, 245
coaxial cable system, 247
CSMArCD, 10BROAD36, overview,
244245
data transport techniques, 551600
-ISDN, 602
784
INDEX
broadband Continued.
ISDN B-ISDN., and ATM, 601678
radio systems, 591592
ersus baseband LANs, 205
broadcast 57
address, 226, 299, 320
domain, 359
BSA boundaries, 327
BSSID basic service set identification ., 319
BT bit times. per meter, 287
Btag beginning tag.:
and Etag, 625
field, 626
buffer allocation size BAsize. indication field,
626
buffer storage, 65
building, campus wiring, cabling, data
communications, 524549
built-in clocking system, 91
burst and wait, 598
burst errors, 65, 611
burst noise:
downstream, 695
upstream, 696
burst profile attributes, table, 709
burst profiles, DOCSIS, 709716
burst timing convention, 711
bursty traffic, 445
of frame relay, figure, 446
figure, 446
bus network, 3, 5
busy hour, 367, 368
bypass in economically evolving nations, 152
bypass, local telephone service, 425
bypassing the local telephone company, 151
byte, 6, 9
interleaved, 571
interleaving, 576
ersus octet, 91
C-bit removed, 150
C-plane ATM., 606
cable:
balanced mode attenuation, horizontal
STP-A cable, table, 539
composition, fiber-optic cable, 563564
delay 287
budgets are smaller, 254
for segment, 286
conversion table, 287
television CATV., 686
plant, 682
terminations, 532
cabling practices, 532533

cache, 336
called station, 336
calculation table, interference, 721
call:
-acceptedrcall-connected packet, 382
and answer modems, 180
-connected packet, 382
control messages, 468, 523
establishment procedure, 468
signaling protocols, 521
-mapping database, 609
reference 470, 517518
flag, 519
values., 517, 646
request, 382
and incoming call packet, ISDN packet
via X.25, 524
format, figure, 383
packet, 521
setup, 382, 383, 520
called party information element coding, table,
479
called party number information element,
figure, 478
calling and working channels, 428
campusrbuilding cable topology, 548
campus star topology:
example, 547
figure, 547
campus wiring and cabling, figure, 529
CAP carrierless amplitude and phase
modulation., 683
capacitance unbalance, 538
carrier:
deference 230231
extension, 228
frequencies, 199
V.34, 180
ersus symbol rate, table, 180
frequency, 96
-frequency agile, DOCSIS, 703
sense, 230, 278, 326
mechanism, 327328
multiple access with collision avoidance
CSMArCA., 323
multiple access with collision detection
CSMArCD., 223232
transmission, 87
carrier-to-ingress power, upstream, 696
carrier-to-interference ratio, downstream, 695
carrier-to-noise ratio, downstream, 695
carrier-to-noise ratio, upstream, 696
INDEX
case A:
configuration, accessing PSPDN services,
520521
figure, 521
CAT 5:
cable 533
UTP, 7
categories of UTP cabling, 533
category 5 UTP cable, 530
CATV, 531
distribution system, early, figure, 687
spectrum assigned upstream and
downstream, figure, 691
CBR constant bit rate., 603
CCITT:
error performance objectives, international
ISDN connections, table, 104
International Alphabet No. 5 IA5., figure,
80
ITU-T, CRC, 63
Signaling System No. 7, 148, 488, 489, 493,
494
slip criteria, 148
voice channel occupies, 116
CDMA code division multiple access., 599
CD0s and CD1s, 241, 242, 243
CDV tolerance, 653, 654
cell, 601
arrival event, 649
B hub transmitter, figure, 719
-based transmission systems, 668
block, 650
boundaries, ATM, 665
may cross a C-4 boundary, 665
conformance, connection compliance,
656657
delay variation CDV., 622, 647, 648, 659
impact on UPCrNPC resource allocation,
654655
origins, figure, 655
sensitivity, 652
delineation, 673
algorithm, 614615
performance improvement, 612
scrambling, 613615
state diagram, 614
discrimination, predefined header field
values, 619
entry event, 647, 648
error ratio, 647, 649650
event, 647
exit event, 647, 648
785
flow, 616
header extraction, 618
generationrextraction, 617
loss priority CLP., 649, 654
field, 610
ratio, 647, 649, 659
objective, 652
misinsertion rate, 647, 650
payload, 647
rate decoupling, 618, 619, 673
rates, various connections, 666667
sequence integrity, 638
structure, ATM, figure, 607
time slots, 619
transfer delay, 647, 648
cells transported contiguously, 603
Cellular Vision USA, 717
central bridge, 232
CEPT30q2 sE1, table, 139
CF period, WLAN, 332
channel attenuation specifications, table, 308
channel coding, 65
with interleaving, 427
channelization, 64-kbps, 493
character, 92
characteristic impedance, 158
check pointing, 47
checkpoint recovery, 53, 54
checksum, 409, 418
UDP, 414
CHECK TOKEN PASS, 293
chromatic dispersion, 560, 562
CIR committed information rate., 445
equals access rate, 457
circuit:
CI, 248
maintenance messages employed, 447
and packet switching, 488
setup signaling procedures, 446
using ATM signaling, 645646
supervision, 603
switched services, 430
-switched data call control, 430
switching, 67, 68
class A format, 396
class B addressing, 396
classful addressing, 396
classless vs. classful, 394
Claude Shannon, 117
clear FECN bits, 453
clear request, 382
CLEC competitive local exchange carrier., 686
client server, 359
786
INDEX
CLLM consolidated link layer management.,

448, 601, 736
use, 452
CLNAP connectionless network access
protocol., 633, 635
protocol data unit PDU. structure and
encoding, 635637
-PDU, 635
-SDU 637
CLNP connectionless network protocol., 364,
415424
frame formats, 416
error report PDU, figure, 417
figure, 416
protocol, ISO 8473, 415424
clock rate, 147
clock recovery, 266, 621, 623
clocking, FDDI, 315
closed-loop system, ATT, 474
CLP cell loss priority., 610, 649
set, 610
CLPs0, 657
CLP bit, 652
value, 654
CLSF connectionless service functions., 633
clumping, 649
CM cable modem., 694
BER performance, 716
table, 716
and CMTS, forwarding agents, end-system
hosts, 696
and CMTS, IP and LLC hosts, 697
generates each burst, 709
CMIP common management information
protocol., 739740
CMISE common management information
service element., TMN, 751
CMOS digital ASIC processes, 276
CMTS cable modem terminator system., 694
MAC forwarder, 699
output, electrical, 714
table, 715
provides time reference, 701
coaxial cable electrical interface requirements,
249
code division multiple access CDMA., 599
code:
groups, 265266
length, 125
rate, 66
vectors, 611
word length, DOCSIS, 711
codes:
cause in CLLM message, table, 451
idle channels, idle slots, interframe time fill,

504, 505
coding, 570, 571
efficiency, 265
gain, 177, 595
of message types, table, 471
steps, 191
VSAT, 427
coexistence, DCF, PCF, 325
colliding transmissions, 245
collision:
back off, retransmission, 231232
delay, 285
detect and deassertion times, 271
detection and enforcement, 231
domains., 256, 271272
individual, 271
enforcement, 231, 246
fragment, 261
handling, 231, 239
presence function, 246
resolution, 60
collision window, 230, 274, 286
collisions, 7
on an ALOHA net, 428
color codes for horizontal 100-ohm UTP cable,
table, 534
COM continuation of message., 602, 604
combined station, 42
command PDU, 216
commandrresponse bit CrR., 441
commandrresponse field bit CrR., 511512
commands and responses, 42, 4952, 216222
commercial building telecommunications
cabling standard, 527
committed burst size Bc., 456, 471
field, 474
committed information rate CIR., 446, 455,
474
common channel signaling, 148
common management information protocol
CMIP., 739740, 747748
common management information service
element CMISE., 753
common mode attenuation, 539
common OAM cell fields, 677
common OAM cell format, figure, 677
common part indicator CPI., 625
field, 630
communication channels, network
management system, 739
communication, relay open systems, figure, 19
communication protocols, last mile, 696700
community name, 743
INDEX
companding, 129
compatibility interfaces, physical layer,
232235
illustrated, 234
compatibility issues, frame relay, 474480
complete packet sequence, 380
complete selection case, ISDN packet, 521
compliant ATM connection., 656
component delays, transmission system model
2, table, 287
concatenation indicator, 577
concentration and expansion, 144
concentrator trunk coupling unit, figure, 307
concept of frame, 131134
concept of the 3-dB power bandwidth, figure,
117
conceptual drawings: FDMA, TDMA and
CDMA, 599
conditioned diphase, 571
conditioning and equalization, 161162
configuration:
BPDU, 346, 347
parameters and format, figure, 350
bridge protocol data units BPDUs., 338
management, 728
messages., 337, 346, 347
confirm, primitive, 34, 209
conformance definition ATM., 656
conformance testing of cells, 653
conformant infrared devices, 334
conforming cell, 658
congested resources, 442
congestion:
avoidance, 446
control, 516
notification, 453
recovery, 448
CONNECT message, 646
connecting grounding bars, 545
connecting hardware for UTP cable, 536537
connection:
admission control CAC., 651, 653
cell rates, various familiar bit rate capacities,
table, 667
closed, 407
control, 10
establishment, figure, 645
ATM, 646
frame relay, 444445
flow control service, 210
management, 38
mode, 27, 620
opening, figure TCP., 405
release, 445
787
reset service, 210

termination services, 210
traffic descriptor, 653
connectionless:
LAN, 603
layer service, 633
mode, 26
data-link data transmission, 28
-oriented, ATM, 603
and connectionless services, 633637
data transfer TCP., 402
service, 8, 69
service, 8
concept, 356
protocol architecture for, figure, 635
functions CLSF., 633
-mode service, 207, 209
provided, 70
connections to the TMGB, 543544
connectivities, 12
connectivity, service equipment, figure, 542
connector pin assignments, 35
connectors, 569
consequences of errors, ATM cell header,
figure, 612
consolidated link layer management CLLM.,
448
messages, 449
constant bit rate CBR., 603
constellation point in, 194
constellation shaping, 179
container SDH., 585
container-n, 587
contention, 232
access, 428
for the medium, 6
process, controlled, 293
services, 324
window CW. parameter, 330
contiguous HDLC flags, 505
continuity cell check, 671
continuous or selective ARQ, 6465
continuous state leaky bucket algorithm, 658
control:
characters, 78
code groups, 268
of data-circuit interconnection, 29
encoding, 243
field C., 36, 44, 45, 47
ISDN, 510
none, 438
format, table, 513
formats, 213214, 510, 512513
of HDLC, figure, 45
788
INDEX
control Continued.
flags, 409
functions incorporated, 437
memory, 142
response times, figure, 652
store, 144
controls the connection establishment, 514
conventional header, 70
convergence sublayer CS., 618, 620
PDU CPCS., 624
convolutional codes, 6566
core and cladding, 557
core-CEI, 444
core-cladding interface, 561
correction mode, 611
correlate fault, performance information, both
directions, ATM, 669
corrupted by noise, 80, 81
corrupted frame, 60
cost, a routing parameter, 346
counter-rotating ring, 308
coupling efficiency, 565
coupling of multi-vendor equipment, 204
couplings between pairs, 536
coverage of CRC in a data-link frame, figure,
63
CP sequences, 193
CPCS common part convergence sublayer.,
616
characteristics, 628
CPCS-PDU, 624
format for AAL-5, 629
payload, 630
structure and coding, 628
CPCS-SDU sequence integrity, 628
CPCS-UU CPCS user-to-user indication ., 629,
630
CPE customer premise equipment., 724
CrR bit, 437
CRC-4:
block errors, 737
remainder, 736
CRC-16, 49
CRC-32 calculation, 637
generator polynomial, 630
CRC cyclic redundancy check., 49, 62
check, ISDN U.S., 499
error, 288
error checking, 503
field, 627, 630
indication bit CIB., 635
procedures, 505
value, 227
critical angle, 556
cross-connect:
blocks, 536
facilities, typical, 537
jumpers, 526, 528, 549
cross-point:
array, figure, 143
gates, 144
matrix, 142
cross-points, 139
crosstalk, 100, 101, 684
energy, 536
and fiber optics, 554
crystal transmit asserted, 304
CSI convergence sublayer indicator., 621
CSMArCA distributed algorithm, 323
CSMArCD carrier sense multiple
accessrcollision detection., 3, 223232
1000 Mbps networks, 274288
basic operation 229232
LAN, 734, 735
MAC functions, figure, 236
sublayer, 230
relationship with OSI reference model,
figure, 224
system designation, 239240
CSUrDSU, 150
CTS clear to send., 319
frame, 331
figure, 322
cumulative latency, 304, 305
curing ills of bandwidth scarcity, 691692
customer-owned OSP cabling, 549
cut-through forwarding, 356
CW max, 331
cyclic redundancy check CRC., 49, 6264
D bits, inverted pointer, 588
D channel:
access control, 497
call establishment, 488
carries out OSI layers 1, 2, 3, 488
correspondence with CCITT SS No.7, OSI
model, figure, 495
failure, 514
ISDN, 484
signaling data, 489
information, 488
protocol, 523
DACS digital cross-connect ., 684
DAMA demand assignment multiple access.,
428, 429
DARPA Defense Advanced Research
Projects Agency., 39
data and PAD fields, 227
INDEX
data burst, 598

