Throughput performance and fairness in multihop wireless mesh networks

In IEEE 802.11-based multihop wireless mesh networks (WMNs), nodes that are a number of hops away from the gateway (GW) face two-fold contentions: namely channel access contention between itself and other nodes, and traffic contention between its own and relayed traffic. As a result, they suffer from low throughput and even starvation, which results in unfairness in transmissions. In this paper, we develop a rate control mechanism that can flexibly coordinate the tradeoff between throughput and fairness to adapt to different application scenarios by combining a rate adaptation algorithm, an adaptive capacity estimation algorithm and a queue management scheme. More specifically, with the adaptation algorithm, nodes can dynamically change their rates according to the number of flows observed or to the updated rate overheard; with the adaptive capacity estimation algorithm, nodes can adjust their rates by varying the fairness index provided that a certain fairness index threshold is satisfied; while the queue management scheme enables nodes to transmit their data packets originated from different flows (including their local flows) in a proportional manner. To evaluate the effectiveness of the proposed scheme, we demonstrate its performance via ns-2 simulations in terms of throughput, fairness as well as convergence property.