Project WEMON

Network monitoring and troubleshooting from within the browser: a data-driven approach

About the Project

WEMON seeks to develop a data-driven light-weight solution for network monitoring and troubleshooting, a solution able to estimate the performance of the network, both mobile and fixed, and troubleshoot it in case of service degradation. By relying on measurement data freely available within the browser (e.g., data on page rendering), this solution has the advantage of passively monitoring the network without the need to actively probe it as with existing solutions (e.g., Speedtest), thus saving network resources and ensuring continuous monitoring of the network. The solution, embedded in a web extension, integrates statistical learning models that transform the passive web measurements into network performance estimators (e.g., latency and download speed) and classifies the network anomalies when they occur according to their origins. We study the feasibility of the approach by collecting data through controlled experiments in the lab, calibrate the learning models and aim at making the web extension public.