Protocol Stack Optimization And Implementation Based On IEEE 802.11n WI-FI Router

Wireless networking, IEEE 802.11 in particular, has been extremely successful. To meet the need from increasing numbers of Wi-Fi devices, diversification of application services, 802.11a/g uses orthogonal frequency division multiplexing(OFDM), while 802.11 n adds multiple-input multiple-output(MIMO) techniques. These technologies allow data portions of frames to be transmitted at high data rates. This decreases the time spent transmitting data, but does not generally decreases the time spent on a variety of overheads. Network throughput is still difficult to improve. As the number of users increasing the access delay is large, especially in hybrid scenario, and the effective amount of accessable devices cannot meet requirements.We conducted a series of optimization to reduce the impact of the frame overhead on transmission performance, which involve improving throughput, reducing user access delay in mixing scenarios and improve users’(especially 802.11n) experience. First, we optimized the timing of frame aggregation, which made the fame of longer waiting time in the queue higher priority than others. Reasonable choice of A-MSDU and A-MPDU aggregation depended on frame error rate as a parameter. Through these two methods we aim to increase throughput. Then, we designed RAS algorithm on the different rate of the devices. The algorithm lower the user delay by optimizing scheduling without affecting the low user delay. The algorithm configurated channel resources reasonablly, make full use of the high-speed channel rate, and improved all kinds rate users’ experience(especially in high-speed rate).In burst frame transmission, we were difficult to analyze factors influencing the burst transmission. This has resulted in difficult determing the number of a burst, high packet loss rate, large end-to-end delay and difficult to improve throughput. After our in-depth analysis of the burst frame end-to-end delay, the relationship between the number of burst and delay, and relationships with delay and channel load, we came up with ABT burst frame transmission mechanism by the number of burst determined reasonably by channel use-ratio. This mechanism improves the burst transmission throughput and reduce the user’s end-to-end delay, gives full play to the 802.11 n performance.Then, we combined RAS frame aggregation scheduling algorithm with ABT burst frame transmission mechanism in mixed scenarios, and applied it in real complex scenario, which timely switched aggregation and burst transmission according to the channel utilization depended on timely adjustment parameters using channel state. By the optimization, the effective amount of accessable users can be increased, the throughput be improved and the delay be reduced.We implemented these methods on Atheros 9340 chips and evaluated it on the performance of End-to-End transmission delay and thoughtput with TCP scripts with Chariot tools in our prototype of 802.11 n Wi-Fi APs. The experimental results shew that proposed methods can improve the effective client number of APs and reduce access delay in the situation of high-density clients.