data circuit, 29
identification, 30
performance, 102107
data code groups., 268, 269, 282
data communication equipment DCE., 35, 108
data communications, office environment:
part 1, 203288
part 2, 289360
data delay, PCS to PMA, 270
data encoding, CSMArCD, 241
data error recovery, 35
data forwarding through CM and CMTS,
697698
figure, 698
data frame intervals., 192, 194, 198
data link:
addresses, 26
connection, 25, 26, 208
endpoint, 28
identifiers DLCI., 25, 27, 439, 440,
506, 507
establishment and release, 28
identifier DLCI., 70, 442
splitting, 27, 28
between two or more SAPs, figure, 506
frame, 370
figure, 36
header, 390
layer, 2629, 3537
reference model, figure, 508
management, 29
messages, 507
sequencing, 210
semantics, 699
message units, 506
protocols, 35
serice data unit SDU ., 25, 27, 28
mapping, 28
3-dB bandwidth concept, 116, 117
data loop back at PMD, 279
data message, 9
data mode modulation parameters, 179
data modems, 165200
data modulation terminology, 96
data multiplexers and statmultiplexers,
366367
data network connectivity, CATV, 692
data network operations, 5771
data nibbles, 262
data occupancy, 155
data offset, 409
data protocols, key to network operation, 711
data quartets, 311
789
data rate, 87
signaling rates, 198
for different K and S, table, 193
token ring, 303
data switching, 6771
methods, summary, table, 70
data terminal equipment DTE., 35
figure, 108
data transfer, 38
data transmission:
I, 7387
II, 89119
on digital network, 149151
data unit identifier, CLNP, 419
database management process, TCP, 404
datagrams., 387, 390, 391
format, IP, figure, 392
maximum length, 393
DC balanced by last bit, 497
DC baseband, 95
DC component, 302
DC resistance:
150-ohm STP, 538
unbalance, 538
DCC data country code., 644
DCE data communication equipment., 35, 108
pads information field, 371
DCF, 326
access method, 331
basic access procedure, 331332
IFS DIFS., 329
DDS digital data system., 150
DE discard eligibility. bit, 442, 455, 458
de-assertion, CRS carrier sense., 260
decimal digital representation, IP addresses,
759760
decimal number:
to binary number in addressing, 757762
to binary sequence, 758759
shorthand binary number, 757
decision, binary receiver, 82
decoding delay, 67
dedicated circuits., 365366
figure, 364
dedicated connecting hardware, 533
dedicated detection of unavailability state, 670
default:
address format, 463
mode, HEC, 611
operational parameters IEEE 802.1D, 348
routing, IP, 397
defer access, figure, 332
deferrals, 324
defragmentation, 325
790
INDEX
degradation in BER, 122

degraded minutes, 103, 104
delay:
in call setup, 494
distortion, 100
equalizer, 163
figure, 164
for network media segments, 251
model 1, table, 285
over satellite circuits, 123
varies from band center, 160
delete bit, 462, 467
delimited by flags, 507
delimiter error, 257
delimiting and synchronization, 28
delivery confirmation bit, 377
delta times consecutively, 614
demand access, 520, 521
demand for frequencies, 595
depressed cladding, 562
derivation of currently transmitted symbol
quadrant, table, 706
design variables, VSAT networks, 427428
designated bridge, 338, 345
designated port, 345
destination:
address DA., 36, 394
and source address, 635
field, 226, 320
figure, 636
specific CLNP., 421
-controlled transmitters, 448
LAN, 337
MAC address, 343
port, 409
UDP, 414
service access point DSAP., 388
and source address fields, 298
unreachable message, 400, 401
desynchronize tributary signal, 575
desynchronizing phase-locked loop PLL., 576
detailed IP operation, 391
detected VPC failure, 671
detection mode, 611
detection of lost packets, 377
developing a hypothesis, 733
development, PCM signal, 123131
3 steps, 124
DHCP dynamic host configuration protocol.,
DOCSIS, 700
diagnostic packet, 375
dial-up connection, figure, 364
diameter of completed cable, 534
differential:
encoding rule, 195

impedance specifications, table, 308
Manchester coding, 302, 308
quadrant coding, 704
V.32 modem, table, 171
differentiated service, 412
diffuse infrared transmission, 334
DIFS DCF interframe space., 328
period, 332
digital:
channel, 483
coding line formats., figure, 570
data cross-connects, 685
data system, 108
North American, 150
exchange, 140
hierarchy, 571
interfaces at 1.544 Mbps, table, 503
LOS microwave terminal block diagram,
figure, 594
modem, 188
encoder block diagram, 192
modulation schemes, 570
modulator, 594
satellite system operation, access, 598599
section level, F2, 668
signals, 570571
subscriber line DSL., 499, 682
switching,
overview, 139146
voice channel, 136
digitally multiplexed signal, 497
digitized program channels for broadcasters,
484
direct routing, IP, 397
direct sequence DS., 599
direct sequence spread spectrum DSSS.
operation, WLAN, 334
directing data message, 59
DISC disconnect command., 51
discard eligibility DE., 456, 601
indicator, 458
bit, 442
discard eligible, 456
discarded frames, 442, 455
disconnect DISC. command, 221222
disconnected mode DM., 52
discrete levels, 126
dispersion effect, 649
distance vector, 399
distortion and dispersion, 558561
differentiated, 558559
distribute CATV signals to customer premises,
690
INDEX
distributed:
coordination function DCF., 323324
frame, 6
initialization, 290
queue dual bus DQDB., 601, 602
random integer, 231
DLCI data link connection identifier ., 70,
439, 440, 442, 506
of the congested circuit, 449
on the D Channel, 442
equal to 0, 465
field, 442
range, 443
value, 444, 460
table, 443
for D Channel, table, 444
DL:
CONNECT, 221
CORE:
frame relay, 436
connection endpoint identifier CEI., 444
control bits, 442, 444
parameters, 444
sublayer, 445
DISCONNECT, 221
DM disconnected mode., 52
DNHR dynamic nonhierarchical routing., 146
DOCSIS data over cable service interface
specification ., 692
reference model figure., 693
specification, 692716
dot fragment bit set, 393
flag, 401
dot11ShortRetryLimit, 331
dotted decimal notation system, 760
double current meaning polar, 83
downstream, 188, 682
frequency plan, 714
PMD sublayer, 714
RF channel transmission characteristics,
table, 695
transmission convergence sublayer, 701
DQPSK modulation, WLAN, 334
DS1:
coding, 131
configuration, separate, 485
connection, 150
ESF extended superframe . ATM., 662
family of waveforms, 128
mapping:
ATM, 660661
ATM cells with PLCP, 662
time slot, 155
DS3:
791
frame:
transporting ATM, 659660
structure used to transport ATM cells,
661
PLCP, 659
tributary, 575
DS series line rates, tolerances, line codes,
table, 138
DSAP destination service address point., 211
address, 217
and SSAP address field formats, figure, 212
DS0rE0 channels, 436
DSL digital subscriber line., 499, 682
modem router, 685
reference model, 685
technique, 485
DSLAM digital subscriber line access
multiplexer ., 684
facilities, 685
DSP domain-specific part., 644
DTE data terminal equipment., 35, 108
address field, 385
and DCE data packet formats modulo 8,
378
full media independence,
-DCE interface, 108116, 375, 377, 379
PVC, 463
X.25, 373
dual cable:
frequency allocations, table, 250
system, figure, 245
dumb bridges, 335, 336
dummy call reference, 461
durationrID field, 326
duty cycle, 138
DWDM dense wavelength division
multiplexing., 569
dwell time boundary, 332
Dykstra algorithm, 399
dynamic mix of contention-, reservation-based
upstream transmit opportunities, 701
dynamic range, 566
WLAN receiver, 333
E bit, 301
E1:
family:
of multiplex, 139
of waveforms, 128
mapping ATM cells, 662664
multiframe, figure, 737
E3, E4 frames, ATM cell mapping, 664
E.164:
addressing, 357
792
INDEX
E.164 Continued.
ATM address format, 644
or X.121 addressing, 384
earth station antenna sidelobes, 597
eavesdropping, 323
EBCDIC code, 90, 91
figure, 81
EBCDIC extended binary coded decimal
interchange code., 79
echo, 159
cancellation, 501
techniques, 178
suppressors at each end, 497
and singing: telecommunication network
impairments, 159
ED alignment error, 303
EDD envelope delay distortion., 165
EDFA erbium-doped fiber amplifier., 568
amplification, 569
effect of source spectral width, fiber index
variations on material dispersion, figure,
560
EFS end-of-frame sequence., 294
EFS error-free seconds., 103
EGP exterior gateway protocol., 398
EIA Electronic Industries Alliance., 35, 108
EIA-530 connector contact assignments, table,
113
EIArTIA:
-232, 85, 93, 108109
electrical interface, 110112
pin assignments, table, 109
-530, 108109
interchange circuits, table, 114
-568A, 531, 532, 534
-607, 543
circuits by category, table, 110
interface standards 530, 422 and 423,
114116
recommends, 528
EIFS extended interframe space., 328
medium idle time, 330
8-ary PSK, figure, 167
8Br10B transmission code, 282
8B6T coding, figure, 163
eight-level coding of North American DS1
system, table, 133
elastic buffer, 303, 305
elastic store, 147
elasticity buffer, FDDI, 315
electrical:
bit decisions, 7982
characteristics, interchange circuits, 181
communication, information, 73
noise, 79
output from CM, table, 714
representation, binary data, 8284
electromagnetic compatibility EMC., 554
Electronic Industries Alliance EIA., 35, 108
electronic mail, figure, 386
element length, type, payload ATM., 637
elements of a MAC frame, 225227
elevation angle, 426
eligibility indicator, 456
elliptical orbits, 597
E & M signaling, 157
embedded delimiter, 311
encapsulate, 387
encapsulation, 9, 10, 268269, 618
encoder, 186187, 199
block diagram, V.34, 186
encoding:
bridge identifiers, 351
decoding. GMII data octets, 278
port identifiers, 352
option parameters, table, 420
root path cost, 351
timer values, 352
end delimiter ED., 291
end of transmission delimiter ETD., 241
end tag Etag. field, 626
ending delimiter ED., 300, 303
FDDI, 313
figure, 304
ending flag, 92
endless loops, 394
endless routing loops, 339
end-of-frame sequence EFS., 294
end-of-stream delimiter ESD., 269270
TrR, 268
end-system function, 38
end-to-end:
connectivity, 68
data delivery, 38
F4 flow, 669
F5 flow, 670
flow control, 24
performance monitoring, ATM, 676
segmenting, 24
transit delay, 453
energy per interference bit, 720
enterprise field, 745
enterprise network, 1
environment, 139
entities:
service access points SAPs.:
and endpoints, figure, 506
and identifiers, figure, 34
INDEX
entrance facilities, 533

envelope delay distortion EDD., 99100, 161,
162
envelope delay values, 163
EOC embedded operations channel ., 499
EOM end of message., 602, 604
EPOs error performance objectives., table,
105
equalization, 136, 163
equipment rooms, 533
equivalent:
bandwidth, function of information rate, 128
binary number, 167
noise temperature, 101
versions, generic cell rate algorithm
GCRA., figure, 658
error:
in ATM cell header, figure, 612
burst, 63, 103
control, 910
correction schemes, 6467
detected bit E bit., 298, 301
detection, 27
and recovery, 28
code EDC.OAM., CRC-10, 678
schemes, 6164
events G.826, 107
notification, 25, 27
performance, 102103, 147, 595
Bellcore perspective, 107
ITU-T perspective, 103107
objectives EPOs., 105
token ring, 307
rate, 31, 87
recovery, 25, 28, 5355, 515
frame relay, 434
procedures, 508
report CLNP., 418
data part, CLNP, 424
function, CLNP, 420
PDU, CLNP, 423424
status field, 745
errored block, 105
errored cell outcome, 648
errored second, 104, 105
error-free:
no system is, 61
operation, 63
seconds EFS., 103
ES ratio errored second ratio., 105
ESF extended superframe ., 736
ESI end system identifier ., 644
essentials, TV signal distribution, CATV
system, 688690
793
establishment of data-link connection, 514

Ethernet, 3, 6, 232
CSMArCD, 204
limits, 391
market share, 290
ether type, 388
assignments, table, 389
Euclidean distance, free, 177
even parity, 61
event, 731
events, actions, time scales, response times,
651652
excess burst size Be ., 455, 471
excess spare capacity, 152
exchange codes, 357
exchange identification:
frame, 449
XID, 52
exclusive-OR, 66
expiration of timer, 304
explicit notification, 447
explorer frames, 354
exponential increase of CW, example, figure,
331
extend upper useful frequency limit, 689
extended binary coded decimal interchange
code EBCDIC., 79
extended IFS EIFS., 329
extended multiframe, 503
extended superframe, 503
extending CATV coverage area, 690692
extending the control field, 45
extension field, 228
extension indicator, 44
exterior gateway protocol EGP., 398
extract cell flow out of transmission frame,
618
extranet, 360
F bit:
acknowledges P bit, 47
set to 1, 216
F4 flow, 619
bidirectional, 669
and F5 flows, 669670
initiated at or after connection, 669
mechanism, 669
relates to virtual path, 671
F4 OAM cells, 619
F5 flow:
mechanism, 670
relates to virtual channel, 671
must be terminated, 670
fading, 103, 595
794
INDEX
fairness, 456
false delineation, 615
far-end fault generate, 266
far-end user, 58
fast collision interference, 326
fast select facility, 382
fault condition notification, 30, 31
fault down to card level, 731
fault management, 728
FC frame control. field, 298, 313
FCS frame check sequence., 36, 37,
62, 300
computation, 228
field, 48, 71, 227
WLAN, 321
processing, 6061
slot, 62
FDDI fiber data distributed interface ., 5, 6
code symbol set, table, 311
frame format, figure, 312
ring operation, 314315
timers, timing budget, 314
token format, 310
topology, figure, 309
FDMA frequency division multiple access.,
428, 598
FEBE bit far-end block error., 499
bit indicates, 500
FEC forward error correction., 6566, 177
bit-to-symbol ordering, 707
encode, DOCSIS, 707
encoder, 66
operation, simplified diagram, 66
FECN forward explicit congestion
notification., 437
bit, 441, 601
bit set, 453
and BECN bits set to 1, 452
indications, 448
usage, 453454
Federal Express, 59
feedback control, 651
FER frame error ratio., 334
FH PHY WLAN, 332
fiber distributed data interface FDDI., 6,
308315
fiber jacket, 564
fiber optic:
amplifiers, 568569
cable, 563564
link, block diagram, 554
receiver, block diagram, 567
transmitter block diagram, figure, 564
transmitters, 564565
transmission, 552571
field and frame boundaries, 148
field descriptions, CLLM message format, 450
field mapping conventions, 440
field number, IP Protocol, table, 395
file servers., 5
file transfer protocol FTP., figure, 386
fill, 296
fillridle patterns, 93
filter roll-off properties, 595
filtered, 337
filtering database, 340, 342
filtering information, 339
final bit, 47, 55, 213
final frame, 53
firewalls, 360
first, second bandwidth approximations,
116117
fixed equalizer, 136
fixed latency buffer, 304
fixed-phase relationship, 589
flags., 91
CLNP, 417418
fragment indication, 393
flag field, 36, 43
flag sequence, 44, 510
flat:
channel, 98
frequency response, fiber optic cable, figure,
689
power spectrum, 100
rate payment, 445
response, 688
floating and locked modes, 578
floating VT mode, 578579
flooded, 399
to all users, 399
flow control, 10, 25, 28, 38, 508, 601
and CLLM, 739
FECNrBECN bits, 453455
mechanism TCP., 403
none in IP, 391
parameter negotiation, 379
ready state, 379
user data, 516
window, 377378
flow label, IPV6, 411, 413
flows F1, F2 and F3, 670
FMIF size, 475
FOIRL, 256
forced insertions, 677
format, 551
effectors, 78
identifiers, 45
INDEX
link layer core parameters information

