An Epidemiological Study of Information Dissemination in Mobile Networks.

We present an analytical, epidemiological, model for the overhead of information dissemination in mobile networks, which finds application in a range of settings, including resource discovery, information dissemination in sensor networks, location services and ad hoc network routing. We consider two principal information dissemination mechanisms, one for which nodes strictly relay information for other nodes, and the other for which they also take advantage of the information already being disseminated to insert their own data to share. The latter mechanism is denoted by path accumulation. We will show that the overhead with path accumulation scales as n^3/2 and without path accumulation as n^5/4. With respect to the connection rate, both scale as radic(lambda). We also identify the cutoff point for path accumulation as xi_c = 1 - radic(2/radic(n)). Below that threshold, path accumulation should be preferred, while above, path accumulation does not provide substantial benefits. Our model proves extremely accurate against simulation data.