Integrating Trafc Estimation and Dynamic Channel Reconguration in Wireless Mesh Networks
Athula Balachandran, A. Antony Franklin, C. Siva, Ram Murthy · 2009
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.