Anti-Starvation Channel Assignment With Global Conflict Set Selection in IEEE 802.11 WMNs
Ying Qu, Bryan Ng, Hang Yu · 2016
Channel assignment algorithms play a critical role in the performance of IEEE 802.11 wireless mesh network by allocating channels and regulating interference between channels. In a channel assignment algorithm, the strategy used to compute a conflict set is crucial because it directly determines the allocation of channels to a given link. This paper investigates the fairness and prevention of flow starvation due to different strategies in selecting conflict sets in a channel assignment algorithm. We apply five different conflict set selection strategies to a vanilla channel assignment algorithm and compare the resulting network performance of each strategy. Our findings show that none of these conflict set selection strategies prevent flow starvation and achieve poor fairness as they only consider "local channel conflict" and lack the consideration of "global channel conflict". We propose a new conflict set selection for channel assignment algorithms called anti-starvation channel assignment (ASCA). Our proposed algorithm considers both local and global conflict to alleviate flow starvation and improves fairness by 62% (with respect to the fairness index) compared with the best result from existing conflict set selection strategies. To the best of our knowledge, the proposed ASCA is the first one to analyse channel assignment algorithms using the concept of local and global conflict sets.