Research On Congestion Control For Layered Multicast Based-on Multi-rate

With the development of Internet, multicast network control attracts lots of researchers interesting. Network prediction and packet marking algorithm in are the most important research points of control system. These two aspects are also the main contents of this dissertation.Firstly, the main algorithms of congestion control on multicast are researched and analysed. These algorithms in the hetero-network are short of fairness and friendliness, however, the flexible and congestion in network are solved inefficiently. A new promotion of congestion control algorithm which is FATM (fair aggregate traffic marker) is provided to solve these problems on multicast. The marking algorithm is researched in DiffServ(different services) network, and the fairness performance is importantly analyzed in marking algorithms. The FATM can improve the fairness among the individual flows within an aggregation, while keeping the fairness among the aggregate traffic flows and total network throughput. The auto-regressive prediction model is used to analyze and predicate the traffic of layered multicast of multi-rate with promotion of the algorithm, which provides reliable data to support test of research.Finally, the NS-2 which is a network tool is used to analyze performance in marking algorithms. Simulation results show that FATM algorithm is more responsive, more stable and fair, and can adapt to network heterogeneity better.