Elastic ring search for ad hoc networks

Abstract

In highly dynamic mobile ad hoc networks, new paths between nodes can become available in a short amount of time. We show how to leverage this property in order to efficiently search for paths between nodes using a technique we call elastic ring search, modeled after the popular expanding ring search. In both techniques, a node searches up to a certain number of hops, waits long enough to know if a path was found, and searches again if no path was found. In elastic ring search, the delays between search attempts are long enough for shorter paths to become available, and therefore the optimal sequence of search extents may increase and even decrease. In this paper, we provide a framework to model this network behavior, define two heuristics for optimizing elastic ring search sequences, and show that elastic ring search can incur significantly lower search costs than expanding ring search.