Software Architecture

and the UML

Grady Booch
Architecting a dog house

Can be built by one person

Requires
Minimal modeling
Simple process
Simple tools
Architecting a house

Built most efficiently and timely by a team

Requires
Modeling
Well-defined process
Power tools
Architecting a high rise
Early architecture

Progress
- Limited knowledge of theory
Modern architecture

Progress
- Advances in materials
- Advances in analysis

Scale
- 5 times the span of the Pantheon
- 3 times the height of Cheops
Modeling a house
Movements in civil architecture
 Bronze age/Egyptian (Imhotep)
 Grecian/Roman (Vitruvius)
Progress
- Imitation of previous efforts
 Byzantine/Romanesque - Learning from failure
- Integration of other forces
- Experimentation
 Gothic
 Mannerism (Michelangelo, Palladio)
 Baroque
 Engineering/Rational/National/Romantic
 Art noveau
 Modern movement (Wright, LeCorbusier)
Neufert Architect’s Data
The Handbook of Building Types

Kinds of civil architecture

 Community
­ houses, flats and apartments, gardens, education,
hospitals, religion

 Commerce
­ shops and stores, restaurants, hotels, office
buildings, banks, airports

 Industry
­ industrial buildings, laboratories, farm buildings

 Leisure
­ sport, theaters and cinemas, museums
Forces in civil architecture

Load
Kinds of loads
- Dead loads
- Live loads
- Dynamic loads
Compression Tension Avoiding failure
- Safety factors
- Redundancy
- Equilibrium

Load

Any time you depart from established practice, make ten times the
effort, ten times the investigation. Especially on a very large project.
- LeMessuier
Brand, How Buildings Learn

Shearing layers of change

Space plan

Services
Stuff
Structure

Skin

Site
Walker Royce

Dimensions of software complexity

Higher technical complexity
- Embedded, real-time, distributed, fault-tolerant
- Custom, unprecedented, architecture reengineering
- High performance
An average software project:
- 5-10 people Defense
- 10-15 month duration Telecom Weapon System
- 3-5 external interfaces Switch
- Some unknowns & risks National Air Traffic
Commercial Control System
Embedded Compiler
Automotive
Software Large-Scale
Lower CASE Tool Organization/Entity
Simulation
Higher
management management
complexity Small Scientific complexity
- Small scale Simulation - Large scale
- Informal IS Application
Defense
- Contractual
Distributed Objects Enterprise IS
- Single stakeholder (Family of IS MIS System - Many stake holders
(Order Entry)
- “Products” Applications) - “Projects”
IS Application
GUI/RDB
(Order Entry)
Business
Spreadsheet

Lower technical complexity

- Mostly 4GL, or component-based
- Application reengineering
- Interactive performance
Forces in Software
Functionality
Cost Compatibility

Capacity Fail safe

Availability Fault tolerance

Performance Throughput

Technology churn Resilience

The challenge over the next 20 years will not be speed or cost or performance;
it will be a question of complexity.
Bill Raduchel, Chief Strategy Officer, Sun Microsystems

Our enemy is complexity, and it’s our goal to kill it.

Jan Baan
Wojtek Kozaczynski

The domain of architecting

The “what” The “why”
Architecture System
Satisfies Features
Qualities

Architecture S/W
Constrain Requirements
Architecture
Representation System
Quality Attributes

Produces Technology Defines

The “who” The “how”
Follows
Architect Process

Skills
Defines role
Organization
Stakeholders
Philippe Kruchten

We all know that ...

Architecture and design are the same thing
Architecture and infrastructure are the same thing
<my favorite technology> is the architecture
A good architecture is the work of a single architect
Architecture is flat, one blueprint is enough
Architecture is just structure
System architecture precedes software architecture
Architecture cannot be measured and validated
Architecture is a Science
Architecture is an Art
Merriam Webster’s Collegiate Dictionary
10th edition

Architecture defined (again)

Architecture n (1555) 1: the art of science of
building, specifically, the art or practice of
designing and building structures and esp.
habitable ones 2 a: formation or
construction as or as if as the result of
conscious act <the ~ of the garden> b: a
unifying or coherent form or structure <the
novel lacks ~>
Mary Shaw, CMU
Grady Booch,

