Evaluating contacts for routing in highly partitioned mobile networks
Chao Chen, Zesheng Chen · 2007
The movement of real users often follows some patterns that can be characterized by certain statistical metrics of the contacts. Such metrics are useful for nodes to make routing decisions, especially in sparse mobile ad hoc networks where instantaneous network connectivity is rare and messages are delivered using store-carry-forward routing. None of the statistical metrics used in existing routing algorithms considers the dependency between neighboring contacts, thus resulting in an inaccurate estimation of message delays over paths. In this paper, a new metric called the expected dependent delay that characterizes the expected delay of a contact dependent on the previous hop is proposed. With this new metric, the closed-form expression for the expected delay of a path is given and can be further simplified in highly partitioned mobile networks. Simulation results show that the proposed algorithm has a significant improvement in terms of the message delay than those that consider only the delivery probability or the expected delay.