The Design And Realization Of High-performance Peer-to-peer Streaming Media Content Distribution Server

With continuously development of Internet, various kinds of web application services appeared on the Internet. The born and development of streaming media technology, which is mainly represented by network video and IPTV, has brought new energy and opportunities. However, traditional stream media systems are based on C/S mode. When the number of clients grows to a large amount, it would suffer from huge load pressure, which may cause a bad QoS. To improve the Quality of streaming media service, the PPSDN will be introduced in this thesis, which is based on P2P and CDN technology. The main work is divided into two parts:First, this thesis made a research on streaming media content distribute mechanism, and then designed an intelligent publishing system based on P2P, which makes streaming media content distribute faster and more efficiently without doing any impact on the whole system's service capability. The distribute speed among servers grows faster with the help of P2P. This thesis implemented two kinds of publish mode: Push and Pull, which control the storage and distribution of the streaming media content. With these two modes, the system made the file stored on different servers. After the research of content distribute strategies, this thesis designed a publish strategy based on server global load situations, which implements the load balance of the system and makes the system more reliable and available.Second, this thesis designed and implemented a high performance P2P streaming media content distribute server, which makes the streaming content distributed faster, and provides users with high quality of streaming media on-demand and download services. The server is designed and implemented based on Epoll and asynchronous IO framework, which handles network IO by the Epoll asynchronous event-driven mechanism, and uses Linux asynchronous IO mechanism to improve the efficiency of disk IO operations. An application layer data caching is designed to avoid repeating IO operation of disk. The server has implemented the function of video on-demand, downloading, publishing, and deleting of the streaming media content. At the same time, in order to improve the performance of the server, this thesis designed an Intelligent Dynamic Speed Limited mechanism, which is based on single connection and single content to reduce the load pressure of the server. A load generator is designed to simulate a large number of users to access the server to test the performance of the server. Through statistic and analysis of the test data, the server has the ability to provide stable streaming media services for large-scale users.