Epidemic spread in mobile ad hoc networks: Determining the tipping point

Short-range, point-to-point communications for mobile users enjoy increasing popularity, particularly with the rise in Bluetooth-equipped mobile devices. Unfortunately, virus writers have begun exploiting lax security in many mobile devices and subsequently developed malware exploiting proximity-based propagation mechanisms (e.g. Cabir or CommWarrior). So, if given an ad-hoc network of such mobile users, will a proximity-spreading virus survive or die out; that is, can we determine the "tipping point" between survival and die out? What effect does the average user velocity have on such spread? We answer the initial questions and more. Our contributions in this paper are: (a) we present a framework for analyzing epidemic spreading processes on mobile ad hoc networks, (b) using our framework, we are the first to derive the epidemic threshold for any mobility model under the SIS model, and (c) we show that the node velocity in mobility models does not affect the epidemic threshold. Additionally, we introduce a periodic mobility model and provide evaluation via our framework. We validate our theoretical predictions using a combination of simulated and synthetic mobility data, showing ultimately, our predictions accurately estimate the epidemic threshold of such systems.