Research On Node Selection Strategy For P2P-based Streaming Media Vod System

With the development of mobile internet,people's access to information is more and more diversified.Lots of short video applications have sprung up in recent years,video is not only just a traditional network television or film but also a sharing medium.Under the mobile Internet era,the video will be everywhere in our life.On this condition,Video on demand system based on P2 P technology has become one of the most widely applications.P2 P model can solve scalable and fault tolerance problems,but it has some challenges such as security,network model and architecture design.According to the characteristics of P2 P network,this article focuses on the evaluation of service ability of resource nodes,which can improve the capability of interaction between each node in the system.To solve problems above,this paper analyzed the research status of video on demand system.Based on the technical characteristics of the system,I analyzed and summarized the existing network model and node selection algorithm.On the basis of BT protocol,this paper designed a distributed video on demand service system,it mainly includes torrent management server,the central management platform,the index server and cache server.The central management platform is responsible for the scheduling of all cache server resources.The torrent management server is responsible for resource storage.In view of the inherent limitations of BT's node selection algorithm,this paper designed an optimization node selection strategy to improve the overall performance respectively.On the server side,I put forward a kind of adaptive node selection strategy based on compensation node.This method maximize the use of common nodes and the cache node resources from the system.On the client side,I propose an adaptive client node selection algorithm according to node upload quota.The system this paper designed introduces the BT protocol to the streaming media on demand system,it not only improves the network resource utilization but also improves the concurrence access.To a certain extent,this system also solved the high consumption of bandwidth when there are lots of users.And at the same time satisfy the demand of the interactive system requirements,users can drag to play at any time.The new node selection strategy improve the timeliness of the system performance and network throughput.And users can easily monitor the use of system resource so that they can adjust it timely.However,there remains some problem to be solved in terms of stability and accidental event processing.