Research On Data Scheduling Algorithm And Implementation Of Streaming VoD System

With the widely development of broadband services, the propotion of streaming media services on the internet services has been increased, and a lot of streaming media applications appeared, such as network television, live broadcast, distance education etc. On traditional streaming media, most of the application system based on the c/s model, the server of data distribute data to all clients while the clients simply receive data. When the user number increased to some point, the server could not serve users well, even a single point failure will appear. In this case, the use of p2p technology brings the dawn for this issue, and became the key point in research of both academic and industry. At present, the stream media business can be divided into general streaming media downloads, live broadcast, demand three categories. The download system broadcasting system based on P2P technology have been developed, because of the interaction and broadcast asynchronous, the implementation of the broadcast system on demand has been increased. Therefore, to build expanded and high reliability, high experience of the broadcast system on demand on the internet has been a hot issue.The contributions of this thesis include:First, we also raised one algorithm based on historical data bandwidth which combined comprehensive consideration, such as, node bandwidth, node service capabilities and candidate nodes number, etc. The full use of the bandwidth reduces both the start delayed and transferring latency, improve the continuity of play.Secondly, a node to node optimization mechanism was raised.the connections between the nodes will be optimized at intervals, a node evaluates the inferior quality of connection to its neighbor by the statistics of the date quantity from that neighbor, to improve the system connection and distribution efficiency.Finally, the two levels management strategy was raised. In memory caching, data are stored to improve the fluency of the play, meanwhile, in the disk cache many video that have been played are also stored to widen the service range of this node to further reduce the load of server and operation costs. Besides, Design and implementation of P2P-based streaming on-demand network systems and related technologies are given the test results.