element, 474
of pseudo-header, figure, 415
formation, ATM cell from user frame,
figure, 624
forming conclusions, 734
fortuitous distortion, 94
forward delay, 350
forward error correction FEC., 6566, 177
forward explicit congestion notification
FECN. bit, 441
forwarding rules DOCSIS., 698
4Br5B code:
table, 311
groups, table, 267
4D symbol interval, 183
four bit nibbles, 265
four-bit bundle, 257
four-pair, 22 AWG cables, 534
four-pair UTP, 528
four-state system, 74
four-wire transmission, 156157
four-wire transmit path, 158
FR network, Pacific Bell, figure, 449
FR switches, 446
figure, 449
FRAD frame relay access device., 437438,
446, 459
fragment:
datagram, 393
number field, 321
offset, 393, 400
zero, 399
fragmentation, 391, 393
needed and DF set, 401
and reassembly, 402
frame, 9, 132
abort, 440
adaptation, 617
or packet of data, generic, figure, 92
PCM, 126
frame alignment, 504
North American, 132
loss, 595
signal, 132
word, 736
frame body, 325
field, 321
frame check sequence FCS., 36, 37, 300, 440
field, 4849, 227228, 510
frame control FC., 297
field, 291
WLAN, 318
frame-copied C. bits, 301
795
frame delimiting, 437

ISDN, 507
frame delimiter, 44
frame elements characteristics, CSMArCD,
240
frame encapsulation, 291
frame encoderrdecoder, figure, 306
frame error ratio FER., WLAN, 334
frame field description, fixed part, CLNP,
416418
frame filtering, 338
frame formats.:
FDDI, 312313
IP datagram, description of functions,
391392
for layer 2 frames, ISDN, figure, 509
WLAN, 318
frame handler, 460
frame length, FDDI, 312
frame numbering, 10
frame operation HDLC., 5255
frame overhead and frame relay, 434
frame reception, 237
frame relay, 60, 69, 70, 433480, 601
alive and healthy, 446
ANSI frame format, figure, 438
circuits, 7
compared to X.25, table, 434
forum, 436, 474
network, figure, 435
revenue bearing traffic, 447
signaling functions, 460480
switches, 455
repetition rate, 504
frame size limits, CSMArCD, 223
frame status FS., 301
field, FDDI, 313
frame structure, 437443
across AUI,
of DS1 to transport ATM cells, 662
of DS3 to transport ATM cells, figure, 661
DOCSIS, 711
ISDN, 504505
of multiplexed bit stream, 2BqD, figure, 496
at 1.544 Mbps interface, figure, 503
at reference points S and T, ISDN BRI, 496
SDH, 588590
frame switching, 179
frame synchronization, loss of, 148
frame synchronous scrambling, 572
frame transmission, 237
frames structure, flexible burst length mode,
DOCSIS, figure, 713
796
INDEX
frame type bits, 297

framing, 199
balance bit, same polarity, 498
bit, 132
functions, 28
and indexing, V.34 figure, 183
parameters, V.34 table, 184
pattern, 659
rate, 133
signal with balance bit, 498
and timing distinguished, 134
V.34, 183184
frequency:
allocation and licensing, 554
division multiplex, 598
hop FH., 599
hopping spread spectrum operation, 332
modulation FM., 96
range:
downstream, 695
upstream, 696
response, 98, 161, 162
reuse, 596, 718724
frequency shift keying FSK., 166
FRMR information field, 222
format, figure, 222
front-line command post, 729
FSK frequency shift keying., 166
modulation, 96
full availability, 139
full-duplex point-to-point link, 232
full mesh network, 425
full status, PVS, 467
functional:
architecture, ATM, 633635
capabilities of MAC sublayer, 236237
description selected interchange circuits, 112
diagram, regenerative repeater, 135
fields, CLLM, 449
functions:
within data-link layer, 27
individual ATMrB-ISDN layers, 616620
within PCS, 279, 281
within physical layer, 31
service primitives, figure, 507
fusion splice losses, 569
G.826 error events, 107

gap between continuous frame services, 323
gateway, 37
routing tables, 401
gating, 125
Gaussian distribution, 101
Gaussian FSK, 333

GBSVC general broadcast signaling virtual
channel ., 641
GCRA generic cell rate algorithm., 657
configuration, 657
general broadcast signaling:
cells, 619
virtual channel, 641
general format identifier, 374, 375
generating polynomial, 62, 63, 175, 510,
614
V.34, 183
generator polynomial, 612, 627
generic:
cell rate algorithm GCRA., 657
communication context, ISDN, figure, 494
flow control GFC., 607
applied, 619
frame, data link layer, figure, 36
information models, 749
primitives, 34
SNMPv.2 managed configuration, figure, 748
genesis of frame relay, 435436
geographically distant LANs, 354
geostationary satellite ., 598
GET, 742
retrieves data, 742
GetBulkRequest, 747
GET-NEXT, 742
GET-RESPONSE frame, 746, 747
gigabit Ethernet CSMArCD., 274, 275
gigabit Ethernet extends, 274
gigabit media independent interface GMII.,
235
global DSAP address field, figure, 212
globally unique, CLNP, 421
GMII data stream, figure, 276
GMII gigabit media independent interface .,
260, 276277
signal CRS, 281
go and return paths, monitoring of, 737
go-back-n ARQ, 47, 48, 65
graceful close, 404
graceful connection close, 405
two examples, figure, 407
graded-index fibers, 559
graphic symbols, 75
ground fault return path, 542
ground reference, telecommunication systems,
540
grounding and bonding infrastructure,
commercial buildings, 540546
grounding electrode system, 541
group 3 facsimile, 178
INDEX
group address used, 299

group delay:
distortion, 161
limits relative to minimum measured, 162
figure, 162
group field value, 450
group TEI, 512
growth management, 729
guard time between bursts, DOCSIS, 711
H0 channel s., 485, 505

H1 channels, 485
H4 pointer, SONET, ATM, 666
H12 channel s., 485, 505
H-channel interface, 484
H. Nyquist, 117
half-duplex and full-duplex, 232
hamming distance, 66, 298
hamming free distance, 177
Handbook of Computer-Communication
Standards, 37
HDB3 high density binary 3., 135
HDLC high level data link control., 9
commands and responses, table, 50
frame format, figure, 44
frame structure, 4349
rules, 511
similarities, 370
stations and configurations, figure, 42
HDSL high-speed digital subscriber line., 684
headend, 687688, 690
location, 244
header, 92
OAM, 678
header check sequence HCS., 602
header checksum, 394, 400
header error control HEC. field, 610
header error control sequence generation,
611612
header extension, 635, 637
header extension length HEL., 635
header information, 103
header pattern for idle cell identification,
table, 613
header structure, ATM cell, figure, 607
health and welfare:
checks, 460
of a circuit, 69
of PVC link, 463
heavy rain storms, 427
HEC header error control.:
-based mapping DS3., 660
bits, 613
797
field, 610611, 612

receiver modes of operation, figure, 611
sequence, 618
value, 610
hello time, 547, 350, 351
HFC hybrid fiber-coax., 690
hierarchical network, 4
high latency, 7
high-level data link control HDLC., 9
typical data-link layer protocol, 4155
high-stability clocks, 148
higher:
capacity STMs, 585
layer protocol identifier HLPI., 635
level:
network synchronization, 148
SONET signals, 572
higher order:
digital multiplexing formats: SONET, SDH,
571591
PCM multiplex, 136139
star network, 35
virtual container-n to VC-n, 587
hold time, 349
hop-by-hop service enhancement protocols, 24
hop count, 398, 399
hop frequencies, WLAN, 333
hop limit, 412
horizontal cabling, 528530
horizontal cross-connect, 528, 530, 532
horizontal UTP cable, 534535
attenuation, table, 535
host, 387
address, 357
portion, 358
ID field, 396
hot-line or complete selection methods,
520
HRC hypothetical reference connection.,
430431
HRP hypothetical reference path., 105
hubs., 5-6, 426
port processorrrepeater, 356
technology, improvements, 355356
-and-spoke, 425
and switching, 355356
HUNT state, 614
hybrid, 157
ring control, 310
transformer, operation of, figure, 158
hydroxyl ions, 561
hypothetical reference connection HRX., 148
for public synchronous data networks, 429,
430431
798
INDEX
hypothetical refererence Continued.

reference path HRP., figure, 106
hysteresis due to inductance, 95
I-format, 214
I format control field, 45
I frames, 45
IA5 code, 78
IAM initial address message., 646
ICMP internet control message protocol.,
399400
fields, 401
figure, 386
message format, figure, 401
messages, 399
protocol, 393
identification and parameter exchange, 29
identifier:
coding, 471
field, 393
IDI international domain identifier ., 644
IDL detection time, 242
IDL line condition, 242
idle:
and busy condition of time slots, 141
or busy, 73
cell:
coding, 619
ersus unassigned cell, 619
code groups, 268
groups conveyed, 270
sequence, 52
state, 52
IEEE:
802 addresses, 357
802 committees, 205, 290
802 frame, LLC and TCPrIP functions,
figure, 390
802-type LANs, 16
802.1 bridge active topology, figure, 339
802.1 bridge application, 338
802.1 bridges, 338352
802.3 MAC, 274
802.5 standard, 308
IETFrInternet Society, 412413
IFIndex interface index., 397
IFS interframe space., 324, 328329
relationships, figure, 328
IHL internet header length., 391, 400
IM noise intermodulation noise., 101
products, figure, 101
impart larger quantum steps, 129
IMPDU initial MAC protocol data unit., 602
implementations:
10 Mbps, parameters, table, 238
1000 Mpbs, table, 239
impulse noise, 100, 102
spike, 82, 103
inbound, VSA, 428
in-building wireless telephone systems, 725
incoming:
call packet, 381
ISDN, 524
committed burst size, 477
time slots, 141
incompatible with voice channel, 155
incomplete:
frames, 10, 60
octets, 10
incorporation of upper layer PDUs into
data-link layer frame, figure, 388
independent mappers, 194
index, propagation profiles, modes, figure, 563
index-matching fluid, 569
indication, primitive, 34, 209
individual addresses identify, 299
individual and group addresses, 299
inductive loading, 122
infinite bandwidth, 553
INFO field, 299300
maximum size, 438
information bits, 66
information element, 462
unique identifier coding, table, 472
other, 520
information field, 36, 48
INFO, 299
ISDN, 510
minimum, 438
information:
modeled flow, 208
payload, SDH, 584, 588
throughput, 67
transfer commandrresponse, 213
transfer format, 214
control field bits. figure, 219
infrared IR. transmission, WLAN, 334
ingress:
and access rate, 477
and egress access rates, 458
node, 457
initial shaping sign bit, 194
initiation, metasignaling for SVCI assignment,
642
inputroutput IrO. devices, 108
insertion losses.:
INDEX
of connectorsrsplices, 569
of DWDM systems, 569
of a hybrid, 158
insertion time of a cell, 652
in-service monitoring, 736
insulation displacement contact IDC., 536
in-sync time ersus bit error probability, figure,
615
integrated services digital networks ISDN.,
483525
intensity modulation of light, 570
intensity range of voice signals, 126
interaction management, 22, 23
interbridge protocol operation, figure, 342
interbuilding backbone cabling, 546
cabling, figure, 529
interchange circuits, 199
V.90, table, 200
interconnect, 151
LANs at a distance, 445
interconnection:
data circuits within physical layer, figure, 30
examples of, figure, 32
STM-1s, 590591
interconnects and bypass, 11152
interface:
2.048 Mbps, 504505
existing North American plant, ISDN, 499
frame structure of SDH, 584587
functions, ISDN, 495497
layers, SONET, 581582
figure, 581
points at PMD sublayer, 702
requirements, CSMArCD, 236
UNI and NNI, 446447
interference level into receiver in cell A, figure,
720
interframe:
fill, 504, 505
idle flags, 455
period, 260, 276277
time fill, 52
interFrameGap, 251
interim interswitch signaling protocol IISP.,
643
interleaver, 427428, 428
intermediate cross-connect, 532
itermediate frame bit I bit., 300
intermodulation noise, 100, 101
intermodulation products, 101
International Alphabet No. 5 IA5. code, 78
international ISDN number, 645
International Telegraph Alphabet ITA. 2
799
code, 77
Internet:
control message protocol ICMP., 399400
datagram envelopes, 402
header and data, 388
plus first 64 bits, 401
protocol IP., 37
address formats, figure, 396
packets transparently, 696
internetworking equipment, 733
interpacket gap shrinkage, 285
due to repeater units, 271
interpreting serial stream of bits, 8993
interpreting AU-n pointers, 588
inter-repeater fiber segment, 2542
intersymbol interference ISI., 94, 100, 136,
177
impact on BER, 117
intra-, inter-building distances, table, 532
intranets and servers, 359360
introduction:
digital networks, 123139
frame relay, 436444
trellis-coded modulation, 177178
invalid:
code groups, 268
frames, 440
MAC Frame, 228
IP4, 357, 391
IP6, 391
IP Internet protocol.:
address, 358, 391, 394
algorithm, 358
connectivity, 686
datagram format, figure, 392
encapsulation, 388
fields, 400401
gateway, 391
header, basic, 399
inter-connect two LANs, figure, 387
operation, detailed, 391
protocol field numbering, table, 395
router, figure, 387
routing, 397398
function, 390391
table, figure, 397
telephony, 725
IPV6 Internet protocol version 6., 410413
addressing, 413
header, 411412
format, figure, 412
packets, 412
routing, 410411
800
INDEX
IPV6 Continued.
IPV4 DOCSIS, 700
IrQ mapping table, DOCSIS, 704
IR PHY, WLAN, 334, 335
ISDN integrated services digital networks.,
483525, 602
access connection element, 491
basic rate, 122
BRI, 682, 683
circuit-mode bearer services, 521
compatible, 487
D channel, 439
generic users, figure, 486
H4 channel, 576
LAPD core functions, frame relay, 437
networks, 490493
numbering plan, 523
and OSI, 493494
packet mode review, 520525
packet service, figure, 490
protocol:
structures, 493505
issues, 488
reference model, figure, 486
64-kbps digital channels, 149
structures, 483485
user channel, 483484
user connectivity to the network, 485
user part ISUP., 489
ISI intersymbol interference ., 94, 100
ISM band, 316
ISM in-service measurement., 104
ISO International Standards Organization., 9
8348, 37
8072, 39
8073, 39
Geneva, 13
isochronous data transfer, FDDI, 310
ISP:
connectivities, 724
to local serving exchange, 682
interexchange:
carrier IXC., 151
data, issues, requirements, 5861
ITA2 code, figure, 76
J and K symbol, 312
jabber function, 246
jam, 230
size, 231
jamming, 231
jamSize, 238
jitter:
accumulation, 553
considerations, CSMArCD, 242

