Applying active measurement in the rate adaptation for 802.11 wireless networks

Rate Adaptation (RA) is a mechanism to choose transmission rate based on the dynamic channel quality in wireless networks. The adaptation algorithm run solely at the sender side in 802.11 networks is studied. The key insight is the inference discrepancy in inferring the relative order of the expected performance of candidate rate, which argues that one can not always reach the correct order based on the channel state information collected exclusively by the sender itself. The consequence is wrong rate decision and significant performance loss. Therefore, a new RA structure is proposed to mitigate such effect by a novel component, rate testing. By embracing the active measurement, a lightweight and effective testing mechanism SFB, short frame burst, is proposed to detect and filter out the unsuitable transmission rate. Finally, an active measurement-based rate adaptation mechanism (AMRA) is designed and implemented. The experiments show that AMRA outperforms many other well-known RA solutions in most scenarios.