An Application Layer Multicast Model Based On Small-World Networks

With the development of computer network technology and multimedia technology, group communication services which require transferring data from point to multipoint or multipoint to multipoint increase rapidly. These services include audio/video conferences, Web cache updates, file distribution, interactive simulation, online live broadcasting and online games, etc. Multicast technology, which transmits the same data from a source to a large number of destination nodes simultaneously, in order to save a great deal of bandwidth and reduce data redundancy in the network, can resolve the problem of bandwidth bottlenecks of the communications in a certain extent, and has become an active research area in Internet.Because of some technology and economy issues, IP multicast services among the whole network haven't been deployed completely. Under the circumstance that IP multicast cannot satisfy these application requirements, researchers begin to consider utilizing network resources of Internet end users to provide multicast services directly in the application level, therefore, the notion of application level multicast has been put forward. Unlike IP multicast, application layer multicast requires no infrastructure support and can be easily deployed in Internet.This paper introduces multicast and application layer multicast technology, and analyzes conditions and characteristics. A comprehensive survey on multicast schemes is given, the advantages and application range of these schemes are discussed respectively, and then a comprehensive comparison is given. Based on these, this paper has put forward an application layer multicast scheme based on small-world networks. An overlay network based on small-world networks is introduced, which can enhance the delay performance of the entire multicastsy system. Meanwhile, regional division can realize self-organization, which not only does great help to flow control , but also significantly saves bandwidth. In addition, weighted random early detection technology based on genetic algorithm is presented to solve the problem of congestion control. The experiment results show that this model improvesperformance preferably, such as link delay performance.