impairment, 236
and optical fiber, 553
joining, fibers of splices, connectors, 569
jumbogram, 393
justification equals stuffing, 138
justification bits, 575
key variable, 599
keying, 96
LAB CSMArCD layers, figure, 258
label M-1, 194
labeling, packets, 410
LAN:
architecture related to OSI, figure, 206
bridges, 335354
drawbacks: proliferation and loops,
337338
protocols relate to OSI, 206
data frame, 354
equivalent, 323
interconnections, 435
interconnectivity, 363, 438
to-LAN connectivity, 335
router, figure, 435
token ring, 294
and WAN, 1
-to-WAN-to-LAN connectivity, 386
LAPB link access procedure B-channel ., 41,
369, 386
discussion, 371372
frame formats, figures, 371
frame information, 370
LAPD link access procedure D-channel ., 41,
435, 505514
address field format, figure, 511
primitives, 514
large grouping of frequency reuse cells, figure,
722
largest frame, 353
laser diode LD., 566
sources, 560
last-mile data distribution systems, 681725
late collision, 228
latency, 7, 60, 303304
equation, 60
layer 1:
basic rate, interface, 494501
primary rate, interface, 502505
layer 2:
interface, ISDN for LAPD, 505514
logical link connection, 519
INDEX
layer 2 Continued.
switch, 356
unnumbered information UI. frames, 460
layer 3:
entities, 506
ISDN Overview, 514520
switching, 356
layer:
descriptions, 1932
interaction, SONET, 583
interface requirements, 236237
layered OSI model, 14
layering, 1415
in cooperating open systems, figure, 14
of D channel, 494
LCN logical channel number., 385
LD laser diode., 564
leaky bucket:
algorithm, 658
concept, 434
leaky modes, 562
learning, 348
bridge, 356
process, 342343
least or most significant bit transmitted first,
758
LEC local exchange carrier., 58, 146
LED light emitting diode., 564
transmitter, 540
and LDs, comparison, table, 566
length:
field, 626, 628, 630
indication LI.:
CLNP, 417
field, 627
type field, 226227
LEO satellites, 598
light and small, optical fiber, 553
light-emitting diodes LEDs., 560, 565
line:
code, 498
configurations, 501
layer, SONET, 583
modulation rate, FDDI, 308
overhead, 583
probing, 179
rates and codes, 139
signaling, 73
signals, V.34, 179180
line-of-sight, 592593
microwave, millimeter-wave transmission
systems, 591600
of each other, 721
801
link:
access procedure:
balanced LAPB., 41
D channel LAPD., 41, 505514
distances not to exceed, 285
encryption system, 175
integrity verification, 462, 466
frame relay, 465
only, 467
layer core parameters information element,
471
specific coding, table, 475
-layer security sublayer, 700
-layer transparent bridging, 698
margin, LMDS, 723
segment, 1000 Mbps, figure, 275
state:
-database, 399
protocol, 399
vector, 399
links, 147
in tandem, no more than 12, 122
LLC logical link control., 41, 206222, 291
commands, related responses, table, 216
control field format, 214
frame, 297, 298
format, 299300
information field, 211
and MAC sublayers, 294
PDU, 214, 229
format, figure, 211
structure, 211
relationship, LAN reference model, figure,
207
sublayer, 231
management:
function, 212
services, 207
services and primitives, 208211
LMDS local multipoint distribution system.
485, 683, 686, 716724
concept, figure, 718
frequency band, 717
IP router, 718
local:
address, 357, 396
area network LAN., 203
COLLISION DET, 248
exchange carrier LEC., 686
flow control, 607
multipoint distribution system LMDS., 485,
683, 686, 716724
locked VT Mode, 580581
logic gates., 142
802
INDEX
logic gates. Continued.

enabled, 144
logical bits or symbols, associated timing
information, 617
logical channel s., 368369
assignment diagram, 385
control, 37
group number, 373375
number, 374, 375, 368
logical data links, 207
logical environment, 32
logical identifier, 459
logical link control LLC., 41, 206222, 295,
700
definition of an Invalid LLC PDU, 213
logical ring, 290
on a physical bus, figure, 290
logical tree structure, 344
logically disconnect mode, 222
LongRetryLimit, 331
look ahead depth parameter, 197
lookup table, 391
loop formation, figure, 337
loopback, 738
examples, 738
loops can occur, 337
LOS microwave, typical installation, 593596
loss:
added, 159
of AU pointer, 673
of cell sync, B-NT2, 675
limits, circuit relative to 1020 Hz, figure, 160
of OAM, 668
of signal or loss of frame ATM., 672
lossless digital exchange, 140
lost and misinserted cells, 622
lost cell outcome, 648
lost frames, 60
low-latency system, 601
lower-order:
virtual container, 587
-n, 587
lowest:
analog video carrier level, downstream, 695
root path cost, 347
LRC longitudinal redundancy checking.,
6162
LSAP layer service access point., 209
LSDU link service data unit., 209, 210
LSDV, 287
LSIrVLSI, VHSIC, 140
LT to NT1 2B1Q superframe technique, figure,
500
M12 multiplex, 137

M and D bits, 380
M-ary:
modulation, 177
techniques, 593
waveforms, 161
M bit, 279, 379
m = n array, 144
M-plane, ATM, 606
MAC medium access control., 39, 206
address, WLAN, 323324
architecture, WLAN, 323
figure, 324
bridge, 338
relays, 339
byte-to-serial upstream convention, 707
client:
data, 225
octets, 226
protocol, 227
sublayer, 236
controllers, 246
data service, 326
destination addresses, pertinent, 336
forwarder, 698699
figure, 699
frame format, 299
figure, 225
frame information field structure, figure, 300
frames, 297
header, 325
-level acknowledgements, 328
overview, DOCSIS, 701
packet delimiting, 279
rPLS service definition, 254
recovery information, 312
SAP address, 211
SDUs, 312
serial bit stream, figure, 277
service data unit MSDU., 318321
state machines, 326
sublayer, 206, 222, 290
DOCSIS, 700701
partitioning, figure, 224
MA FLAG set, 301
MAC control, 293
frame, 293
main cross-connect, 530
maintenance:
bits, 499
cells, 668
of data integrity, 6061
managed objects, 748
INDEX
management information base MIB., 740,

741742
manager-to-manager communications, 747
Manchester:
code, 570
coding, 225
figure, 242, 302
encode, 247
encoding, 241
manual conditioning, 164
many-to-one correspondence, 23
map ATM cells into DS3 format, 659
MAP multi-access protocol., 709
mapper, 194
mapping:
ATM cells into:
155.520 STM-1 frame, figure, 664
622.080 Mbps frame structure, figure, 665
E1 directly, figure, 663
SDH, 665667
SONET, 666
figure, 666
STM-4, 665666
between data units, adjacent layers, figure,
18
frame, 183
in a data frame, 185
switching, 184185
V.34, 184
parameters, 192
process, 575
marginal transmission channels, 67
mark:
was one, 89
and space terms, 84
-space convention, table, 8485
-to-space transition s., 93, 134
marking bias, figure, 93
marking condition s binary 1, 112
mask number, 358
material dispersion, 559560, 562, 566
matrix employing VRC, LRC together, figure,
62
MAU:
collision detection function requirements,
248249
receive function requirements, 247248
transmit function requirements, 246247
max age, 347
maximum:
bridge:
diameter, 348
transit delay, 349
803
channel width, table, DOCSIS, 706

frame mod information field size, 471
info field, frame relay, 439
latency, 314
length:
datagram, 393
fragment, 286
range, commanded nominal receive power,
DOCSIS, table, 714
transmitter output power, WLAN, 333
window size, 46
MaxValidFrame, 336
100-Mbps implementations, parameters, table,
238
100-Mbps multiple-collision-domain topology,
multiport bridge, figure, 272
MC-ASSIGN primitive, 444
MCNS multimedia cable network system., 692
MC-REMOVE request primitive, 445
MDF main distribution frame., 686
MDI medium dependent interface ., 243, 275,
279
mean cell transfer delay, 647, 649
mean time:
between failures MTBF., 730734
to repair MTTR., 730734
measurement interval, 456
T, 455, 476
mechanisms to provide OAM flows, 468
media independent interface MII., 235,
254255
medium:
access control MAC., 39, 206, 223232
token passing schemes, 289335
access protocol, WLAN, 327
attachment unit MAU., 10BASE5, 244
dependent interface MDI., 235, 258
interface connector MIC., 305
meridional rays, 556
occupation limitations, 325
reservation information, 326
message:
age, 350
increment overestimate, 349
timer value, 351
for call connection control of SVC, 468
for circuit-mode connection control, table,
516
definition, 460463
formats, 399400
mode, 623
service, 628
organization example, 519
804
INDEX
message Continued.
specific information elements, 462
in status enquiry message, table, 461
transfer part of CCITT SS No. 7, 489
type, 470
metallic cable installations, 528
metallic raceways, 543
metasignaling, 641642
cells, 619
network functions, 642
procedures, 640
protocol, 642
requirements, 642
VCI and VPI, 642
user access:
functions, 641
signaling configuration, 641642
metric IP routing table, 397
MIB management information base., 645
tree, 742
microwave:
LOS typical Installation, 593596
rmillimeter-wave links, 595
terminal block diagram, 594
MID message ID., 602, 604
value, 626
mid-bit transition, 302
MII:
attaches to any one, 255
cables for 100BASE-T, 272
clocking, 260
connector, 260
frame:
format, figure, 260
structure, 260261
location, protocol stack, figure, 259
nibble stream, 261
minFrameSize, 228
minimum acceptable throughput, 476
mini-slot stream in the upstream, 701
misalignment due to slip, 148
misinserted cells, 650
outcome, 648
mismatch in hybrid, 159
MIT management information tree., 748
MLP multilink procedure., 372
MMPDU, 331
model of operation, 802.1, 340341
modems., 9596, 155
mode-setting commands, 4951
modes of:
operation used with HDLC, 43
service, AAL-3r4, 623
modulation:
and coding, 569571

formats and bursts, DOCSIS, 704
generic types of, 165
modulo 2 division, 510
modulo-8, 370
and 128 operation, 374
sequence numbering, 45
modulo 128 operation, 371
modulo 32768, 371
operation, 374
modulus ;
-2 addition, 194
encoder, 192
minus ONE 1., 46, 215
of the sequence number, 46
monitor bit M., 296, 297
monomode fiber, 562
more data mark, the M bit, 379380
more fragments, 319
more segments, CLNP, 418
Morse code operation, 96
MPDU, 331
MPEG Motion Picture Experts Group., 622
downstream, 701
-2 frame video, 717
MSAP MAC service access point., 699
MSDU MAC service data unit., 325, 318321
MSVC meta-signaling virtual channel .
bandwidth, 642
MTBF mean time between failures ., 730734
MTTR mean time to repair., 730734
mu-law North American companding, 130
multicast group address, 320
multicasting, 396
multicollision-domain:
topology using multiport bridge, 273
networks, 283285
multiframe structure, 503
table, 504
multilevel network, network management
system, figure, 735
multilink:
control field format, figure, 373
frame formats, 372373
figure, 373
functional organization, figure, 372
multimode:
distortion, 559
pulse broadening, intersymbol
interference, figure, 558
fiber, 539, 540, 562
graded-index, 562
step-index, 562
multiple data rate, multi-collision-domain
INDEX
topology, figure, 284

