Load-Aware AP Scheduling in Green WLANs

Due to the extensive deployment of WLANs all over the world, the related energy consumption issue is becoming more and more wide-reaching and calls for careful attention. To cut unnecessary energy consumption, green WLANs have managed to employ AP on/off scheduling in response. However, the existing AP scheduling methods yield an uneven distribution of AP loads which causes traffic congestion. Hence, in this paper we propose a load-aware AP scheduling mechanism which is divided into two steps. In the first step, we put forward an optimal greedy algorithm which minimizes the number of active APs under a certain load constraint without hampering user performance. Furthermore, in the second step we extend a load-balancing algorithm based on bipartite graph semi-matching to prevent the decrease of the network performance posed by the improper migration of users between APs. Consequently, we can achieve load balancing among APs and at the same time, avoid escalating the number of active APs. Finally, we perform comprehensive experiments to evaluate the performance of our algorithms. According to the experimental results, the feasibility and effectiveness of the algorithms are verified