The Effect of Locality on Coordinate Dissemination

Abstract

In age-based routing protocols, nodes forward data toward a position which is an estimation of the destination's coordinates. The fresher the estimation, the higher the probability to reach the destination. In such a class of solutions, it is fundamental that accurate estimations of nodes' positions be disseminated throughout the network in order for the forwarding algorithm to converge in a reasonable time and with as few hops as possible. In general, dissemination is done following uniform distributions. In this paper, we argue that coordinate dissemination should rely on measured parameters (thus adaptive to the network conditions), such as the mobility of a node. We investigate the impact of node locality on the performance of dissemination algorithms and propose a new algorithm that adjusts the dissemination level of a node's coordinates based on the displacements of the node. Through extensive simulations, we show that our proposal leads to both faster discovery and significant reduction of the overhead when compared to flat solutions