LiveNet: Using Passive Monitoring to Reconstruct Sensor Network Dynamics

Abstract

Understanding the behavior of deployed sensor networks is difficult as they become more sophisticated and larger in scale. Much of the difficulty comes from the lack of tools to provide a global view on the network dynamics. This paper describes LiveNet, a set of tools and techniques for reconstructing complex dynamics of live sensor network deployments. LiveNet is based on the use of passive sniffers co-deployed with the network. We address several challenges: merging multiple sniffer traces, determining coverage of sniffers, inference of missing information for path reconstruction and high-level analyses with application-specific knowledge. To validate LiveNet’s accuracy, we conduct controlled experiments on an indoor testbed. Finally, we present data from a real deployment using LiveNet. The results show that LiveNet is able to to reconstruct network topology, bandwidth usage, routing paths, identify hot-spot nodes, and disambiguate failures observed at application level without instrumenting application code.