As computer networks grow in scale and complexity and as the volume of traffic traversing them increases prodigiously, the concept of network monitoring has emerged and developed into a critical tool by which traffic patterns and network behavior can be observed and analyzed. Several solutions currently exist in the research and commercial domains to provide network monitoring services in both hardware and software; however, the infrastructure requirements, processing models, and general inflexibility of these systems can make it difficult to leverage their monitoring capabilities in a lightweight way or in real time.
To address these problems, we have a system that we have developed for monitoring 802.11 wireless network traffic in real time and in a highly flexible manner. Our system provides a lightweight and flexible framework for streaming network monitoring that can be deployed and configured easily while still supplying many of the same robust monitoring capabilities found in more elaborate commercial or research systems. Our work focuses on collecting, analyzing, and reporting on packet traces and other traffic data collected from 802.11 wireless networks. We are primarily interested in extracting relevant data from packet headers at the network and transport layers of the protocol stack and in providing user tools that allow flexible access to both the raw packet data and calculated aggregate values in real time.
This work was conducted as part of the "Internet Scale Sensor Networking" topics course at Harvard University in the Fall of 2007.
The paper I co-authored with Darren Baker about this work can be found here.