Medium Access Control for Achieving Airtime Fairness and Throughput Maximization in IEEE 802.11

Abstract

IEEE 802.11 Wireless LANs (WLANs) is a popular wireless data communication protocol for local area networks and is installed in many mobile devices, such as notebook computers, PDAs, MP3 players, smart phones, smart pads, etc. Providing fairness and maximum throughput is the most important performance issues in IEEE 802.11 WLANs, because channel is accessed through competition. However, the binary exponential back-off scheme, which is adopted for multiple access in the IEEE 802.11 distributed coordination function (DCF), causes unfairness problems. As the IEEE 802.11 standard continues to evolve, IEEE 802.11 physical layer (PHY) supports multiple transmission rates and changes the rates depending on the underlying channel condition via link adaptation. When some stations transmit at low data rates, the performance of the high data rate stations degrades signi_x000C_cantly, and this phenomenon is known as the performance anomaly. As a solution to the performance anomaly, the airtime fairness has been proposed. However, the DCF of IEEE 802.11 cannot provide airtime fairness to all competing stations because the protocol is designed to ensure fair attempt probability. The purpose of this dissertation is to improve airtime fairness of IEEE 802.11 protocol and to provide high throughput at the same time with a minimal modification of DCF. First, we propose the notion of successful transmission time fairness (ST-fairness) as a new criterion for airtime fairness and analyze the ST-fairness in the carrier sense multiple access with collision avoidance (CSMA/CA) mechanism. Next, we propose a new medium access control (MAC) scheme, the successful transmission time fair MAC (STF-MAC), which is fair in terms of successful transmission time and also provides the maximum aggregate throughput of a network. In STF-MAC, a station calculates its backoff value according to its transmission duration and stations' transmission rates distribution of the network in a distributed manner. STF-MAC can also be easily applied to solve the uplink/downlink fairness problem in infrastructure mode. Furthermore, we proposed a modi_x000C_ed version of STF-MAC so that the protocol works when the WLAN stations are within the carrier sensing range of each other. The scheme, which is called additive increase multiplicative decrease ST-Fair Medium Access Control (AMID STF-MAC), is based on the principle of STF-MAC, and determines the attempt probability of a station using the additive increase multiplicative decrease algorithm. To evaluate the performance of STF-MAC and AIMD STF-MAC, NS2 simulations were performed for various network environments. Through the simulations, we demonstrate that STF-MAC and AIMD STF-MAC not only remedy the performance anomaly but also maximize the aggregate throughput under the fairness constraint.