Balanced association algorithm for IEEE 802.11 extended service areas

Most wireless local area network performance estimations are done with the assumption of uniformly distributed (UD) users. In practice, however, stations (STAs) are distributed unevenly among access points (APs) in an extended service area, causing congested hot-spots (HS) and under-utilized APs. Considering a typical network is made up of multiple APs, having some nodes carrying excessive loads degrades the overall network performance. The system performance can be improved by associating STAs efficiently throughout the network, in a sense sharing the network resources fairly among APs and thus relieving congestion. The association algorithm currently employed in IEEE 802.11 systems, that is specifically designed for residential and small office environments, takes into account signal strength as the only parameter and associates STAs to the closest (in signal strength sense) AP, ignoring its load. Novel user association algorithms are required to solve the problems commonly seen in corporate network environments spanning multiple APs, namely congestion relief and resource sharing. In this work, a distributed and online association algorithm is proposed that demonstrates improved average throughput performance, a balanced load distribution as well as fairness across the network compared to the conventional algorithm. Copyright (C) 2008 John Wiley & Sons, Ltd.