The Study Of The P2P Transmission Technology For High Definition VOD

High-definition (HD) television channels are increasing rapidly, which need much more bandwidth of the networks and the higher performance of the servers. Thus the quantity of the concurrent users is reduced in several times for the online VOD (video on demand) technologies of IPTV and IP-QAM. Especially, IPTV cannot carry the real HD-VOD for the lack of QoS. So it's appropriate to play the programs after downloading for moving the downloading traffic from the peak time. Then P2P technology can be employed to decreasing the demand for the server's performance. The HD-VOD method based on P2P is put forward in MAN, which uses the central resource indexing method due to no piracy. The home server is built on NAS as the P2P node, and serves the several STBs as the VOD server.The jam would happen in the backbone of MAN due to the random transmission of the traditional P2P. MAN is a structured network in the small scale, and the distribution characteristics of the P2P flow is analyzed by simulating. A tree structured network model of MAN is built, and then an algorithm is proposed to select the nearest neighbors. Multi-dimensional vector is applied to descript peer node's location. The topology of MAN and peer node vectors were saved in the topology server. To avoid the negative effects of rare piece, the seed node's operating mechanism was optimized; which can balance the amount of rare pieces by giving priority to the nodes requesting the rare pieces. To distribute the resources to cover MAN rapidly, a scope-first algorithm is proposed to give priority to the remote neighbors.In order to solve the shortage of center serving mode, Microsoft network load balancing technology is adopted to achieve concurrent index service. By all the servers in the NLB cluster share the same virtual IP address and the same MAC address, the service for the peers' resources information updating is built with the port out of NLB for receiving the messages' broadcasting from the peers, which ensures the provision of continuous and reliable service. At last, the paper emulates scenarios of this system, executes and gives a performance test. The results of tests indicate that the system is feasible and the overall performance is good in reliability and scalability.