Quality Of Services In Mobile Ad Hoc Networks

As a kind of the access networks of B3G networks, problems as terminal mobility, providing quality of service, security and charging must be solved for mobile Ad Hoc networks. Researching on QoS issue is an important aspect of the researches on all of these issues. This article is about quality of services in mobile ad hoc networks.QoS is realized mainly by setting different transmitting parameters for different classes of services, and reserving resource for accepted services. SWAN model is designed for QoS in mobile Ad Hoc networks. This signaling model includes a module that performs admission control for real time services and another mechanism which provide rate control function for best effort services. This thesis studied many aspects of QoS in mobile Ad Hoc networks, and proposed an improved model to overcome the disadvantages of the SWAN model. The SWAN model adopts AIMD rate control algorithm, however, this algorithm does not control rate according to the traffic situation of the whole network, and due to its locating in upper layer, it cannot respond to the changes of network quickly. The proposed model use EDCA which is the MAC layer mechanism of 802.11e protocol as the rate control function. EDCA replaced the AIMD algorithm and the shaper, the model becomes simple and more reliable. This thesis also proposed a new algorithm for estimating the link bandwidth between two nodes in 802.11e networks in cope with the special MAC layer mechanism. The algorithm have also taken the collision in to consideration, and modified the result of bandwidth estimation to make it more accurate in heavy traffic situation. The collide probability also used to determine the portion of bandwidth each AC can obtain.Through analysis and simulation, the new model is proven efficient and reliable of providing QoS in mobile Ad Hoc networks, and the method of replacing rate control algorithm with MAC layer differentiation is viable. The new model raised the throughput of best effort traffic and reduced the end to end delay. The new model has some improvements on adapting to the changes of network and on the reliability of QoS.