On Nodal Encounter Patterns in Wireless LAN Traces

Abstract

In this work we study WLAN traces from five different sources and focus on investigation of encounter patterns between users. We find that typical wireless LAN users encounter with a small portion of the whole population (no more than 60% in all traces, and on average between 1.88% to 6.70%). Total encounters of MNs follow BiPareto distribution. These few encounters are sufficient to build a connected relationship network, which is a Small World graph. We further investigate the potential of node-to-node information diffusion, and find that the richness of encounter pattern provides a reliable platform on which information diffusion without infrastructure is feasible and robust.