FREEZE: Rate adaptation in wireless LANs using channel contention estimates

Abstract

IEEE 802.11 provides different data rates for a device to utilize based on the channel conditions. Rate adaptation is an important feature of wireless LANs and most of the work in this area use frame error rate as an indicator of signal to noise ratio (SNR). SNR is then used to base the choice of data rate used for transmission. In infrastructure based 802.11 LANs, frames can be lost due to collisions and hence increase the error rate, although the SNR might be high. Thus using frame error rate alone as a deciding factor for rate adaptation would not be a good choice. In this paper we use the 802.11 protocol features to estimate the contention in the system. The contention can be determined by measuring the freezes a node experiences while transmitting. We further show that our estimate is a good indicator of the contention in the channel. This estimate along with the frame error rate is then used in making rate adaptation decisions. Our simulations show 35 percent better performance in throughput as compared to other rate adaptation techniques such as ARF.