Architecture defined (yet again) Philippe Kruchten,

Rich Reitman
Kurt Bittner, Rational

 Software architecture encompasses the set

of significant decisions about the
organization of a software system
­ selection of the structural elements and their
interfaces by which a system is composed
­ behavior as specified in collaborations among
those elements
­ composition of these structural and behavioral
elements into larger subsystem
­ architectural style that guides this organization
Mary Shaw, CMU
Grady Booch,

Architecture defined (continued) Philippe Kruchten,

Rich Reitman
Kurt Bittner, Rational

 Software architecture also involves

­ usage
­ functionality
­ performance
­ resilience
­ reuse
­ comprehensibility
­ economic and technology constraints and
tradeoffs
­ aesthetic concerns
Mary Shaw, CMU

Architectural style
 An architecture style defines a family of
systems in terms of a pattern of structural
organization.
 An architectural style defines
­ a vocabulary of components and connector types
­ a set of constraints on how they can be
combined
­ one or more semantic models that specify how a
system’s overall properties can be determined
from the properties of its parts
Architecture metamodel
Software Software
Architecture Architects
is part
of are actors in
System
is represented by
architecture
Architecture
Design Process

Software produces
Architecture Logical view
has Description

Process
is made of
view
relates to
Architecture is a Implemen-
Style guide tation view
Architectural Deployment
view view
has
Use case
Architectural style is made of view
has
constrains

is a Form Connection
depicts
Architectural
Pattern
Component Constraints

satisfies
Architectural
constrains Blueprint
Requirements
Models
 Models are the language of designer, in many
disciplines
 Models are representations of the system to-
be-built or as-built
 Models are vehicle for communications with
various stakeholders
 Visual models, blueprints
 Scale
 Models allow reasoning about some
characteristic of the real system
Many stakeholders, many views
 Architecture is many things to many different
interested parties
­ end-user
­ customer
­ project manager
­ system engineer
­ developer
­ architect
­ maintainer
­ other developers
 Multidimensional reality
 Multiple stakeholders
multiple views, multiple blueprints
Architectural view
 An architectural view is a simplified
description (an abstraction) of a system
from a particular perspective or vantage
point, covering particular concerns, and
omitting entities that are not relevant to this
perspective
Architecturally significant elements
 Not all design is architecture

 Main “business” classes

 Important mechanisms
 Processors and processes
 Layers and subsystems

 Architectural views = slices through models

Characteristics of a Good Architecture
 Resilient
 Simple
 Approachable
 Clear separation of concerns
 Balanced distribution of responsibilities
 Balances economic and technology
constraints
Representing System Architecture

Logical View Implementation View

End-user Programmers
Functionality Software management

Use Case View

Process View Deployment View

System integrators System engineering
Performance System topology
Scalability Delivery, installation
Throughput Communication

Conceptual Physical
Relation Between Views

Logical view Component view

Process view


Deployment view
How many views?

 Simplified models to fit the context

 Not all systems require all views:

­ Single processor: drop deployment view
­ Single process: drop process view
­ Very Small program: drop implementation view

 Adding views:
­ Data view, security view
The Value of the UML
 Is an open standard
 Supports the entire software development
lifecycle
 Supports diverse applications areas
 Is based on experience and needs of the
user community
 Supported by many tools
Creating the UML

UML 1.3
OMG Acceptance, Nov 1997
Final submission to OMG, Sep ‘97 UML 1.1
public First submission to OMG, Jan ´97
feedback
UML partners UML 1.0

Web - June ´96 UML 0.9

OOPSLA ´95 Unified Method 0.8

Other methods Booch method OMT OOSE

UML Partners
 Rational Software Corporation
 Hewlett-Packard
 I-Logix
 IBM
 ICON Computing
 Intellicorp
 MCI Systemhouse
 Microsoft
 ObjecTime
 Oracle
 Platinum Technology
 Taskon
 Texas Instruments/Sterling Software
 Unisys
Contributions to the UML
Harel
Meyer Gamma, et al
Statecharts
Before and after Frameworks and patterns,
conditions
HP Fusion
Booch
Operation descriptions and
Booch method message numbering

Embley
Rumbaugh
Singleton classes and
OMT
high-level view

Jacobson Wirfs-Brock
OOSE Responsibilities
Shlaer - Mellor Odell

Object lifecycles Classification

Overview of the UML
 The UML is a language for
­ visualizing
­ specifying
­ constructing
­ documenting

the artifacts of a software-intensive system

Overview of the UML
 Modeling elements
 Relationships
 Extensibility Mechanisms
 Diagrams
Modeling Elements
 Structural elements
­ class, interface, collaboration, use case,
active class, component, node

