Prof. Joel Sommers was recently awarded a three-year grant from the NSF for a project that seeks to develop methods for automatically capturing and analyzing metadata collected during Internet performance experiments.
The extraordinary growth of Internet connectivity and bandwidth has helped to fuel economic activity, educational opportunities, and social connections around the world. The network research community has continued to play a critical role in assessing current Internet performance and in using measurements of network performance and behavior toward the design of new and improved infrastructure, protocols, and applications. Even with advances in network performance measurement techniques over the years, however, it is easily possible for measurements to exhibit bias or otherwise be error-ridden. The primary objective of this project is to develop methods for capturing important contextual information about a network performance measurement experiment and to create algorithms to assess whether the measurements exhibit bias or other problems. Extensive experiments will be carried out in both a campus network and in the wide-area Internet to evaluate the tools developed through this project. The outcomes of this project will potentially improve the quality and reproducibility of performance measurements produced through efforts such as the Federal Communication Commission’s Measuring Broadband America as well as other prominent network measurement efforts.