Packet-Centric Tradeoff and Unfair Success Region in IEEE 802.11 WLANs

In this paper, we present a packet-centric analytical study of carrier sense multiple access with collision avoidance (CSMA/CA)-based medium access in wireless local area networks, specifically using the IEEE 802.11 distributed coordination function (DCF). The packet-centric viewpoint of analysis adds a new dimension of insights over those available in the existing body of knowledge in this domain. We propose an absorbing Markov chain-based model for analyzing the behavior of head-of-line frames of each station in such networks. We generalize the long-term and short-term unfairness induced by the randomness of the DCF protocol and depict the actual unfair success region to be avoided in fair resource allocation schemes. One of the primary objectives of any CSMA/CA-based medium access control (MAC) protocol is to increase the frame delivery probability, while reducing the average delay of delivery. With such considerations, we show a tradeoff relationship between the delivery probability and access delay with respect to the maximum retransmission limit in a packet-centric approach. We numerically evaluate the packet-centric behavior of the DCF protocol for single-hop single access point networks involving both the saturated and unsaturated stations, imperfect wireless channel, and varying payload sizes.