Integrating Trafc Estimation and Dynamic Channel Reconguration in Wireless Mesh Networks

Careful Channel Assignment (CA) and Link Scheduling (LS), tuned to the trafc demand in the network, are required to efciently utilize Multi-Channel Multi-Radio (MC-MR) Wireless Mesh Networks (WMNs). In a dynamic network, where the trafc demand keeps changing with time, we need to recongure the CA and LS to suit to the changing demands. But, a change in CA leads to disruption of trafc resulting in a less reliable and lossy network. In this paper, we propose a theoretical framework to evaluate the efciency of channel reconguration by taking into consideration the two conicting objectives of maximizing network throughput and minimizing the re- conguration overhead. A channel reconguration scheme that takes into account the current state of the network cannd a new CA with signicantly less overhead caused by reconguration. We propose and evaluate polynomially bounded heuristic algorithms for performing demand- based and state aware channel reconguration. Further, in a highly dynamic network scenario, performing recon- �guration very frequently to suit every trafc demand will lead to high reconguration overhead. Whereas doing it less frequently will lead to underutilization of the network. Hence, we propose a scheme that employs prediction techniques to estimate the future trafc demands in order to reduce the frequency of reconguration considering the long term trafc demand and c onduct simulation studies to evaluate this scheme.