Distributed Qos Routing Algorithm

In recent years, with the rapid development of the network technology, more and more multimedia applications such as digital video and audio are applied on Internet, in addition, the quality of service are needed much higher than before. In order to provide high quality of service on the network to satisfy all the requirements, QoS(qualiy of service) has been proposed in recent years. There are several metrics of QoS, such as bandwidth, delay, delay jitter and cost. All of these metrics are very important elements on multimediate applications. Therefore, how to choose a path to satisfy the QoS requirements and how to combine QoS routing with resource reservation become an essential part of network researches.Researchers and experts propose many creative and practicable QoS routing algorithms, which can be divided into two directions: source routing algorithms and distributed routing algorithms. There are also some algorithm that use source routing in local area, and use distributed algorithm in global area. Though having some defects, these algorithms contribute a lot to network researches, and will be very efficient if they are used correctly.In this paper, we firstly show theorems and proofs of the theory on QoS. Then expatiate the characteristics of source routing and distributed routing in general. Later we enumerate and analyze several distributed QoS routing in great details. After finding defects of these algorithms, we propose a new ticket-based multi-path distributed QoS routing algorithm to improve the performance and efficiency of algorithm in some extend. In the new algorithm, every node only need to maintain local state information of its adjacent nodes, ticket is used to decrease the extension of flooding, while multi-path can increase the probability of finding the feasible paths. At last, we use OPNET as the simulator to realize the new algorithm in virtual network environment. Simulation results indicate that new algorithm can improve the performance of network without increase the overhead. So, we can draw a conclusion that this new algorithm is feasible and efficient.