A new metric for admission control in multi-rate 802.11 WLANs

We present a lightweight algorithm for the estimation of the achievable throughput in IEEE 802.11 wireless networks. In particular, we consider a WLAN operating in infrastructure mode, where, through our algorithm, the access point (AP) can: (i) estimate the available bandwidth that a new station wishing to join the network can use, and (ii) evaluate the impact on the system performance of admitting the new station in the network. The algorithm can be autonomously run by the AP without requiring the cooperation of the wireless stations, and it accounts for a multi-rate scenario, considering the effects of automatic data rate adaptation mechanisms. Through the proposed algorithm, we also define a new metric, called β-metric, quantifying the impact on network performance that a newly arrived station would have. We maintain that the β-metric can be effectively used for admission control and load balancing in WLANs. Indeed, simulation results obtained using Omnet++ show that the β-metric accurately reflects the system throughput behavior, even in dynamic multi-rate environments where stations join/leave the network and vary their transmission rate.