Performance Analysis of Aggregation-Enabled IEEE 802.11 WLANs With Variable Aggregation Size

Abstract

Aggregation-enabled wireless local area network (WLAN) technologies such as IEEE 802.11n and 802.11ac are densely deployed for high-speed and wireless Internet access. Conventional approaches for performance modeling of 802.11n/ac assume fully saturated traffic, where stations always have enough packets to transmit aggregate frames of full or fixed aggregation size. However, such an assumption is unrealistic, since stations usually transmit aggregate frames with different aggregation sizes depending on several dynamic factors, such as offered traffic load, number of active stations, transmission rate, and random backoff time. In this work, we propose a new performance model of aggregation-enabled 802.11n/ac WLANs for different traffic loads and buffer sizes with the assumption that stations always have at least one packet to transmit. Unlike conventional models where stations always contend to transmit aggregate frames of maximum or other fixed aggregation size, the proposed model dynamically determines the aggregation size based on given offered traffic load, number of stations, transmission rate, and random backoff process. Performance evaluations showed that the proposed model produces remarkably accurate performance estimates for throughput, especially for higher traffic loads.