The emergence of nomadic applications has raised intense research interest in providing Quality of Service (QoS) in Mobile Ad hoc Networks (MANETs). In the networking protocol stack, a QoS capable medium access control (MAC) protocol plays an important role. However, currently, the widely deployed MAC protocol IEEE 802.11 Distributed Coordination Function (DCF) does not provide any QoS guarantee and suffers from a problem of fairness. In this thesis, we formulate the design of MAC protocols as the problem of bandwidth allocation to individual links. The concept of conflict graph is adopted to model the contention relations between link flows in ad hoc networks, and based on this modeling, different algorithms are proposed to address the problem in such a way that the obtained MAC protoc...[ Read more ]

The emergence of nomadic applications has raised intense research interest in providing Quality of Service (QoS) in Mobile Ad hoc Networks (MANETs). In the networking protocol stack, a QoS capable medium access control (MAC) protocol plays an important role. However, currently, the widely deployed MAC protocol IEEE 802.11 Distributed Coordination Function (DCF) does not provide any QoS guarantee and suffers from a problem of fairness. In this thesis, we formulate the design of MAC protocols as the problem of bandwidth allocation to individual links. The concept of conflict graph is adopted to model the contention relations between link flows in ad hoc networks, and based on this modeling, different algorithms are proposed to address the problem in such a way that the obtained MAC protocols for MANETs satisfy QoS objectives such as fairness or a minimal bandwidth guarantee.

Firstly, a distributed max-min fair allocation algorithm is proposed to enable each node in the ad hoc network to calculate the max-min fair share of its links based on the two-hops away link flow information. Once each link flow is assigned its fair share and those of its neighbors, a measurement based backoff algorithm presented in the thesis can be invoked to achieve weighted fairness. Secondly, in order to reduce the communication overheads and meet different fairness objectives, the fair bandwidth allocation problem is also modeled as a general utility based constrained maximization problem. By invoking the Lagrange relaxation and duality, both a non-cooperative and a cooperative game formulation of the problem are proposed. The algorithm derived from the non-cooperative game framework does not depend on any local flow information or contentions. The algorithm derived from the cooperative game framework depends only on one-hop link flow information. The fairness objective of these algorithms is controlled by the adopted utility function. Simulations and numerical results have shown that all these solutions not only provide better fairness than the IEEE 802.11 DCF but also have the capability of quantifying the bandwidth, which is crucial to other processes such as admission control. As an alternative to this class of algorithms that are all based on random access technique, a series of collision-free scheduling algorithms are proposed to guarantee a minimal bandwidth. In these algorithms, with the help of one-hop local flow information, each node determines a local collision-free transmission schedule for all its link flows.