Client-driven load balancing through association control in IEEE 802.11 WLANs

The growth of IEEE 802.11 wireless local area networks (WLANs) (Wi-Fi) brings new possibilities of getting connected in public spaces, known as Hot Spots. Current client-access point associations are an interesting research topic because in these scenarios, users tend to be ‘gregarious’ and essentially static. Under IEEE 802.11 standards, association and roaming decisions are made by client devices and most implementations are based on signal strength measurements; i.e. a client station selects the access point (AP) that provides the strongest signal, which leads to an uneven distribution of clients and load between neighbouring APs. As it can be observed in practical scenarios, the default AP-client association scheme followed in IEEE WLANs, produces unfair situations. This work provides means to effectively alleviate this performance issue and also gives details for a feasible implementation. In this paper we analyse how new IEEE 802.11 standards will allow new radio measurements to provide more efficient association decisions. We propose a new load metric that will produce client-driven associations that ensure greater fairness and throughput. Copyright © 2008 John Wiley & Sons, Ltd.