Medium access control mechanisms for quality of service in wireless computer networks.

The fast growth and development of wireless computer networks and multimedia applications means it is essential that these applications can be transmitted over the standard IEEE 802.11 Medium Access Control (MAC) protocol with high Quality of Service (QoS). The lack of QoS in the standard IEEE 802.11 Distributed Coordination Function (DCF) results in applications with considerably different QoS requirements receiving similar network services. This means, the performance of time-sensitive applications with stringent delay, jitter, and packet loss requirements will not be met. Even time-insensitive applications will be treated unfairly because access to the medium is on a random basis. Therefore, the main aims of this thesis are: (i) to investigate the limitations and performance of the IEEE 802.11 DCF, (ii) to develop a comprehensive solution that provides effective and efficient QoS provisioning in IEEE 802.11 DCF scheme in a fair, scalable, and robust manner. The latter is achieved by developing novel MAC mechanisms for providing QoS in the IEEE 802.11 DCF for multimedia transmission.The scarcity of channel capacity, unfairness and hidden terminal problems, multiple hops, and other conditions and parameters that affect QoS in a wireless network are issues which require in depth investigations and analysis. In this thesis, extensive investigations using the network simulator 2 (NS-2) package were carried out to examine the impact of these issues on the main QoS parameters (throughput, delay, jitter, packet loss and collision). The findings revealed that the IEEE 802.11 DCF protocol performed inadequately when transmitting various applications due to the limitations inherent in its operation. The performance of the IEEE 802.11 DCF protocol was also investigated by studying the impact of varying the values of minimum Contention Window (CW[min]) and the Distributed Inter Frame Space (DIFS). The study shows that inappropriate values of CW[min] and DIFS resulted in significant network performance degradations and demonstrated that it was important to select an appropriate set of MAC protocol transmission parameters in order to provide improved QoS.Artificial Intelligence (AI) techniques using fuzzy logic and Genetic Algorithms (GAs) for assessing and optimising MAC protocol transmission parameters were developed and their effectiveness evaluated. The study confirmed that the application of AI techniques significantly improved the QoS for audio and video applications by more than 50% and fairly shared the channel access among the contending stations as compared to the standard IEEE 802.11 DCF scheme. Ratio based and Collision Rate Variation (CRV) schemes were developed to dynamically adjust the CW and DIFS values according to the current and past network conditions. Using these schemes significant improvements with service differentiation were achieved in an Independent Basic Service Set (IBSS). A queue status monitoring technique was devised for the intermediate stations. This provided QoS differentiation at the MAC layer for multi-hop networks. Autoregressive (AR) models that accurately predicted the network parameters were also developed. These enabled the MAC protocol transmission parameters to be adjusted in an improved manner. Using these models, average QoS was improved by more than 60%; average delay, packet loss and collision were reduced by more than 50% compared to IEEE 802.11 DCF scheme.This led to the development of novel MAC mechanisms to provide QoS in IEEE 802.11 MAC protocol. The mechanisms support multiple QoS metrics and consider traffic history and predict future network conditions. The schemes also are characterised by the simplicity, robustness, and ease of implementation. The contribution of this thesis is the development of a comprehensive solution to provide effective and efficient QoS differentiation in IEEE 802.11 DCF scheme for multimedia transmission in a distributed, fair, scalable, and robust manner. Furthermore, through the use of these approaches, the findings of this study provide a framework that also contributes to the knowledge concerning the QoS over the IEEE 802.11 MAC protocol.