A fast-CSMA based distributed scheduling algorithm under SINR model

摘要

There has been substantial interest over the last decade in developing low complexity decentralized scheduling algorithms in wireless networks. In this context, the queue-length based Carrier Sense Multiple Access (CSMA) scheduling algorithms have attracted significant attention because of their attractive throughput guarantees. However, the CSMA results rely on the mixing of the underlying Markov chain and their performance under fading channel states is unknown. In this work, we formulate a partially decentralized randomized scheduling algorithm for a two transmitter receiver pair set up and investigate its stability properties. Our work is based on the Fast-CSMA (FCSMA) algorithm first developed in [1] and we extend its results to a signal to interference noise ratio (SINR) based interference model in which one or more transmitters can transmit simultaneously while causing interference to the other. In order to improve the performance of the system, we split the traffic arriving at the transmitter into schedule based queues and combine it with the FCSMA based scheduling algorithm. We theoretically examine the performance of our algorithm in both non-fading and fading environment and characterize the set of arrival rates which can be stabilized by our proposed algorithm.