The Study On Multicast Routing Based On Quality Of Service

With the speed development of the Internet, the multimedia data such as common data, audio, image and video, etc, become the main object transferred through the Internet. These new multimedia applications require that Internet Service Provider (ISPs) provide quality of service (QoS). These QoS requirements are network bandwidth, transmitting cost and delay, delay jitter and the loss rate. But the Internet is single transmission model now and just provides "best effort" for applications, this model cannot meet every kinds of QoS for distinct multimedia applications that request QoS. So, the QoS problem in IP network for applications above is becoming the one of the most important and attractive core studying focus in computer network. But it is very difficult to solve this QoS problem for all kind of different multimedia applications.Nowadays, some multimedia applications which send IP packets for single point to multiple receivers, such as video on demand, visual video-conferencing, real-time cooperation work and government affair system, adapt multicast technique to save network resource and share network links. But compared with the unicast applications, the multicasting applications require more strict QoS requirements. Though some papers have proposed more solution to solve QoS problem for network applications, e.g., admission control routing selection traffic shaping congest control negotiation mechanism, etc, the issues exist universally, the disadvantages of these study are: the algorithms are complex, difficult to realize, can not adopt the real network and also can not fit the specialty of multicasting applications.So, this dissertation employs different thought to design several algorithms for multicast routing trees with QoS constraints for multicasting applications, and solve the shortcut in other algorithms. The main constrains in proposed algorithm are transmitting delay and degree in node, and these algorithms are named as static and dynamic multicast routing algorithms. The KEY management problem which is also a urgent requirement for multicasting applications will also be proposed.In introduction, the paper first presents the background, the content and significance of this study. The chapter 2 presents the multicasting applications, QoS requirements and concepts, the mechanism of QoS guarantee. The QoS realization strategies, multicasting technique and methods of KEY management for multicasting session are also presented. In chapter 3, multicasting event and model are presented in detail, some multicast routing algorithms with constraints are introduced and evaluated.The chapter 4 proposes an improved Genetic Algorithm version for multicast routing tree with delay and degree constrains, the proposed algorithm is named static algorithm. The tree constructed using the improved version has better performance. The chapter 5 just employs genetic operations but not genetic algorithm as heuristic mechanism and also resolve the problem above, heuristic mechanism used here can save running time; the chapter 6 proposes a dynamic multicast routing algorithm which only take delay into account, compared with the two proposed static algorithm, this proposed algorithm simulate real dynamic multicasting. A heuristic strategy in chapter 7 is also a dynamic solution to multicasting with delay and degree constraints. In chapter 8, KEY management in multicasting security is studied, a dynamic strategy based on CBT is proposed to manage multicasting session KEY when membership changes. The chapter 9 concludes all the paper.The several algorithms for multicast routing tree proposed in this paper adapt novel thought, compared with the similar methods in other papers, it can achieve multicasting tree with less transmission cost and meet delay constrained, the algorithms can be realized easily. Especially the algorithm of dynamic multicast routing with delay constraints in the chapter 6, it adapts the advantage in the other papers and proposes a concept "balancing factor" to evaluate quality of a tree or a branch. This algorithm can simulate real multicasting scene and can get small multicasting cost, it generate less number of rearrangement, rearrangement is to rebuild a multicast routing tree when the quality of a tree degrade to a fix value. Less rearrangement affect multicasting session a little.