SDN-Based Centralized Downlink Scheduling with Multiple APs Cooperation in WLANs

Abstract

Conventional DCF and RTS/CTS mechanisms perform the channel contention by a distributed and independent manner, which can lead to severe cochannel interference and low channel utilization in multiple APs dense deployment scenario. In this paper, we propose a channel scheduling cooperation algorithm called CCT-SDN (centralized concurrent transmission based on SDN) that enables multiple APs (Access Points) to perform cooperatively a centralized downlink transmission control, thus achieving higher system throughput and channel utilization by avoiding cochannel interference and implementing concurrent transmission. This design inherits the merit of the conventional distributed random channel access and adopts standardized OpenFlow protocol and Software Defined Network (SDN) architecture to make a centralized concurrent downlink traffic transmission decision among APs. Meanwhile, we also present a novel neighborhood relation storage scheme called SPRIM to enhance the retrieving efficiency of SDN controller, which enables CCT-SDN to perform a real-time control. Moreover, we also develop a theoretical model to prove the improvement of CCT-SDN. Furthermore, our solution does not require any modifications to existing ubiquitous 802.11 terminal devices and thus is likely to be widely deployed. Finally, extensive simulation results on Mininet-WiFi verify that CCT-SDN can achieve significant performance in terms of aggregate throughput, channel utilization, and packet loss rate in different deployment scenarios.