 Behavioral elements
­ interaction, state machine

 Grouping elements
­ package, subsystem

 Other elements
­ note
Relationships
 Dependency
 Association
 Generalization
 Realization
Extensibility Mechanisms
 Stereotype
 Tagged value
 Constraint
Models, Views, and Diagrams
A model is a complete
description of a system
from a particular
State
perspective State
Diagrams
Class
Diagrams
Use Case Diagrams
Use Case
Diagrams State
Use Case Use Case
Diagrams State
Diagrams
Use Case Diagrams Object
Diagrams
Diagrams
Sequence Diagrams
Diagrams
Diagrams

Scenario State
Scenario
Diagrams State
Diagrams
Collaboration
Diagrams Models Component
Diagrams
Diagrams Diagrams

Scenario Component
Scenario
Diagrams
Component
Diagrams
Deployment
Statechart
Diagrams Diagrams
Diagrams Diagrams
Activity
Diagrams
Diagrams
 A diagram is a view into a model
­ Presented from the aspect of a particular
stakeholder
­ Provides a partial representation of the system
­ Is semantically consistent with other views

 In the UML, there are nine standard diagrams

­ Static views: use case, class, object, component,
deployment
­ Dynamic views: sequence, collaboration,
statechart, activity
Use Case Diagram
 Captures system functionality as seen by
users
Use Case Diagram
 Captures system functionality as seen by
users
 Built in early stages of development
 Purpose
­ Specify the context of a system
­ Capture the requirements of a system
­ Validate a system’s architecture
­ Drive implementation and generate test cases

 Developed by analysts and domain experts

Class Diagram
 Captures the vocabulary of a system
Class Diagram
 Captures the vocabulary of a system
 Built and refined throughout development
 Purpose
­ Name and model concepts in the system
­ Specify collaborations
­ Specify logical database schemas

 Developed by analysts, designers, and

implementers
Object Diagram
 Captures instances and links
Object Diagram
 Shows instances and links
 Built during analysis and design
 Purpose
­ Illustrate data/object structures
­ Specify snapshots

 Developed by analysts, designers, and

implementers
Component Diagram
 Captures the physical structure of the
implementation
Component Diagram
 Captures the physical structure of the
implementation
 Built as part of architectural specification
 Purpose
­ Organize source code
­ Construct an executable release
­ Specify a physical database

 Developed by architects and programmers

Deployment Diagram
 Captures the topology of a system’s
hardware
Deployment Diagram
 Captures the topology of a system’s
hardware
 Built as part of architectural specification
 Purpose
­ Specify the distribution of components
­ Identify performance bottlenecks

 Developed by architects, networking

engineers, and system engineers
Sequence Diagram
 Captures dynamic behavior (time-oriented)
Sequence Diagram
 Captures dynamic behavior (time-oriented)
 Purpose
­ Model flow of control
­ Illustrate typical scenarios
Collaboration Diagram
 Captures dynamic behavior (message-
oriented)
Collaboration Diagram
 Captures dynamic behavior (message-
oriented)
 Purpose
­ Model flow of control
­ Illustrate coordination of object structure and
control
Statechart Diagram
 Captures dynamic behavior (event-
oriented)
Statechart Diagram
 Captures dynamic behavior (event-
oriented)
 Purpose
­ Model object lifecycle
­ Model reactive objects (user interfaces,
devices, etc.)
Activity Diagram
 Captures dynamic behavior (activity-oriented)
Activity Diagram
 Captures dynamic behavior (activity-oriented)
 Purpose
­ Model business workflows
­ Model operations
Architecture and the UML

Design View Implementation View

Classes, interfaces,
collaborations Components
Use cases

Use Case View

Process View Deployment View

Active classes Nodes

Organization Dynamics
Package, subsystem Interaction
State machine
Software engineering process
A set of partially ordered steps intended to
reach a goal. In software engineering the
goal is to build a software product or to
enhance an existing one.

 Architectural process
­ Sequence of activities that lead to the
production of architectural artifacts:
 A software architecture description
 An architectural prototype
Rational Unified Process
 Iterative
 Architecture-centric
 Use-case driven
 Risk confronting
Focus over time

Discovery Invention Implementation

Focus
Key concepts
When does
 Phase, Iterations architecture happen?

