Channel Access Fairness in IEEE 802.11ac

Abstract

High throughput wireless access networks based on IEEE 802.11ac support a number of protocol enhancements at the physical and medium access control sublayer for supporting data rates in the order of Gigabits per second. These include multiple antenna technologies, wider bandwidth via channel bonding, reducing access overhead via short guard intervals, higher order modulation and coding rates, frame aggregation and block acknowledgements. As these features have their internal trade-offs based on channel conditions, the protocol generally employs a rate/link adaptation technique (sometime called dynamic bandwidth channel access) at the data link layer, that dynamically selects the channel access parameters based on the environment condition. However, this paper shows that such heterogeneity in selecting channel access parameters among neighboring wireless access points results in severe unfairness. In this paper, we address such unfairness in channel access, and develop an intelligent decentralized link parameter selection procedure that significantly improves the protocol performance in terms of fairness and overall network throughput. The proposed scheme, FairHT-MAC, has been implemented in a 26-node (6 access points and 20 client stations) indoor testbed, and the performance is analyzed and compared with other state-of-the-art link adaptation methods, like SampleLite and Minstrel-HT. We observe that FairHT-MAC significantly boosts up access fairness and overall network throughput, while keeps the access overhead (in terms of channel access delay) and average power consumption almost same that of Minstrel-HT and SampleLite.