Distributed channel aware link scheduling for CSMA based wireless networks with time-varying channels and delay sensitive applications

Abstract

In wireless networks, interference between neighboring links is an important issue. The link scheduling algorithm controls the interference between neighboring links such that no adjacent links can be concurrently active. Distributed throughput optimum algorithms for the link scheduling problem have been proposed in the literature. However, the maximum packet delays of these distributed throughput optimum algorithms can become arbitrarily large, which significantly degrades the performances of delay sensitive applications such as “Skype”. In this thesis, we propose two distributed link scheduling algorithms: a full opportunistic algorithm and a delay based adaptive algorithm. The proposed algorithms, while maintaining throughput optimality, increase the average delay performance of the previously proposed throughput optimum scheduling algorithms by 20% under the fading radio channel. We propose a new metric “Effective Goodput”, which measures the rate of packets that are successfully received before their respective playout times for delay sensitive applications. The delay based distributed adaptive scheduling algorithm proposed in the thesis increases the “Effective Goodput” by nearly 100% compared with the throughput optimum scheduling algorithms proposed in the literature.