multiple frame operation, 509
multiple port switch, 142
multiplex higher-order comparison, table, 137
multiplexrdemultiplex timing, 138
multiplexing:
function, 31
identification MID. field, 627
mechanism TCP., 403404
method directly from:
container 1 SDH., figure, 585
container 3 using AU-3, figure, 586
primary and auxiliary channel bits, 185186
multipoint circuit, figure, 364
multipoint connectivity, 3
multiport bit rate capacity, 273
multiport bridge, 237, 272
figure, 284
multirate-capable PHYs, 328
MUX, 198
N391 polling cycle, 465
Nq1 entitity, 33
figure, 16
Nq1 layer, 14, 15, 16
Ny1 layer, 14
Ny1 service data unit, 18
N-connection endpoint identifier, 33
N-entity, 15, 17
N-interface data, 17
unit, 17, 18
N-layer, 1417
N-protocol:
data unit, 18
between N-entities, figure, 17
NR. transmitter receive sequence number.,
213
-1, 47
count, 47
less than the NS., 53
receive sequence number, 219
sequence number, 214
NS. transmitter send sequence number., 213
and NR., 213
sequence number, 215
values, 48
N-service data unit, 17
NA numerical aperture ., refractive index,
557557
NACK negative acknowledgement., 64, 78
narrow emission line, 560
narrow positive clock pulses, 136
narrowband ingress, 704
NASP numbering plan., 644
805
national transit:
network, 491
rtandem connection element model, ISDN,
figure, 492
NAV state, 328
NDC national destination code., 645
NDF, set to 1001, 588
NDM normal disconnected mode., 54
nearest neighbors, 178
negative:
polarity voltage, 81
ranging offset, DOCSIS, 710
threshold, 136
negotiated parameters frame relay., 456
negotiated throughput, 452
NEMA bolt hole:
size, 544
sizing and spacing, 543
network:
200 m, 262
access:
overview, 67
provider, 685
addresses, 25
allocation vector NAV., 324, 327
applications, 733
capacity, 729
clock, 730
component delays, transmission system
model 2, 274
table, 287
congestion, 402
use, CLLM, 452
connection, 25
multiplexing, 515
defect, 105
-dependent value, 457
design considerations, initial, 7
hardware, 733
layer, 2426
DOCSIS, 700
layer 3, 37
management, 25
protocol identifier, 416
loops, example condition for, figure, 699
management:
application entity, figure, 752
breakout by tasks, 728729
definition of, 727
for enterprise networks, 727755
functions of upper OSI layers, 751755
provisioning, aids in, 735739
protocols, 739749
to network information transfer, 638
806
INDEX
network Continued.
performance issues, 656
procedures for setting BECN bit, 455
protocol software, 733
recovery, 294
resource management NRM., 650
responsibilities, 459460
server, 5, 6, 359
-service data units, 25
service provider, 685
signaling protocols, 29
synchronous information payload for virtual
container, 587
topology, 36, 7
traffic, observation of, figure, 341
-transport-layer boundary, 24
usage data, 729
utilization, 733
networking with VSATs, 424429
neutral and polar waveforms, 8283
waveforms compared, figure, 83
neutral ersus bipolar bit streams, figure, 134
new bit, 467
next NEXT.:
header, 412
hop, 398
router, 397
loss, 539
for connecting hardware, 537
sequence number expected, 406
nibble at a time, 260
nibble structure of multiframe, 660
NNI:
network-network interface, 447
network-node interface, 619
NOC network operations center., 740
nodal transmit antenna, 721
noise, 97
fiber-optic receivers, 567568
spike, 81
thresholds, 101
noise burst, 292, 293
nominal burst timing, figure, DOCSIS, 712
nonassured data transfer, 628
noncompliant connections ATM., 656
nondata:
J, 301
K, 301
symbol, 298
J and K, 303
nondefault QoS, 410, 413
noninformation bits, 138
nonlinear characteristics noise., 101
nonlinear encoding, 179
nonprinting characters, 76
nonpublic data network, 203
non-return-to-zero NRZ., 83
nonuniform attenuation, 98
normal response mode NRM., 43
normalized frequencies, V.34, 181, 182
North American PCM:
hierarchy and multiplexing plan, figure, 137
system, 128
notation:
conventions 8Br10B code, 282283
OSI model, 14
notes on:
bit positions:
groups, BRI frame direction:
NT to TE, table, 498
TE to NT, table, 497
crosstalk interference, 536
NRM normal response mode., 43
NRZ binary bipolar, 570
NRZ bit streams, 134
NRZ code bits, 312
NRZ code bits, FDDI, 311
NRZ non-return-to-zero ., 83
NRZ waveform, 149
NRZI code bits, 312
NSDU network service data unit., 418
NT network termination.:
derives its timing, 505
interface, ISDN, 494
to LT 2B1Q superframe, overhead, figure,
501
NT1 network termination 1., 486
in Europe, 497
in figure, 492
NT2 network termination 2., 487
NTSC National Television System
Committee., 690
null DSAP address, 218
null function, 327
numbered supervisory functions, 219
numbering of packets, frames, blocks, 10
numbering plan identification NPI., 382, 383,
479
numerical aperture, 557, 558
Nyquist, H., 117
Nyquist:
2-bit rule, 117
intersymbol interference, 117
rate, 133
rNyquist bandwidth, 117
sampling:
rate, 191
theorem, 124
INDEX
OAM operation, administration,

maintenance .:
cells., 619
extracted, 669
before leaving span, 670
inserted, 654
overhead, 666
type, 678
flows, 667
functions., 667
ATM layer, 671677
cell-based physical Layer, 674
physical layer, 670
SDH-Based Physical Layer, 672
system management, 670
supported solely by flows F1 to F3, 671
type, 678
VCC F5 Flows., 677
information, 616
levels in the B-ISDN, 667668
decreasing order, table, 667
type identifiers, table, 677
OCF IFS PIFS., 329
OC-N line rates, current allowable, table, 572
OC-N regenerator, 583
octet, 6, 9
53- cell, 602
-aligned, 371
boundary alignment, 229
rnibble transmit, receive order, figure, 261
offset, 422
odd bit differential coding, table, 195
odd parity, 61
offered load, 456
offered rate, 454
offset, 393
150-ohm transmission performance
requirements, 538539
100-ohm UTP cabling systems, 533535
one bit per Hz, 95
one-to-one pairing, 57
on-hook, 69
onroff concept, 74
open:
control flag, 406
shortest path first OSPF., 399
specification, 13
systems, 13
interconnection OSI., 9
opening and closing flags, 438
operable conditions, ATM, 654
operating wavelength, 564
operation of a hybrid, 157159
operational aspects, token bus, 292293
807
optical fiber, 552

advantages, 553554
attenuation ersus wavelength, figure, 561
cabling systems, 539540
composition, 555556
link, overview, 554555
segment, 1000BASE-LX, 285
as a waveguide, 555
optical line:
coding and transmission rate, 571
rate versus electrical rate, 571
optical power ersus drive current, figure, 565
optimizing bandwidth usage VSAT., 428429
options:
part, CLNP, 419420
TCP, 410
order, 319
order of bit transmission, 52, 228, 440
ordered delivery, 10
origins of cell delay variation, figure, 655
OSF operations system function., TMN, 751
OSI open systems interconnection ., 9
application environment, 748
connectionless service, 448
discussion of layers 14, 3439
data-link layer consists of, 223
environment OSIE., 18
and frame relay, 434
layer1, the physical layer, 3435
layer 3 signaling functions, 460
model, 9
data link layer, 1339
network management system, 749
reference model, 33
and ISDN, 494
and token bus, 291
layers, 224
to LAN model, figure, 295
with protocols, figure, 19
systems management architecture, 749
OSP outside plant., 681
cabling infrastructure defined, 546
topology, 546547
OSPF open shortest path first., 399
protocol, 392
other methods to breach the last mile, 724725
other station modes, 54
OUI organizationally unique identifier ., 357
outbound VSAT., 428
outdoor telecommunication pedestal, 546
outgoing:
bit stream, 142
and incoming Be magnitude, 478
time slot, 140
808
INDEX
outlet:
connectors, 529
ports, 142
time stage, 145
out-of-sequence frames, 60
output order, 142
overall loss, 561
overflow or underflow, 147, 148
overhead:
amortize, 92
bits, ISDN BRI, 495
information, 71
versus revenue-bearing traffic, 739
overview of WEP, 323
p and t data, 494

P bit, 47
binary number, 185
set to 1, 216
P-NNI Private network-network interface .,
643
PABX private automatic branch exchange.,
155
telephone system, 730
packet:
assemblerrdisassembler PAD., 910
data, 367385
communications, 373, 523
format and timing diagram, figure, 249
rframe relationship, X.25, 370
-handling PH.:
function, 522, 523
figure, 490
port, 522
layer protocol, 370
loss rate, 694
network, data communications, 367385
networks, 363
sequence numbering scheme, 377
service, FDDI, 309
-switch offering, 488
-switched services, 430
virtual connection based, 369385
switching, 67, 68, 516, 517
facilities, 488
network, 6869
-type identifier, figure, 376
PAD packet assemblerrdisassembler ., 9, 10
field, 227, 637
length, 635
padding, 410
to 32-bit boundaries, 624
CLNP, 420
PAD field, 626, 630
PAM wave, 126

of sampling simple sinusoid, figure, 124
formation, figure, 125
PAR mechanism, 403
parallel-to-serial conversion, 85
parameter s.:
alpha, table, 182
beta and tau, V.34, Table, 182
field, 217
length, CLNP, 420
permitted CLNP., 420421
value field, CLNP, 420
parameter values, 349
parity:
bit, 61, 91
checks, 616
parsing input sign bits, shaping frames, table,
195
passive:
open, 405, 406
state, 83
path:
cost, 348
parameter values, 349
delay, 292
per head, 573
layer, SONET, 583
overhead, 583
lookup tables, 633
overhead, 581
setting-up function, 522
terminations, 573
pathway s. existence, 57, 5859
payload, 573
capacity, 575
data unit, 36
demapping, 575
pointer, 573
type PT. field, 610
PCF point coordination function., 324
access method, 332
PCI protocol control information., 622
PCM pulse code modulation.:
bit stream, figure, 125
channel banks, 123, 149
code:
sets, 192
sign bits, 194
signs, 197
codewords, V.90 table, 189190
frame duration, 140
line rate, 150
multiplex:
comparison, table, 137
hierarchy, 136, 137
INDEX
multiplexing, 131
sign bits, 196
switching, 139146
advantages and issues, 139140
symbol magnitudes, 197
time slots., 107, 144
word, 130
PCR peak cell rate., 657
PCS physical coding sublayer.:
carrier sense, 262
deserializes code bits received from the
PMA, 271
encapsulation, figure, 269
error sense, figure, 262
reference diagram, 270, 283
transmit function, 262
PDH plesiochronous digital hierarchy ., 137
transmission systems, ATM, 668
PDU protocol data unit.: 17, 221222,
297
checksum CLNP., 418
lifetime, 417
segment length CLNP., 418
-type field values, table, 745
PDV path delay value., 286, 288
peak cell rate, 653, 656, 657, 659, 666
peak responsivity, 566
peak traffic conditions, 68
peers., 33
entities, 15
-level communications, 515
protocol data exchanges, 208
percent efficiency value, 92
performance:
802.1d, 348349
management, 729
monitoring, 671
period of a bit, 161
periodic polling, 465
periodic switching pattern SWP., V.34, 184,
185
perishability, 364
permanent virtual circuit PVC., 7, 374, 379,
380, 437
PrF bit, 213
bit set to 1, 214
frame, 53
PH-connection, 444
PH-DATA request, 496
phase distortion, 99100, 160161
Nyquist, 117
delay distortion, 97
phase encoding of V.29 modem, table,
69
phase-locked loop design, 303
809
.phase modulation PM., 96

phase shift keying PSK., 166
philosophy, data signaling ersus telephone
signaling, 6971
PHY:
clocking requirements, FDDI, 312
layer, 254
serial stream of symbols, 301302
sublayer, 213
physical coding sublayer PCS.:
100BASE-X, 264271
1000BASE-X, 277279, 280, 281
physical connections., 30
activation and deactivation, 31
endpoints, 30
connections, 30
physical layer, 2932, 616617
DOCSIS, 701
characteristics, 35
entity PHY. sublayer, 252
management, 32
mechanisms, ATM, 668
signaling, 225, 236
systems, 255256
SONET, 581
physical media for OSI, 19
physical medium PM.:
attachment PMA., 230, 264
sublayer:
100BASE-X, 264271
1000BASE-X, 277279
sublayer, 616
physical service:
access point, 30
data units, 30
transmission, 31
physical signaling PLSrAUI, selected
discussion, 240249
physical size, network, 271272
PIFS PCF interframe space., 328
PIN photodiodes, 566567
placement, TMGB, 545
PLCP frame, DS3, figure, 660
plesiochronous digital hierarchy PDH., 137,
552
plesiochronous operation, 148
PL-OAM physical layer OAM., 668
cell recognition, 675
PLS physical layer signaling.:
performs, 229
service definition, 257
plug-in module technology, 568
PM-dependent functions, 616
PMA physical medium attachment .:
code bit stream, 265
810
INDEX
PMA Continued.
interface, 270
service interface, 262, 279, 283
specification, 264
requirements, 306
PMD physical medium dependent .:
includes, 702
sublayers., 279
overview, 701702
1-point CDV, 649
2-point CDV, 649
32-point signal constellation, V.32 modem,
figure, 173
point coordination function PCF., 324
point-to-point:
connectivity, 3, 7
data connectivity, 107108
signaling virtual channel, 641
SVCs of signaling entity, 640
polar waveforms, 8283
polarity, 570
polarization isolation, 596, 719
isolation, LMDS, 723
policing actions, 455
policing a frame relay network, 455458
poll:
poll-based approach, SNMP, 743
command, 214
frame, 47
rfinal bit, 47, 213,
PrF bit, 215
polling, 366
and contention access schemes, 428
master, 324
verification timer T392, 465
port identifier, 346
port:
parameter, 348
state information, 340
states, 347
figure, 348
portal, 318
ports, TCP identifiers, 403
positive:
clock pulses, 136
constellation points, 194
-going voltage is a 0, 85
-intrinsic-negative PIN., 566
justification operation, 588
polarity voltage, 81
portion, segmented approximation, mu-law,
figure, 132
pulse, 80
possible sublayers, 1617

