Service differentiation and QoS support for wireless networks

Abstract

Quality of service (QoS) support in wireless networks is becoming more and more important in recent years. In this thesis, we discuss the service differentiation and QoS support in wireless networks, specially in the most popular two technologies: IEEE 802.11 based wireless fidelity (WiFi) and IEEE 802.16 based worldwide interoperability for microwave access (WiMAX). Enhancing the IEEE 802.11 protocol, the IEEE 802.11e standard has been designed to overcome the QoS problem in wireless local area networks (WLAN). Many performance analysis researches for IEEE 802.11e assume that the network is saturated. However, the saturation assumption is not always valid in real IEEE 802.11 networks. In this thesis, we introduce a new analytical model for IEEE 802.11e enhanced distributed channel access (EDCA) under the non-saturation and saturation conditions. The model captures all major QoS features of EDCA, including the number of access categories, contention window sizes, arbitration interframe spaces, and virtual collisions. Using the model, we derive the performance analysis for the throughput and average delay for all service classes. The model shows how the traffic loads affect the overall performance of service classes and improves the understanding of EDCA. Simulation results show that our model is more accurate than that of previous works. Extending the model, we introduce a parameter control algorithm which dynamically assigns the QoS parameters to achieve desired throughput ratio. Admission control and rate control (ARC) algorithms are also proposed to decide whether a new flow can be accepted or rejected and to prevent the starvation based on the QoS of current flows. Throughput and/or delay-based ARC algorithms are proposed and the algorithms control admissions of the highest priority class and control rates of lower priority classes. To support the QoS in IEEE 802.16 network, we propose a QoS architecture which consists of...