A Network Forensics Tool - To visualize a Packet Capture offline as a Network Diagram including device identification, highlight important communication and file extraction

Wiki has more help too.

Given a Pcap File, plot a network diagram displaying hosts in the network, network traffic, highlight important traffic and Tor traffic as well as potential malicious traffic including data involved in the communication.

• Investigation of a Pcap file takes a long time given initial glitch to start the investigation

• Faced by every forensics investigator and anyone who is analyzing the network

• Location: https://github.com/Srinivas11789/PcapXray

• Make a network diagram with the following features from a Pcap file Tool Highlights:
• Network Diagram – Summary Network Diagram of full network
• Information:
  • Web Traffic with Server Details
  • Tor Traffic
  • Possible Malicious traffic
  • Data Obtained from Packet in Report – Device/Traffic/Payloads
  • Device Details

• Network Diagram
• Device/Traffic Details and Analysis
• Malicious Traffic Identification
• Tor Traffic
• GUI – a gui with options to upload pcap file and display the network diagram

• Python 3

apt install python3-pip
apt install python3-tk
apt install graphviz
apt install python3-pil python3-pil.imagetk
pip3 install -r requirements.txt
python3 Source/main.py

( Make sure to escalate privilege to allow file creations - Run with sudo )

For MAC:

brew install graphviz

• Python 2

apt install python-tk
apt install graphviz
pip install -r requirements.txt
python Source/main.py

( Make sure to escalate privilege to allow file creations - Run with sudo )

• Tkinter and TTK – Install from pip or apt-get – Ensure Tkinter and graphviz is installed (Most Linux contain by default)
  • apt install python-tk
  • apt install graphviz
  • apt install python3-tk (for python3 support)
  • Sometimes ImageTk errors are thrown in python3 env --> use apt install python3-pil python3-pil.imagetk
• All these are included in the requirements.txt file
  • Scapy – rdpcap to read the packets from the pcap file
  • Ipwhois – to obtain whois information from ip
  • Netaddr – to check ip information type
  • Pillow – image processing library
  • Stem – tor consensus data fetch library
  • pyGraphviz – plot graph
  • Networkx – plot graph
  • Matplotlib – plot graph (not used as of now)

• Clone the repository
• pip install -r requirements.txt
• python Source/main.py

• Tested on Linux
• Options for Traffic include - Web (HTTP and HTTPS), Tor, Malicious, ICMP, DNS

• Unstability of the TK GUI:
  • Decision on the GUI between Django and TK, settled upon tk for a simple local interface, but the unstability of the tk gui caused a number of problems
• Graph Plotting:
  • Plotting a proper network graph which is readable from the data obtained was quite an effort, used different libraries to arrive at one.
• Performance and Timing:
  • The performance and timing of the total application was a big challenge with different data gathering and output generation

• Memory Hogging

  • Sometimes memory hogging occurs when lower RAM is present in the system as the data stored in the memory from the pcap file is huge
  • Should be Fixed by moving data into a database than the memory itself

• Race Condition

  • Due to mainloop of the TK gui, other threads could undergo a race condition
  • Should be fixed by moving to a better structured TK implementation or Web GUI

• Tk GUI Unstability:

  • Same reason as above

• Code:

  • clumsy and unstructured code flow

• Current Fix in rare occasions: If any of the above issue occurs the progress bar keeps running and no output is generated, a restart of the app would be required.

• Dockerfile present in the root folder was used to build images
• Already built docker images are found at dockerhub
  • srinivas11789/pcapxray-1.0
  • srinivas11789/pcapxray-2.2
• Performing the steps in run.sh file manually would work to launch the tool via docker (I can help with errors)
• Running run.sh scripts is an attempt to automate (would not work 100 percent)
  • tested on mac and linux - will be better soon!...

• Clean up code (beautify code base from being a prototype)
• Report generation on unique folders for all assets of a packet capture
• Suspicious activity detection
• Support more pcap reader engine
• Traffic support: ICMP, DNS
• Known file type detection and Extract
• Python2 and Python3
• Interactive map

• Structured and clean code flow
• Change the database from JSON to sqlite or prominent database, due to memory hogging
• Change fronend to web based such as Django
• Make the application more stable
• More protocol support
• Clean up code

• Thanks for making it better,
  • Professor Marc Budofsky
  • Kevin Gallagher
• Thanks for all the dependent libraries used
• Logo created with logomakr.com and www.inkscape.org