Postal Service, U.S., 59
postamble BEOFD, 247
post-dial delay, 494
post-equalization, 164
POTS plain old telephone service., 67, 621,
683
lines, 717
operation, 692
power sum crosstalk energy, 536
practical signaling considerations, ATM,
643646
preamble PREAMBLE., 241, 261
CSMArCD, 240
erosion, 315
field, 225
prepend, DOCSIS, 708
preassigned VP, VCI, PT and CLP values,
UNI, table, 608609
precedence:
bits, 412
IP, 392
precoding, 179
pre-emphasis, 199
prefix routing, 410411
presentation:
entities, 22
layer, 2022
services, 20
PRESYNC state, 614
PRI primary rate interface . ISDN., 484
primary:
channel rates, V.34, 180
rate B-channel interface structures, 484
service 23BqDr30BqD, 493
station, 42
tool, detect transmission errors, 37
prime power;
generation, 597
on satellite, 597
primitive, 495
descriptions, 34
primitives, 34, 208
associated with Layer 1, ISDN, table,
496
associated with the data-link layer,
table, 513
between layer 1, other entities, 495
processing of, 515
print server, 5, 560
priority:
bits, 297
control, 651
private:
INDEX
data networks, 424

-line geosatellite network, 59
ISDN access connection element, figure, 493
network, 152
probability of error, 103
process port not active, 401
proliferation and loops, 337338
propagating topology information, 346347
propagation:
delay, 539
on any pair, 536
and LMDS, 721
time delay, 251
protocol, 394
application, 415
data unit PDU., 17, 312313
discriminator, 468, 517518, 519
identifier, 350
octet, 373
layers, relationship of TCP, other layered
protocols, figure, 402
stack on DOCSIS RF interface, figure, 697
structure, connectionless data service, figure,
634
timer expiry, 348
version, WLAN, 318
identifier, 351
provisioned circuits, 446
proxy ARP, 358
service, 699
pseudo-header, 414
pseudo-noise, 599
pseudorandom binary sequence PRBS., 738
pseudoternary code is used with 100% pulse
width, 498
PSK phase shift keying., 166
PSPDN packet-switched public data network.,
520, 521
PS-POLL control frames, 327
PSTN public switched telecommunication
network., 97
PT payload type identifier ., 610
coding, table, 610
field, 609
public packet data switching center, 373
public switched telecommunication network
PSTN., 121123
pulse amplitude modulation PAM. wave, 124
pulse code modulation PCM., 121
pumping principle, 569
push service, 404
PVC permanent virtual circuit., 7, 69, 437,
446
active, 467
811
asynchronous status report, 463

exists, 69
management frame format, 460
figure, 461
management procedures, 460468
status, 463
with 2-octet address, figure, 464
and SVCs, 446, 643
physical connections, 25
Q.2931, 645
Q adapter function, TMN, 751
QAM quadrature amplitude modulation., 168,
178, 595
schemes, 177
16-QAM:
differential-coded symbol mapping, figure,
705
figure, 168
Gray coded symbol mapping, figure, 705
signal constellation, 168
signals, 694
QoS quality of service., 392, 422
classes, 652
objectives of a compliant connection, 657
parameters, 392
requirements, ATM, 615
QPSK modulation, figure, 167
QPSK symbol mapping, figure, 705
quadrature:
amplitude modulation QAM., 168, 178
modulation, diagram, 167
phase-shift keying QPSK., 598
qualifier Q. bit, 377
qualifty of service QoS.:
ATM, 647650
guaranteed, 603
parameters, 25, 27, 30, 31
for frame relay, 458459
on a transport connection, 23
quantization, 125126
and encoder, 125
noise, 126
and resulting coding, figure, 127
quantizing:
coding, European E1 PCM system, figure,
129
distortion, 126, 128
steps, 132
quantum:
level, 131
noise, 567, 568
step, 128
812
INDEX
quarantine service, 23
quasi-omnidirectional antennas, 598
quaternary:
elements, 501
symbols every 1.5 ms, 499
queue regime, 367
Radia Perlman, 336, 631
comments, 623, 624
radio:
distance to horizon, 593
frequency spectrum as seen by a TV set,
figure, 688
horizon, 592
line-of-sight ersus optical, figure, 593
rainfall:
fading, 427
study mandatory, 721
ramp:
down time, DOCSIS, 710
time, DOCSIS, 710
random access, 6
ersus controlled access, 7
rcontention method, MAC, 289
random backoff, 330
algorithm, 332
procedure, 326
random errors, 103
ranging offset, DOCSIS, 709
rate, 571
adaption is required, 149
-based control, 453
enforcer functions, 456
sequence exchange, 178
of operation MII., 260
ray beamwidths, 595
Rayleigh scattering, 561
RD request disconnect ., 51
RD routing domain., 644
reach of the system, 683
ready state, 381
real:
circuit, 57
connection, 69
time service, 410, 413
reason for discard:
CLNP, 423
parameter values, table, 423
reassembly, 9
of fragments, 391
receive:
function, MAU, 246
Manchester-encoded, 246
not ready RNR., 47, 220

ready RR., 220
packets, 379
sequence:
counter, 465
number:
NR., 46
PR., 377, 378
received, 466
state variable VR., 46
receiver:
address RA., 322, 323
field, 321
bit stream synchronized, network clock, 502
in exact synchronism, 89
fiber-optics, 566567
reception without contention, 229
recognized cable media, selection, 54
reconciliation sublayers RS., 259
gigabit media-independent interface
GMII., 276277
media independent-interface MII., 254261
recording of route CLNP ., 422
redundant no-break power, 731
Reed-Solomon generator polynomial,
DOCSIS, 707
reference:
configuration:
for connectionless data service, figure, 634
of public ISDN connection, figure, 492
for traffic control ATM., figure, 651
flag, 519
model, 13, 14
phase, 96
points T, S and R, 487
reflection, refraction, critical angle, figure, 556
refractive index, 557, 559
profile, 561
variation, 592
regenerative repeater s., 135, 136, 554555
regenerators.:
binary digital, figure, 82
digital, 147
section, 672
level, F1, 668
register, digital storage device, 85
REJ:
command response, 54
frame, 48
PDU, 220
REJECT frame, 448
reject recovery, 54, 57
relationship:
INDEX
of 1000BASE-X and PMD, figure, 280

among data units figure., 17
of MII and GMII, 260
among parameters, 456457
figure, 457
between VCrVP and transmission path,
figure, 637
relative envelope delay, 99
relay:
bearer connection, 444
open systems, 18
relaying:
decisions, 338
MAC frames, figure, 341
release of data-link connections, 514
remote:
bridges, 354
bridging concept, figure, 354
operations:
protocol, TMN, 754
service element ROSE., 753
terminal unit RTU., 683
repeated collisions, 230
repeater s.:
100BASE-T, 256
bridges, routers and hubs, 335
component delays, 272
delay, 286, 288
extends LAN range, 335
with internal crossover, 263
fiber-optic, 568
diagram, 568
set, 255, 256, 257
CSMArCD, 253
spacing, 553, 568
report type information, 462
request REQUEST., primitive, 34, 209, 507
request with response, 292
require real-time failure information, 670
required margin above noise,interference, 696
requirements:
802.1d, 349
interchange of data, 5758
signaling virtual channels, 640641
rerouting of signaling messages, 514
reservation bits, 297
reserved bits, 301
reset, 25, 27, 28, 29
resets connection identifiers, next link, 619
resistance-capacitance balancing network, 158
resource congestion, 452
response:
primitive, 34, 209
813
frame, 47
PDU, 215, 216
window, 293
to unnumbered commands, 5152
restart, ISDN, 516
restart packets, 375
retard phase, 166
retransmission interval, TCP, 403
retry, 319
return:
channel, 65
loss:
connector, 537
structural, horizontal UTP cable, worst
pair, table, 535
to repeat TRR., 304
-to-zero, 83
revenue-carrying data payload, 377
reverse address resolution protocol RARP.,
398
reversing the sense, 84
RF radio frequency.:
bandwidth wasteful, 595
channel assumptions, 694696
collision enforcement, 246
RFC request for comment. 391, 792, 400, 401
1819, 410
791 for IP4 specification, 391
RFI-NSI forwarding process, figure, 699
RI routing information., 352
field, 352353
right to transmit token bus., 292
ring:
latency, 304305
maintenance functions, 289
network, 4
transmission, FDDI, 310311
RIP routing information protocol., 399
builds table, 399
risetime, fast, 566
RNR receive not ready. frame, 47
root:
bridge, 345
designated bridge, 345
identifier, 350, 351
mean square rms. signal-to-peak impulse
amplitude ratio, 102
path cost, 346, 347
port, 338, 347, 348, 354
ROSE remote operations service element.,
TMN, 751
round-trip:
collision delay, 274, 285286
814
INDEX
round-trip Continued.
delay, 159, 425
route:
age, 398
designators, 353
routers., 356358
functions as gateway, 37
or gateway, 391
and hubs, 335
routing:
algorithms conditional statement, 358
based on cost, 399
function, 25, 397
of virtual channels, 638
information protocol RIP., 399
with IP, 398
information RI. field, FDDI, 313
loops, endless, 339
mask, 398
parameter, 346
prefix, 411
protocol, 398
and relaying, 2526, 27, 29, 515
tables, 394, 411
RR receive ready., 47
RTS request to send., 319
rCTS:
exchange, 326
frames, 327
mechanism, 327
frames, 324
figure, 322
RXD:
and GXD data bundles, 260
7:0 bundle of 8 data signals, 277
RX ER, 268
RZ return-to-zero ., 83
1s density, 175
S consecutive frames, 454
S format control field, 45
S frame supervisory frame., 45
S-format, 214
S and T functions, BRI consist of, 495
SA field, 301
SABM set asynchronous balanced mode., 49
rSABME command, 509
sampling:
rate, 133
sequence, 131
the voice channel, 124
SAP service access point., 16, 33, 34, 206, 208
identifier SAPI. 442, 512
-TEI, 512
SAR segmentation and reassembly., 9, 617

-PDU, 627
format for:
AAL-1, figure, 621
AAL-2, 622
AAL-3r4, figure, 625
AAL-5, 631
header, 628
payload, 622
field, 627
structure and coding, 631
sublayer, 620
user information, 622
satellite:
access techniques, 598599
capacity, 596
communication payload, figure, 596
communications, 596600
connectivity, uplinks, downlinks, figure, 597
orbits, 597598
SBS selective broadcast signaling., 641
scalable interleaving to support low latency,
DOCSIS, 714
scanning sequence, 131
scenario, TCP, 404405
scheduling quantum, 231
scope of meta-signaling, 642
SCR sustained cell rate. 657
and burst tolerance, 657, 659
scramble NRZ data, 246
scrambler, 199
DOCSIS, 707
V.34, 183
block diagram, 708
seed value, 708
scrambling:
and its rationale, 174175
period, 174
SDH synchronous digital hierarchy ., 136, 552,
583591, 602
-based interface, 614
transmission systems, 668
bit rates, table, 584
multiplexing:
method, container 1 using AU-3, figure,
586
structure, figure, 584, 591
standard bit rates, 584
rSONET envelope, 618
SDLC synchronous data-link control., 41
loop operation, 55
variations with HDLC, 55
SDP severely disturbed period., 105
SDU service data unit., 18, 312
INDEX
second differential encoding, 195, 196

secondary station, 42
second-order products, 101, 102
section:
definition, 583
error monitoring, 581
layer, SONET, 581582
overhead SOH., SDH, 584, 588
security field format, CLNP, 421422
security management, 729
15-segment approximation, logarithmic
mu-law, 131
segment:
F4 flow, 669
F5 flow, 670
offset CLNP., 419
type ST. field, 627
segmentation:
part, 419
permitted, CLNP, 417
and reassembly SAR., 9
of CPCS user PDU, figure, 631
segmented approximation of mu-law, 132
segmenting and blocking, 25
selecting cable media, 530531
selection:
of media, 529530
method, V.34, 181
selective:
ARQ, 47, 6465
broadcast signaling virtual channels, 641
selective reject recovery, 54
self-adaptive equalizer, 165
semi-permanent:
access, ISDN packet, 520
transparent network connection, 523
send:
sequence:
counter, 465
using modulo 256, 466
number:
NS., 46
PS., 377
received, frame relay, 463
of transmitted messages, 45
send state variable, 217
VS., 46
sending sequence number, 46
sense, signal, 35
sense of bits, 83
separate channel signaling, 493
sequence control, 28, 507
control field, 321
figure, 321
815
sequence errors, 47
sequence-length value, 449
sequence number SN., 65, 409
field, 321, 627
PR. and PS., 377
processing, 623
sequence numbering, 378
in HDLC, 4546
list, 46
sequenced information, 509
sequencing, 30, 31, 515
sequential write, random read, 141
serial:
bit stream, 91
format of data, 85
incoming CPCS-PDU, 631
quotient, 63
server;
intranet, 359360
managed system, 740
service:
access point SAP., 16, 206, 356, 505506,
618
identifier SAPI., 442, 512
aspects, ISDN packet service, 523525
classification for AAL, 620632
table, 621
connection-oriented, connectionless,
633637
equipment power. ground, 540
level agreement SLA., 7, 58, 471
primitives, 507
figure, 208, 507
provided to:
data-link layer, 2930
network layer, 26
serving node closer to customer, 691
SES severely errored second.:
event, 107
ratio SESR., 105
session:
address, 22
connection, 22
establishment, 23
synchronization, 22
layer, 2223
set asynchronous balanced mode SABM., 49
command, 220221
set asynchronous response mode SARM., 49
SET changes data, 742
setting:
BECN bit, 455
this bit, 441
set-top box, 717
816
INDEX
setup and release of VCCs, 640

