The Research On QoS-Based Multicasting Routing Algorithm

As an important communication model in which a sender concurrently transfers data to multiple receivers in computer network, multicast technique is mainly used for transmitting the distributed real-time and multimedia applications such as audio/video conferencing, long-distanced teaching. With the rapid development of multimedia and high commercialization of Internet, QoS, quality of service, for real-time and multimedia applications should be guaranteed stringently in underlying network. Hence, QoS-based multicasting routing problem increasingly becomes the focus of attention and remains to be solved. Nowadays, some heuristic algorithms are proposed to work out this problem. These algorithms can construct a multicast tree satisfying the required QoS and optimize the cost of network but high complexity of themselves make them difficult to be implemented in distributed environments. So this thesis studies the interrelated problems and algorithms of QoS-based multicasting routing with the purpose of designing a multicast routing algorithms characteristic of simpleness, speediness, high efficiency, easy implementation and supporting the QoS requirements.The primary contributions of this thesis contain the following two points. Firstly, a multicast routing algorithm based on main-bone tree is proposed, which can minimize the cost of multicast tree and timely deal with the dynamic changes of member. However it does not consider the QoS requirements of real-time applications. Secondly, to guarantee the required QoS of real-time applications, two multicasting routing algorithms with QoS constraint are proposed. One is multicast routing algorithm with delay constraint. Mainly based on idea of MPH algorithm, this heuristics establishes multicast tree merely with a source node at thebeginning and finishes till the tree holds all members. When a member wants to join the multicast tree , it should satisfy the delay constraint from source to itself and the cost of path from tree to itself is less than the cost of path from tree to other members that are not on the tree. This heuristics can find a least cost multicast tree with delay constrained if such a tree exists. The other is a share multicast tree algorithm with delay constrained, in which a center node is precisely selected for share tree according to the topology of networks at first and then the member dynamically joins or leaves the tree. This heuristics has good capability of finding route in the networks and satisfies the delay constraint from end to end. The above algorithms are simulated in many random network models with the results that they show some advantages such as simpliness, low complexity, good performance and easy to be implemented in the distributed environment. So they can be applied to the realistic environments.