Delay properties of opportunistic back-pressure routing in CSMA-based wireless mesh networks

In wireless mesh networks, scheduling is the bottleneck of the cross-layer design approach. Thus, it is surprising that CSMA is utility-optimal if it is used traffic-adaptive. However, the price to pay is short-term fairness and thus end-to-end delay. In this paper, we propose a step size adaptation algorithm that combines the advantages of small and large step sizes in a way that the convergence delay is considerably reduced and the steady state queue variability is low and controllable. In addition, we propose a delay-limiting algorithm that determines upper limits for the physical queues according to a given per-hop delay limit. The algorithm controls the burstiness and delays at the expense of throughput, so that the end-to-end latency can be reduced to a level of practical relevance.