SETUP message, 468, 470, 471, 474, 646
content, table, 469, 518
seven-layer OSI reference model, figure, 19
severe absorption, 596
severely disturbed period SDP., 105
severely errored seconds. SES., 104, 105,
733, 736
cell blocks, 649
outcome, 648
ratio, 647, 650
SFD, CSMArCD, 240
SFS start-of-frame sequence., 294
and EFS fields, 298
Shannon, 117118
limit, 118
signal-to-noise ratio, 117
shaping frame, 195, 196, 197
shell mapping, 179
shift keying, 96
shift register used with CRC-CCITT, figure,
63
Shockley, Bell Laboratories, 124
short:
burst traffic, 602
IFS SIFS., 329
retry counter, 332
shortened codewords:
capability not available, 711
last, mode, 707
shot noise, 566
shrinkage of the interpacket gap IPG.,
251
SIFS short interframe space., 322, 328
interval, 322
timing, 329
sign assignment, 198
sign inversion rule, 197
signal:
characteristics, CSMArCD, 241244
constellation, 167
for 4800 bps operation, V.32 modem,
figure, 171
and mapping, trellis-coded modulation,
figure, V.33 modem, 175
element, 112
encoding, CSMArCD, 241
impairments, 79
-to-noise ratio, 103
processing sequence, DOCSIS, 708
quality error, 248
sense, 35
space diagram, figure, V.29 modem, 170
state mappings, V.32 modem 9600 bps, table,

172
transfer points STP., 489
signaling:
activity status, PVC, 467
channel, 484
in header, 603
rate, 243244
required for SVCs, 468474
requirements, 640646
virtual channels VC., 640641
connections, 619
requirements, 641
silence, 241
SIM set initialization mode., 49
simplified:
concept, ISDN switching, figure, 488
diagram, FEC operation, 66
singing, 159
single:
cable:
frequency allocations, table, 250
system, figure, 245
-collision-domain multisegment networks,
284285
-fiber buffers, 564
-link:
and multilink procedures, 372373
procedures, 371
-mode fiber, 561, 562563
-point-failure, 425, 426
SLA service level agreement., 7, 58, 471
slips., 147, 148
objectives, North American, 148
rate performance, 148
slot, 601
slot times., 231
definition, 239
slotTime, 228, 231, 238
slot time, 292
slow-start mechanism, 453
SLP single link procedure., 372
SLRC station long retry count., 330
smart bridge, 336
routers and switching hubs, 71
SMASE systems management application
service element., TMN, 751
SMDS switched multimegabit data service.,
602
transport, 659
SMI structure of management information.,
740
smooth earth, 592
INDEX
SMT station management., 309

Snells Law, 555, 556
SNMP simple network management protocol.,
394, 645, 700, 740746
error codes, table, 745
message embedded in a local network frame,
figure, 743
overview, 740746
PDU, 743
structure for GetRequest, figure, 744
pictorial overview, figure, 742
three components, 740
trap:
codes, table, 746
PDU format, figure, 744
uses TCPrIP for communication, 739
version 2, 746747
SNP sequence number protection., 621
SNRM set normal response mode. command,
49
socket, TCP, 403
SONET synchronous optical network., 136,
552, 571, 602
network element NE., 583
payload, ATM, 666
rates and formats, 571572
rSDH frame, 604
source, 36
address field, 36
figure, 636
identifies, 299
address specific CLNP., 421
clock frequency recovery, 622
-controlled transmitters, 448
coding, 65
port, 409
UDP, 414
quench message, 400
routing, 344
bridge, 352354
CLNP, 421
header, RI Field, 352353
traffic descriptor s., 654, 659
space:
network time slot, 145
-to-mark transitions, 91
segment charges, 427
-stage time slots, 145
switch, 142144
allows time slot interchange, figure, 143
-switching stage, 144
-time-space STS. switch, figure, 146
span interleaving., 427
817
spanning tree, 344, 354

algorithm, 335, 337338, 340, 343, 344350,
355
explorer, 353
figure, 345
paths, 347
protocol, IEEE 802.1d, 698
spatial representation of data tone, BPSK, 166
SPE:
assembly process, 573574
figure, 575, 576
special conditioning, 161
specialized common carrier, 151
specific comments, OSI, 3339
specific high-speed modems, 168174
specific layers, OSI reference model,
1819
specifically routed, 353
specification, output ports, 502503
spectra, three DSL waveforms, figure, 684
spectral;
bandwidth, 532
efficiencies, 178
line widths for LED and LD, figure, 565
metric, 196, 197
shape filter, 198
shaper, 196197
shaping, 195
DOCSIS, 706
frame, 197
metric, 198
width, 559
speech memory, 141
speed tolerance, 138
speeding up networks, 433
splices, 569
splitting downstream 800-MHz bandwidth,
figure, 722
spread, symbol in time domain, DOCSIS, 711
spread spectrum, 599
SREJ selective reject., 54
command, 48
rREJ, 5354
SS7 ISDN user part ISUP., 643
SSAP source service access point., 211, 388
field, 216
SSRC STA short retry count., 330
SrT interface with reference model, 494
STA:
originating, 326
short retry count SSRC., 330
stackables hubs., 6
Stallings, William, 37
818
INDEX
standardized PTI, for end-to-end

communications, 670
standby transponder, 426
star:
backbone topology, 546
configuration, 546
network, 425
simple, 4
topology of hub, figure, 355
start:
delimiter SD., 291
element, 90
frame delimiter SFD., 241, 247
field, 225
-space, 90
-stop:
character, 89
transmission, 8991
-of-stream delimiter SSD., 269
rJrKr, 268
starting:
delimiter SD. 295, 296, 303
figure, 304
and ending delimiter patterns, figure, 304
2-state trellis diagram, 196
state, learning allowed, 342
state variables, 45, 46
static entries, 342
does not exist, 342
station:
and configurations, 42
functional organization, token ring, figure,
306
long retry count SLRC., 330
organization, token ring, 305306
-to-TCU connection, figure, 307
statistical time division multiplexing, 366367
statmultiplex concept, frame relay, 435
STATUS ENQUIRY message, 460462,
465468
STATUS message, 460, 462, 468, 471, 474
table, 462
status:
of signaling link, 465
request, 54
step-index fiber, 558
STM-1 bit rate, 593
STM-N frame structure, figure, 588
STM payload, 587
stop element, 90
stop-and-wait ARQ, 64
stop-mark, 90
store-and-forward:
switching, 67
systems, 68, 368
techniques, 355
STP shielded twisted pair., 528
-A, 532
stream of alternating 1s and 0s, 80
streaming mode, 623
service, 628
strictly ordered service class, 319
stripping, 301
structure:
cell flow, 618
CLNAP-PDU, figure, 636
virtual tributary, 577581
STS space-time-space .:
switch, 146
architecture, figure, 146
STS synchronous transport symbol.:
-1 envelope capacity, 573
-1 frame structure, 572573
figure, 572
-1 path overhead POH., 574
-1 SPE:
interior, STS-1 frame. figure, 574
-1 SPE with STS-1 POH, payload capacity,
figure, 574
-3c transport overhead assignment, figure,
577
frame boundary, 666
-N frame:
figure, 576
structure, 576577
-N signal, 583
-Nc, 576
-Nc SPE, figure, 578
-Ne SPE, 577
path overhead, 581
stuffing:
bits, 575
equals justification, 138
sublayering:
AAL, 620
OSI layer 3, internetting, figure, 38
subnet:
address field, figure, 358
masks, 358
subnetwork, 24
connections, 25
independent convergence protocol SNICP.,
415
subtree, 742
subvector 1, 300
successful cell transfer outcome, 647
successive OAM cells, 650
INDEX
superconstellation, 186, 187

figure, 187
superframe:
V.34, 183
VT, 578
supertrunks, 692
supervisory:
commandrresponse, 213
format control field bits, figure, 220
format S, 214
frames, 4748
function bit, 213
signaling, 73, 133
surface leakage noise, 567
survivability:
where network management pays off,
729730
of VSAT network, 425
sustainable cell rate SCR., 657, 666
SVC switched virtual circuit., 69. 446, 468
bandwidth, 642
requires circuit setup, 460
SVCI signaling virtual channel ., 642
switch buffer overflow, 649
switched ATM connections, 659
switched virtual circuit SVC., 69, 446
switches in tandem, 121
switching, simplified concept for ISDN, figure,
488
SWP, represented, 184
symbol:
decoder, 303
encoding, figure, 302
mapping, DOCSIS, 704705
symbol rates, 199
V.34, 180
table, 180
timing, 303
symmetric configuration, 42
symmetrical operation, 682
SYNCH state, 615
synchronization word, 499
synchronizing clock frequency, 505
synchronizing word, 133
synchronous:
data-link control SDLC., 41
digital hierarchy SDH., 139151, 552,
583591
optical network SONET., 136, 552, 571583
payload envelope SPE., figure, 573
primary channel, V.34, 178
residual time stamp SRTS., 621
transmission, 91
FDDI, 315
transport module STM., 584
819
SYNC UNITDATA.indicate primitive,

281282
syndrome equals zero, 614
system:
above 28 GHz, 592
considerations, multisegment:
100BASE-T networks, 271274
10 Mbps broadband networks, 249251
1000-Mbps networks, 283288
-defined timeout, 53
depth, network management problem,
734739
management message
components, figure, 754
figure, 755
manager, 294
model 2 single repeater, figure, 286
systematic distortion, 94
T391 expires, 465
T and S architecture, 140
TA interface, ISDN, 487
table of mark-space convention, 84, 85
tag cells, 654
tapped delay line, 163
TBB telecommunications bonding backbone.,
542
TBBIC telecommunications bonding
backbone interconnecting conductor.,
544
TC sublayer, 619
TCM trellis-coded modulation., 177178
systems, 177
TCP transmission control protocol.:
assigns sequence numbers, 403
connection, 403
defined, 402403
entity state diagram, 410, 411
header format, 408410
figure, 408
rIP internet protocol., 37, 38, 386413
and data-link layers, 387390
family of protocols, 39
IEEE 802 relating to OSI, ULP, figure,
389
relationship, data-link layer frame, figure,
388
mechanisms, 403404
receptive, 404
TCU trunk coupling unit., 305
TDM link, 428
TDMA time division multiple access., 428,
598
burst, 711
TDR time-domain reflectometer ., 734
820
INDEX
TE1rTA, NT2 not synchronized to network

clock, 502
TE2 adapts equipment, 487
TEI terminal endpoint identifier ., 512
Telecommunications.:
bonding backbone TBB., 541,
542543
building layout, major elements of, 528
closet TC., 530, 532533, 541
and equipment rooms, 546
connection, typical long-distance, figure,
157
entrance facility TEF., 540, 541, 545
grounding busbar TGB., 542, 544
description TGB, 544
main grounding busbar TMGB., 543544
management network TMN., 749755
network, data transport vehicle, 121152
TELEDESIC satellite . system, 598
ten-bit interface TBI., 235
term set, 157
terminal endpoint identifier TEI., 511, 512
server, 360
terminate OAM flow, 669
ternary signals, 571
ternary symbol transmission rate, 263
test command, 217218
TEST response, 218
testing hypothesis, 733
TFTP trivial file transfer protocol., 700
TGB telecommunications grounding busbar.,
542
theoretical arrival time TAT., 657
thermal noise, 100101, 567
third-order products, 101
three-way handshake, 404, 405
throughput, 103
time availability, 595, 723
time to charge capacitance, 95
time-division:
circuits, 58
PCM switching, 139
space switch cross-point array, figure, 143
time division multiplexing TDM., 124, 366
of data, figure, 367
time duration of time slot DS1, 141
time exceeded message, 400
time to live TTL., 394
time-sequence diagram, primitives, 210
time slots., 16, 67, 132, 140, 504
0, 133, 504
for synchronization, framing, 736
allocated, 366
assigned, 134, 598
assignment, 24 to D-channel, 503
counter, 141, 142

duration, 367, 598
interchange:
figure, 140
time switch, figure, 141
among trunks, figure, 143
interchanger TSI., 140, 144146
passes, 144
time-space-time switch TST., 144145
figure, 145
time switch, 140142
time slot interchanger:
random write, figure, 142
sequential read, figure, 142
time-switching stage, 144
timeout:
mechanism, 342
recovery, 53
value, 343
timestamp, 400
field, 746
timing:
circuit, 135
considerations, 505
ISDN BRI, 498499
discontinuities, 575
distortion, serial bit stream, 9394
error, CM and CMTS, 703
frequency, 136
jitter, 136
out of entries, 343
recovery, 243
refers to synchronization, 134
relation between source and destination, 620
and synchronization, 91
TMGB telecommunications main grounding
busbar., 542
description, 543
TMN telecommunications management
network., 749755
management message structure description,
753
relationship with network it manages, figure,
750
TOA type of address ., 382384
rNPI address format, 382384
token:
bit, 297
bus, 290
frame formats, 291292
control right to access, 289
control signal, 294
format:
FDDI, figure, 310
figure, 291
INDEX
token ring figure, 295

holding:
time, 296, 294
timer, 290, 310
passing, 6, 292293
bus, 290293
bus architecture, 291
ring, 7, 293308
configuration, figure, 293
schemes, 289335
ring, 5, 7
comparison, table, 308
frame format, 295301
figure, 296
symbol transmission characteristics,
301302
sending station, 292
time, definition, 314
tone energy, 166
tonterias, 734
topology:
applicable to transmission model 1, 252
backbone cabling, 530
changes., 343
announcement, 351
notification BPDU, 348
FDDI, figure, 309
torn-tape center, 68
total:
dispersion, 561
distortion, 94
internal reflection, 556
within fiber, figure, 557
length field, 388
traffic class:
bits, 413
IPV6, 411, 412413
traffic contract, 652, 659
violation, 610
traffic control, congestion control, ATM,
650659
traffic counter parameter specification, 659
traffic descriptors and parameters, ATM,
653
traffic flows, 410
traffic loops, disastrous, 355
traffic management functions, 619
traffic parameters, 653
traffic profile of a public switched telephone
network, figure, 445
traffic shaping, policing, 619
trailing control indicators, 313
trailing flag, 92
trailing signal element, 302
821
training period for equalizers, 165