 Process Workflows What does

­ Activity, steps happen?

 Artifacts What is
produced?
­ models
­ reports, documents
Who does
 Worker: Architect it?
Lifecycle Phases

Inception Elaboration Construction Transition

time

 Inception Define the scope of the project and

develop business case
 Elaboration Plan project, specify features, and
baseline the architecture
 Construction Build the product
 Transition Transition the product to its users
Major Milestones

Inception Elaboration Construction Transition

time

Vision Baseline Initial Product

Architecture Capability Release
Phases and Iterations

Inception Elaboration Construction Transition

Prelim ... Arch ... Dev Dev ... Trans ...

Iteration Iteration Iteration Iteration Iteration

Release Release Release Release Release Release Release Release

An iteration is a sequence of activities with an established plan and

evaluation criteria, resulting in an executable release
Architecture-Centric

 Models are vehicles for visualizing, specifying,

constructing, and documenting architecture
 The Unified Process prescribes the
successive refinement of an executable
architecture
Inception Elaboration Construction Transition

time
Architecture
Unified Process structure
Phases
Process Workflows Inception Elaboration Construction Transition

Business Modeling
Requirements
Analysis & Design
Implementation
Test
Deployment
Supporting Workflows
Configuration Mgmt
Management
Environment
Preliminary Iter. Iter. Iter. Iter. Iter. Iter. Iter.
Iteration(s) #1 #2 #n #n+1 #n+2 #m #m+1

Iterations
Architecture and Iterations

Use case Design Implementation Deployment Test

Model Model Model Model Model

Content
Architectural design
 Identify, select, and validate
“architecturally significant” elements
 Not everything is architecture
­ Main “business” classes
­ Important mechanisms
­ Processors and processes
­ Layers and subsystems
­ Interfaces

 Produce a Software Architecture Documen

Architectural design workflow
 Select scenarios: criticality and risk Use case view

 Identify main classes and their responsibility

Logical view
 Distribute behavior on classes
 Structure in subsystems, layers,
define interfaces Implementation view

 Define distribution and concurrency Deployment view

 Implement architectural prototype Process view

 Derive tests from use cases

 Evaluate architecture
Iterate
Sources of architecture

Method Method

Theft Intuition Theft Intuition

Classical system Unprecedented system

Patterns
 A pattern is a solution to a problem in a
context
 A pattern codifies specific knowledge
collected from experience in a domain
 All well-structured systems are full of
patterns
­ Idioms
­ Design patterns
­ Architectural patterns
daVinci

Mechanisms
 Screws Brakes
 Keys Pipes
 Rivets Valves
 Bearings Springs
 Pins, axles, shafts Cranks and rods
 Couplings Cams
 Ropes, belts, and chains Pulleys
 Friction wheels Engaging gears
 Toothed wheels
 Flywheels
 Levers and connecting rods
 Click wheels and gears
 Ratchets
Design Patterns
Gamma et al

Design patterns
 Creational patterns
­ Abstract factory
­ Prototype
 Structural patterns
­ Adapter
­ Bridge
­ Proxy
 Behavioral patterns
­ Chain of responsibility
­ Mediator
­ Visitor
 Mechanisms are the soul of an architecture
Modeling a design pattern
Modeling a design pattern (cont.)
Modeling a design pattern (cont.)
Software Architecture
Shaw and Garlan

Architectural patterns Buschmann et al

A System of Patterns
Buschman et al
Booch

 Distributed  Layered
 Event-driven  MVC
 Frame-based  IR-centric
 Batch  Subsumption
 Pipes and filters  Disposable
 Repository-centric
 Blackboard
 Interpreter
 Rule-based
Real-Life Object-oriented Systems
Soren Lauesen

Complex business system

Customer
name : String Sales
Order GUI layer
Address : String product : Product
date : Date
save()

ServiceAgent Observer

purchase(customer, product, items) update()

Middle layer

Customer Order Line Product

name : String items : Product name : String
Address : String price : Currency
getName()
save() updateName() getName()
getName() updateName()
updateName()

Customer Order Line Product SQL Database

* *
Logical application architecture

Graphical Graphical Graphical

User User User
Interface Interface Interface

Relational Business Business

Database Object Object
Model Model

Relational
Database
Relational
Database
Physical application architecture
Thinner client, thicker server

