Medium access control techniques for QoS provisioning in adhoc networks

The object of this dissertation is the study of medium access control techniques for QualityofService provisioning in ad hoc networks. As ad hoc are characterised the wireless networks that are not supported by preexisting infrastructure. This study focuses on two layers of the ISO/OSI (Open System Interconnection) networking stack, namely the data link and the network layer. The attention is drawn to distributed medium access control algorithms that are capable of providing QoS assurances. Medium access control techniques are of paramount importance in ad hoc networks, since they determine the effective use of the available raw bandwidth. Initially, the problem of determining the optimal backoff window that maximizes the throughput of distributed medium access control protocols is analytically studied. The validity of the derived analytical value is corroborated by simulation experiments. Aiming at improving the efficiency of medium access techniques, a distributed backoff mechanism is proposed, which dynamically calculates the backoff parameters based on the estimation the offered load. The positive characteristics of the designed protocol are revealed by means of comparison with the most wellknown algorithms that have been proposed in the literature. Next, the issue of providing QoS support at the distibuted wireless MAC layer is investigated and two distributed MAC protocols are proposed that aim at providing soft "perflow" QoS guarantees. To assess the performance of these protocols, two analytical models were devised that permit the performance assessment of the throughput of the protocols for different traffic conditions (number of active stations, offered load), as well as the assessment of their ability to schedule the contending packets based on their QoS requirements. The doctoral dissertation is concluded with a comparative study of the proposed protocols with other wellknown protocols, where the efficiency of each access scheme in maximizing the utilization of the wireless bandwidth while providing QoS support for heterogeneous applications is assessed and the limitations and merits of each mechanism are discussed