transit delay, 31
transit network selection, 646
transition density to facilitate clock recovery,
266
transition from off to on, 113
transmission:
characteristics, table, 534
control protocol TCP., 3738, 402410
Internet protocol TCPrIP., 386413
convergence TC. sublayer, 616
functions, 617618
data over:
analog voice channel, 155200
longer distances, 9596
digital data, 94107
downstream, 694
levels, 694695
loss with frequency, 98
models, maximum network size, 272
order, 283
over optical fiber, 555
path:
example with three repeater sets, figure,
256
level, F3, 668
maximum, figure 254
with four repeater sets, figure, 257
with three repeater sets, figure, 255
performance requirements, 535537
150-ohm STP, 538539
rate, 31
reduction, 455
system:
model 1, 251252
model 1 and model 2, 284, 285286
upstream, 694
windows, 561
transmit function, MAU, 246
transmit spectra templates, V.34, figure, 181,
182
transmitter address TA. field, 321
transparency, 440
ISDN, 511
transparent, 44
bridge, 336
frames, 354
transponder, 596
bandwidth, 598
transport address es., 22, 23
transport layer, 2324
layer 4, 38
transport:
-connection-establishment, release, 24
822
INDEX
transport Continued.
overhead, 573
channels, 576
protocols, 23
of virtual containers, 591
transporting ATM cells, 659667
transversal equalizer, 163
transversal filter, figure, 164
traps., 742
format, 743
manager, 747
transmit function requirements, CSMArCD,
figure, 247
traveling-wave tube TWT., 595
tree network extension of bus network, 3, 4
trellis:
-coded modulation TCM., 174, 177178
coding:
all data rates, 178
V.32 modem, 170
diagram, figure, 197
encoding, 179
tributary:
information, 580
signal, 573
unit group TUG., 587
unit-n, TU-n., 587
trivial file transfer, 415
troubleshooting, 732734
truncated binary exponential backoff, 231232
trunk coupling units, figure, 306
TSI time slot interchanger ., 144
TST switches., 140 144, 145
TTL time to live., 394
on routers to prevent endless loops, 394
TUBA, see CLNP, 415
TU-n pointer, 589
TV signal distribution, CATV system, 688690
twist and sway maximum, 595
2BqD, 483, 484, 485
configuration, figure, 496
customer data, 500
frames, 499
23BqD, 484, 485
30BqD, 484, 485
2B1Q, 683
coding, 499
waveform, 501
2D symbol intervals, 183
two:
contiguous bits, 74
network management protocols, 739740
-stage multiplex, figure, 590
-state information, 73
-way:
data transfer, 405
two examples of, figure, 406
digital service, 690
window turns, 453
-wire transmission, 156
ersus four-wire operation, 156158
TX EN transmit enable., 260
TX ER signal, 268
type 1:
operation commands, responses, 216217
operation parameters, 214
operation response, 218
control field bit assignments, figure, 218
type 2:
operation, 211
and parameters, 214215
types.:
code CLNP., 418, 422
field, 394
and instance, 15
optical fiber, 562563
of service, 392
and subtype field, 318
timing distortion, figure, 94
typical differential delay across voice channel,
figure, 100
typical X.25 transmissions exchange of packets,
381382
U chord, 188
U format, 214
control field, 45
U frame unnumbered frame., 45, 48
U interface, 499
ISDN, 487
U-plane, ATM, 605
UA unnumbered acknowledgement., 51, 220
Ucode, 188, 194
UDP user datagram protocol., 414415
DOCSIS, 699
datagram header, figure, 414
header, 414
-IP interface, 415
port 161, 162, 747, 745
numbers, 699
UI unit interval.:
command, 217
frame, 48, 461, 508
ULP upper layer protocol., 389, 404
connection, TCP, 403
modes, TCP, 404
INDEX
synchronization, TCP, 404

UMD timer function, receive, 248
unscrambled postamble, 247
unacknowledged:
connectionless-mode service, 209210
frame, 46, 509
operation, 508
unassigned cell versus idle cell, 619
unbalanced:
configuration, 42
digital interface circuit EIArTIA-423, 115
undefined format, 298
Ungerboeck, 177
UNI user-network interface ., 447
ATM, 605
NNI interfaces, figure, 447
ersus NNI, 468
measurement point, 647
unicast, 413
uniform power loading per unit bandwidth,
696
unique:
binary sequence, 291
field flag., 91
identification, bridge, 344
path identifier, 742
unit interval, 91, 571
United Parcel Service, 59
unnumbered:
command, response control field bit
assignments, figure, 221
format commands, responses, 220
format U, 214
information UI.:
command PDU, 217
frame U., 48, 465
poll UP., 55
unscrambled mode delimiter UMD., 246
unsigned:
binary number, 351
unsigned integer, 412
unsigned mapper outputs, 198
unused OAM cell information field
unwanted disturbances superposed, 79
UP unnumbered poll., 51
UPC usage parameter control., 610
function, 656
rNPC:
function, 656
mechanism, 654
parameters, 654
uplinks, downlinks, 597
uppercase, 76
823
upstream, 682
channel characterized, 701
CM signals, 694
concept, figure, 692
demodulator input power requirements, 711
direction, 191
features, 702
electrical output, CM, 711
frequency agility, range, DOCSIS, 706
modulator must provide, 707
PMD sublayer, 706
QPSK symbol mapping, 704
RF channel transmission characteristics,
table, 696
urgency, 5960
urgent pointer, 410
usage network parameter control UPCrNPC.,
651
usage parameter control UPC., 610
use of:
DLCIs, 443
Windows, to rate-based control, 453454
useful bits, 103
user access capabilities, ISDN packet, 524
user behavior, receipt of FECN bit, 453
user congestion in popular RF bands, 595
user-data cells, 619
user data field length, 379
user datagram protocol UDP., 414415
user information, 637
ISDN, 516
user interface, UDP, 415
user-network:
access, ISDN, 484
interface UNI., 442, 459, 643
configuration and architecture, 605606
interfaces, 487
traffic contract, 654656
-to-network signaling procedures, 640
user part, CCITT SS No. 7, 489
user plane, ATM, 605
user reaction to BECN and FECN, 448
user response to congestion frame relay., 448
user unique burst parameters, DOCSIS, table,
710
user-to-user signaling configuration, 642
user-VCs belonging to, 640
utilization factor, 368
UTP unshielded twisted pair., 528
category definitions, 533
V.24 circuit nomenclature, 110
V.29 modem, 168170
824
INDEX
V.34:
definitions, 179
mode, 198
modem, 178187
operation, 179181
V.90 modem, 187200
V.110 techniques, 150
validating, sequence number, 407
variable:
binding, 745
bit rate VBR.:
services, 603
support, 631
depth interleaver, downstream, 702
-length frames, 623
size data units, 633
variants, CSMArCD, table, 233
VBR variable bit rate., 603
scenario, 622
VC-4 container, 665
VC-4-4c:
boundary, 666
path overhead, ATM cell mapping, 665
VC virtual channel .:
cross-connect, 638
-n associated with each TU-n, 590
-switching, 609
VCC endpoints, 638
VCI virtual channel identifier ., 607
fields, 609
identifies, 637, 638
ver version., 391
vertical redundancy checking VRC., 6162
virtual call, 375, 379, 459
virtual carrier sense, 326
virtual channel, 637
connection, 640
level, 637638
virtual circuit, 58, 67, 381, 446
frame relay, 435
and logical channels, 368369
virtual connection, 67, 69
virtual container VC.:
-N VC-N., 585
POH appropriate to level, 587
virtual local area network VLAN., 369
virtual path VP., 638
rvirtual channel, 671
connection VPC. operation flow F4 flow.,
619
identifier VPI., 637
level, 638639
virtual scheduling algorithm, 657
VLAN virtual local area network., 359
tag, 359
VLSIrVHSIC, 67
voice channel, 95
defined, North American, figure, 99
impairments to data transmission, 97102
limited bandwidth, 155
traversing telecommunication network,
figure, 97
voice telephony, 531
voltage level, 125
measurement, 130
VP virtual path.:
-AIS:
alarms, 671
cell detection condition, 671
cell generation conditions, 671
cells, 671
release condition, 671
performance management functions,
676677
switchrcross-connect, 638, 639
-switching, 609
and VC switching hierarchy, 640
figure, 639
VCI, PT and CLP values at the UNI, table,
608609
VPC virtual path connection.:
continuity check, 676
endpoints, 638
-FERF, 671
alarms, 671
cell detection condition, 676
cell generation condition, 676
release condition, 676
field, 609
value equals zero, 642
and VCI values, 607
VPI virtual path identifier ., 607
rVCI:
fields routing field., 607608
value used in both directions, 640
values, 646
VRC vertical redundancy checking., 61
rLRC used in tandem, 62
VSAT very small aperture terminal., 424
network, 151, 599
description, 425
disadvantages, 425426
networking, 424429
VT :
group, 577
payload:
capacity, 578
pointer, 578
INDEX
POH byte, 580

sizes, figure, 579
superframe, figure, 580
W-index fiber, 562
WAN wide area networks., 363432
encapsulation, 354
constrained by intervening media, 204
and LAN, 1
waveforms line codes., binary digital
transmission, figure, 84
waveguide dispersion, 560
wavelength interval, 565
WDM wavelength division multiplex., 568
weak link, 597
WEP wired equivalent privacy., 319
algorithm, 323
wide area networks WAN., 363432
wideband radio or wireless systems, 683
window, 46, 409
of frames, 46
open, 46
size, 379, 454
TCP, 403
windowed protocol, 453
2-wire full duplex, ISDN, U.S., 501
wire pair:
capacitance, 93
loop, 683
noisy, 122
use of 100 Mbps CSMArCD, figure, 263
systems serving last mile, 681
transmission medium, 204
wired equivalent privacy, 319
wireless:
CATV technique, 717
LAN, 316335, 724
medium, 318
PABX, 725
wiring closet, 5
WLAN wireless local area network., 316321
MAC, 323
work station, 3, 5, 203
working window, 453, 454
worst-case:
burst timing, DOCSIS, figure, 713
825
path delay value, 286

writing incoming time slots, 142
X.21 and X.31, 370

X.25, 369385
architecture, OSI reference model, figure,
369
call:
clearing, figure, 381
request packet, 523
setup, figure, 381
compared to frame relay, table, 434
data communication equipment, 489
octet:
contained in LAPB frames, 370371
conventions, 380
packet:
embedded, LAPB frame, figure, 370
layer procedures, 522
typical, 377380
type, figure, 376
protocol, ISDN packet, 524
slower, 433
X.121:
address, 382
addressing, 357
X.301 addressing, 382
XDSL, 683
XID exchange identification ., 51
control field, 450
frames, 449
information field basic format, figure, 218
response, 217, 218
and TEST commandrresponse, 214
XMIT flag, 282
Z control bits, 297
Z overheads, 660
zero:
-bit:
extraction, 459, 460
insertion, 43, 44, 440
crossings, 136
dispersion near 1300 nm, 560561
zeros, long string, 135

Practical Data Communications Roger L Freeman

Uploaded by

Copyright:

Available Formats

Practical Data Communications Roger L Freeman

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Practical Data Communications Roger L Freeman

Uploaded by

Copyright:

Available Formats

Practical Data

For my son, Bob

Enterprise Network Environment

The OSI Model and the Data-Link Layer

Internet Protocol IP. and Gateway Layer

High-Level Data-Link Control (HDLC) Typical

Error Correction Schemes 64

Error Performance from an ITU-T

The Telecommunications Network as a Vehicle for

7.3.3 Time Switch 140

The Transmission of Data over the Analog

Scrambling and Its Rationale 174

Data Communications in the Office Environment,

9.4.2.3 Control Field Parameters 214

Typical Topologies Applicable to

Data Communications in the Office Environment,

Protocol Data Units 312

Wide Area Networks (WANs)

Economy Versus Data

ULP Synchronization 404

12.5 PVCs and SVCs 446

Interface Functions 495

Building and Campus Wiring and Cabling for

14.6 Entrance Facilities 533

Broadband Data Transport Techniques

STS-1 Frame Structure 572

Broadband ISDN (B-ISDN) and the Asynchronous

16.4.3 Idle Cells 613

Circuit Setup Using ATM

Mapping ATM Cells into SDH 665

Last-Mile Data Distribution Systems

Extending CATV Coverage Area

Network Management for Enterprise Networks

SurvivabilityWhere Network Management

PREFACE TO THE SECOND EDITION

PREFACE TO THE SECOND EDITION

PREFACE TO THE SECOND EDITION

PREFACE TO THE SECOND EDITION

1.1 ENTERPRISE NETWORKS

THE ENTERPRISE NETWORK ENVIRONMENT

Figure 1.2. An example of a WAN.

TYPES OF NETWORK TOPOLOGY

1.2 TYPES OF NETWORK TOPOLOGY

Figure 1.3. Point-to-point connectivity.

Figure 1.4. Multipoint connectivity (sometimes called multidrop).

THE ENTERPRISE NETWORK ENVIRONMENT

Figure 1.5. An example of a tree network. It is an extension of the bus network.

Figure 1.6. A ring network.

TYPES OF NETWORK TOPOLOGY

Figure 1.8. A higher-order star network. Such a network

THE ENTERPRISE NETWORK ENVIRONMENT

A hub is a centralized connecting point for a LAN. It is a multiport

1.3 NETWORK ACCESS: AN OVERVIEW

INITIAL NETWORK DESIGN CONSIDERATIONS

1.4 INITIAL NETWORK DESIGN CONSIDERATIONS

Latency r turn-around time

Low when no collisions;

THE ENTERPRISE NETWORK ENVIRONMENT

1.5 CONNECTION-ORIENTED AND CONNECTIONLESS SERVICE