Research On Key Techniques Of QoS-based Multicast Routing

Multicast routing can be described as follow: according to a source node s, a group of destination node set D, a serises of limited condition C and some optimal objects , a tree can be found to cover up s and D and is satisfied with C. Because there would occur to congestion and queuing in network, the large delay, delay jitter and data loss would happen during the time of data packet transformation. But the measurement and control network traffic can not tolerate these faults, beacuse the measurement and control network traffic usually needs the network to offer certain quality of service (QoS), such as smaller time delay, the smaller data packet loss rate, security of information, etc.. Nowadays the study of multicast is consentrated as follows: bandwidth constraint multicast routing, delay constraint multicast routing, delay constraint and least cost multicast routing, delay and dely-jitter constraint multicast routing. The solution for this kind of problem is NP-complete. It means that there is no definition polynomial complexity solution. The QoS-based multicast routing to support real-time traffic and finding a relative simple algorithm is researched in this article. Intilligent algorithm and graph theory is applied for the solution of multicast routing and the feasibility is analysed according to domestic and overseas research information. On the base of it the genetic algorithm is proposed to solve the problem. An improved binary code method is adopted. This algorithm is simple and fast to converge to the optium solution. Simulated annealling algorithm is also proposed to solve the multicast routing problem. It iterally produces new solution, judges the result, accepts or aborts the solution. At last it will obtain a global optimal solution. Simulated annealling algorithm is a random search algorithm and can converge to a good solution. With the theory the computer simulation has been done. For supporting the real-time traffic in this article a hybrid genetic algorithm is proposed for QoS-based multicast routing. Compared to GA and SA, hybrid algorithm is the most suitable for the real-time traffic. Finally, a key management approach of multicast is provided in this article to improve the security of multicast. It is based on the approach of assignment key to every group member through key center. In view of some management schemes where members join, leave or are deleted, key service center must distribute a new key through unicast another time. The bigger amount of members, the greater expenses will be spent. In this article with member varying their upper key service center still distribute the new key through multicast and an ID is assigned to every member to identify their transmission message so as to implement data origin authentication. The essential principle of this approach is to distribute a key generator for each member.