Empirical Analysis of Network Latency under Controlled Traffic Conditions using Wireshark
This video presents my Computer Networks project on network latency analysis using Wireshark under different traffic conditions. The project is inspired by the SharkFest session “Back to the Packet Trenches”, which focuses on packet-level analysis to understand real-world network behavior.
Project Overview
In this project, latency is measured using Round Trip Time (RTT), which represents the time taken for a packet to travel from the source to the destination and back.
The analysis was performed under three controlled traffic conditions:
Normal Traffic
Medium Traffic
Heavy Traffic
Traffic was generated using command-line tools, and packet data was captured and analyzed using Wireshark.
Key Concepts Covered
TCP Packet Analysis
RTT (Round Trip Time) Calculation
Traffic-based Latency Variation
Queueing Delay and Congestion
TCP Behavior and ACK Timing
Analysis Performed
Generated more than 20 RTT graphs using multiple TCP streams
Compared latency under different traffic loads
Observed stability, variation, and congestion effects
AI-Based Analysis
To enhance the analysis, Julius AI was used to:
Generate comparative graphs such as line charts, box plots, and bar charts
Analyze latency trends and variability
Identify congestion patterns
Validate observations obtained from Wireshark
Key Findings
Latency variability increases with traffic load
Heavy traffic leads to unpredictable network behavior
Queueing delay becomes dominant under high load
Average latency is not a reliable performance metric
Conclusion
This project demonstrates that network latency is influenced not only by average delay but also by variability and stability. By combining packet-level analysis with AI-based insights, a more comprehensive understanding of network performance was achieved.
Tools Used
Wireshark
Command Line tools (ping, curl)
Julius AI
Python for data visualization
Resources