An intelligent backoff scheme for IEEE 80211 DCF wireless LANs

Abstract

The wireless MAC Protocol in IEEE 802.11 uses a distributed contention resolution mechanism for sharing the wireless channel. In this environment, the CSMA/CA scheme requires nodes contending for access to the channel to wait for a Backoff interval, randomly selected from a specified range of contention window (CW), before initiating a transmission. Every time a node packet is involved in a collision, the CW size for that node is doubled up to a maximum CW, and the new CW is used for the following transmission attempt. After a successful transmission a node resets its CW to zero. The random selection of Backoff time and the resetting mechanism causes inefficiency and unfairness in transmitting nodes. It may also cause unfairness with optimized nodes obtaining an unfair advantage over well behaved nodes. In this paper, we propose to replace the random selection of Backoff with an algorithm that will intelligently determine the Backoff time. The algorithm uses machine learning, Bayesian approach. We implement our proposed intelligent Backoff algorithm in the Opnet Modular simulator. Results from simulation experiments reveal that the proposed algorithm achieves higher throughput than the Binary Exponential Backoff (BEB).