In this PhD thesis, we focus on the design and performance optimization of multi-channel multi-interface wireless mesh networks. To take advantage of the increased capacity in such networks, a number of issues has to be handled properly. The first contribution of this thesis is a novel classification and formal evaluation of different channel and interface assignment strategies. In particular, we focus on connectivity in terms of topology formation, density of connections, and neighbor discovery. Our second contribution presents broadcast algorithms able to handle any of the multi-channel multi-interface assignment strategies. These algorithms guarantee a broadcast packet to be delivered with a minimum probability to all neighbors. The third contribution of this thesis consists in evaluating the network capacity (i.e., throughput) obtained through the different channel and interface assignments schemes. More specifically, we propose three mixed integer linear programming formulations to model the routing and bandwidth sharing constraints in presence of interference. We derive then upper and lower bounds for different MAC strategies. The fourth and last contribution of this thesis is the development of a novel cross-layer routing solution for multi-channel multi-interface mesh networks. In particular, we propose a link-quality aware metric to estimate the residual bandwidth of a link. An on-demand routing protocol selects the routes offering the best throughput. All our contributions are validated through extensive simulations that demonstrate the efficiency of our solutions. In summary, this thesis provide insight into the improvement of multi-channel multi-interface wireless mesh networks, as well as guidelines for network designers in planning efficient deployments.