Client A Client B Client C

Application Application WWW Browser

Business Object DCOM

CORBA Beans
Services ADO/R

Business Object
Engine Web
HTML
Business COM Beans Server CGI ASP Java
Object Server MTS ETS

Business Object Business Object

Services Services

Business Object Business Object

Engine Engine

Relational Database Server(s)

The Second Wave
Paul Dreyfus, Netscape

Complex Internet system

Dynamic HTML, JavaScript, Java

Client
plug-ins, source code enhancements

Java, C, C++, JavaScript, CGI

Server

Application Java, C, C++, JavaBeans, CORBA, DCOM

Server

Fulfillment Financial Inventory RDBMS Native languages

System System System Server
Who are the architects?
 Experience
­ software development
­ domain

 Pro-active, goal oriented

 Leadership, authority
 Architecture team
­ balance
Architect
 Not just a top level designer
 Need to ensure feasibility

 Not the project manager

 But “joined at the hip”

 Not a technology expert

 Purpose of the system, “fit”,

 Not a lone scientist

 Communicator
Software architecture team charter
 Defining the architecture of the software
 Maintaining the architectural integrity of the
software
 Assessing technical risks related to the
software design
 Proposing the order and contents of the
successive iterations
 Consulting services
 Assisting marketing for future product definition
 Facilitating communications between project
teams
Architecture is making decisions

The life of a software architect is a long (and

sometimes painful) succession of suboptimal
decisions made partly in the dark.
Futures
 ADL: Architecture Description Languages
­ UML, UniCon, LILEAnna, P++, LEAP, Wright,
µRapid

 Standardization of concepts
­ IEEE Working Group on Architecture
­ INCOSE Working Group on System
Architecture

 Systematic capture of architectural patterns

References (Architecture)
 Len Bass, Paul Clements & Rick Kazman, Software Architecture in
Practice, Addison-Wesley, 1998.
 Frank Buschmann, Régine Meunier, Hans Rohnert, Peter Sommerlad,
and Michael Stahl, Pattern-Oriented Software Architecture - A System
of Patterns, Wiley and Sons, 1996.
 Christine Hofmeister, Robert Nord, Dilip Soni, Applied Software
Architecture, Addison-Wesley 1999.
 Eric Gamma, John Vlissides, Richard Helm, Ralph Johnson, Design
Patterns, Addison-Wesley 1995.
 Philippe Kruchten, “The 4+1 View Model of Architecture,” IEEE
Software, 12 (6), November 1995, IEEE.
­ http://www.rational.com/support/techpapers/ieee/

 Eberhardt Rechtin, Systems Architecting: Creating and Building

Complex Systems, Englewood Cliffs NJ, Prentice-Hall, 1991.
References (Architecture)
 Eberhardt Rechtin & Mark Maier, The Art of System
Architecting, CRC Press, 1997.
 Recommended Practice for Architectural Description, Draft 2.0 of
IEEE P1471, May 1998
­ http://www.pithecanthropus.com/~awg/

 Mary Shaw, and David Garlan, Software Architecture—

Perspectives on an Emerging Discipline, Upper Saddle River,
NJ, Prentice-Hall, 1996.
 Bernard I. Witt, F. Terry Baker, and Everett W. Merritt, Software
Architecture and Design—Principles, Models, and Methods,
New York NY, Van Nostrand Reinhold, 1995.
 The World-wide Institute of Software Architects
­ http://www.wwisa.org
References (UML)
 Grady Booch, James Rumbaugh, Ivar Jacobson, The Unified
Modeling Language User Guide, Addison-Wesley, 1999.
References (Process)
 Barry Boehm, “A spiral model of software development and
enhancement,” IEEE Computer, May 1998.
 Barry Boehm, “Anchoring the software process,” IEEE Software, July
1996.
 Grady Booch, Object Solutions, Addison-Wesley, 1995.
 Philippe Kruchten, “A Rational Development Process,” CrossTalk, July
1996.
­ http://www.rational.com/support/techpapers/devprcs/

 Philippe Kruchten, The Rational Unified Process - An Introduction,

Addison-Wesley, 1999.
 Rational Unified Process 5.0, Rational, Cupertino, CA, 1998
 Walker Royce, Software Project Management: a Unified Framework,
Addison-Wesley, 1998
 The Software Program Manager’s Network
­ http://www.spmn.com