P2P vs C/S networks for chat applications

In the First Biweely update we’ve developed a testing framework to evaluate the performance of Client/Server (C/S) and Peer-to-Peer (P2P) chat systems under real-world network conditions. Using Mininet and iPerf, we simulate varying levels of latency, bandwidth, and packet loss to measure key metrics like Round Trip Time (RTT), message success rate, bandwidth consumption, and setup time. So far, we’ve successfully configured our environment and implemented automated pipelines to ensure consistent data collection and analysis. 

Over the past two weeks, we focused on collecting data for key performance metrics (RTT, success rate, bandwidth usage, and setup time) across both Client/Server and Peer-to-Peer architectures. We simulated real-world conditions including packet loss, bandwidth limitations, and latency. We configured user loads and tested scenarios like server overload and peer disconnects. Our findings show that C/S performs well under stable conditions, while P2P adapts more effectively to congestion and dynamic environments. 

We completed full data collection for both architectures across baseline and impaired network scenarios. Using color-coded time series plots, we analyzed how each model performs. Our findings show that while C/S excels in clean networks with lower RTT and faster setup times, P2P consistently outperforms under high latency, packet loss, and low bandwidth. Overall, P2P proves more resilient in unpredictable conditions, while C/S remains optimal for structured, high-throughput systems.