Distributed Rerouting for Multiple Sessions in Cognitive Radio Networks

This paper proposes a distributed rerouting scheme in cognitive radio networks (CRNs). Specifically, we consider how we can reroute for multiple sessions in a situation where switching channels cannot ensure end-to-end connectivity. We propose a ring searching-based framework for rerouting, which is distributed and considers both the channel availabilities and potential rerouting choices of other sessions. Our framework can be applied for the rerouting of multiple sessions in CRNs and does not require any communication among them. The objective in our model is to minimize the queueing delay for each session that needs rerouting. We propose a novel way to estimate the delay for each session. Each node, in our model, that will be receiving RREQ (Route Request) messages maintains two queues: one queue on each node contains a list of nodes whose RREQ is received, and the other one contains the probabilities of it being chosen. We apply an opportunistic overhear approach to predict the probabilities. Also, we take the channel availabilities into account when constructing the rerouting path. The simulation results prove the high performance of our scheme. We also show the influence of different parameter settings, including both network environment parameters and scheme parameters.