Research And Design Of Efficient P2P Streaming Media On-Demand System

With the spread of video applications in internet and battlefield network, the research on streaming media on-demand is focused. Traditional method with poor scalability is based on C/S model. The IP Multicast proposed later network layer is too complicated, and the cost of CDN distributing network is too high. On contrary, in P2P network, the peer not only provides but also receives services, then the idle network resources are fully utilized, and the load of streaming media server has been reduced, and the system's scalability has been improved.Based on the analysis and comparison of the three models of P2P streaming media on-demand content distribution——single multicast tree model, multi multicast tree model and mesh model, an efficient P2P streaming media on-demand system based on mesh model is designed in this dissertation, and the self-adaptive system index topology is proposed and constructed. The details as following:designing peer operation such as joining, leaving, maintenance, search and positioning in order to maintain the system topology dynamically; proposing index peer to support the VCR operation better; choosing LDAP as index server database; putting forward the concept of "building group" and "built group" to optimize the peer management in the design process of the database.From the simulation experiment of the self-adaptive system index topology, we found that the system topology has certain superiority not only in the node efficient search and positioning but also in the adaptability. Finally, a detailed hierarchical design of the system node function is proposed in this dissertation.