Packet-Level Channel Switching in Multichannel 802.11 WLANs

In the last few years, the number of wireless devices has increased significantly. Today’s enterprise networks are facing challenges to manage such a large number of devices with the limited wireless resources. To increase the capacity, today’s enterprise networks use a dense deployment of Access Points (APs) [5]. In this scenario, adjacent and overlapping WiFi hotspots (AP-STA clusters, where STAs are associated with an AP) are assigned to different channels (CHNLs) to avoid congestion and to handle the load. It is possible that hotspot A, assigned in CHNL A, serves a lot of stations (STAs) and becomes congested whereas its adjacent other hotspots (say hotspot B, assigned to CHNL B) is currently handling a light load. In this scenario, hotspot A might transfer some of its load in CHNL B without significantly hampering hotspot B’s performance. In our current work we propose a design where, the overly congested hotspot A transmits part of its packets in CHNL B to balance the load, but at the same time it ensures that this extra load in CHNL B is not become a burden to hotspot B. In [4], authors evaluate the performance of an opportunistic multi-channel Medium Access Control (MAC) protocol and compare it to that of the corresponding single-channel MAC and non-opportunistic multi-channel MAC. The study shows that opportunistic access increases bandwidth utilization which reduces the system’s total busy time. On the other hand, the much slower rate of access on the control channel imposes overheads. In SAM [6] and MOAR [3], authors introduce dynamic multi-channel MAC schemes that allow transmitters to dynamically switch between channels in search of good instantaneous channel condition. Such opportunistic approaches, though intuitively appealing, also require reserved control channels, which incur additional overheads. In our previous works Token DCF [2] and O-ACK [1] we designed Medium Access Control (MAC) protocols by incorporating packet overhearing and piggybacking. Those protocols significantly improves the throughput. In our current