Resource sharing and allocation in wireless mesh networks

The onset of Wireless Mesh networks and the increased density of wireless base stations and users call for algorithms that enable collaborative resource sharing in order to enhance the performance. Partisan nature of wireless nodes in such networks, impedes the applicability of network wide sharing methods. We consider groups of nodes in the network with a logical association and thereby a common interest. We explore resource allocation mechanisms in the context of power sharing and arrive at pareto-efficient distributed strategies for guaranteeing fairness. We evaluate the strategies using analysis and simulation under a diverse set of network environments. We next study how frequency sharing can be leveraged for enhanced user performance in the presence of multiple access points in mesh networks, in the face of rapidly varying channels. Towards this, we formulate a partial information based stochastic control problem and quantify the best possible performance achievable as a function of the available information. We demonstrate that owing to partial information, the relation between fundamental QoS metrics gets altered. Using a combination of analysis, simulation and experiments we identify strategies for spectrum utilization and quantify the performance as a function of system parameters. We use channel models developed from measurements carried out in wireless environments for evaluation. The channel models characterize the temporal variation of signal to noise ratio, and we present a methodology to parameterize them using traces measured in a variety of Wi-Fi environments