ES2: Managing link level parameters for elevating data rate and stability in High Throughput WLAN

Abstract

High Throughput (HT) IEEE 802.11n/ac wireless networks support a large set of configuration parameters, like Multiple Input Multiple Output (MIMO) streaming, modulation and coding scheme, channel bonding, short guard interval, frame aggregation levels etc., that determine its physical data rate. However, all these parameters have an optimal performance region based on the link quality and external interference. Therefore, dynamically tuning the link parameters based on channel condition can significantly boost up the network performance. The major challenge in adapting all these parameters dynamically is that a large feature set need to be enumerated during run-time to find out the optimal configuration, which is a not feasible in real time. Therefore in this paper, we propose an estimation and sampling mechanism to filter out the non-preferable features on the fly, and then apply a learning mechanism to find out the best features dynamically. We apply a Kalman filtering mechanism to figure out the preferable feature sets from all possible feature combinations. A novel metric has been defined, called the diffESNR, which is used to select the best features from the sampled feature sets. The proposed scheme, Estimate-Sample-Select (ES2) is implemented and tested over a mixed wireless testbed using IEEE 802.11n and IEEE 802.11ac HT wireless routers, and the performance is analyzed and compared with other related mechanisms proposed in the literature. The analysis from the testbed shows that ES2 results in approx 60% performance improvement compared to the standard and other related mechanisms.