The Research On Key Algorithms Of P2P Streaming Media

The transmission and broadcasting of the audio and video information which are large scale and high concurrent on internet have become staple applications in today’s network. P2P streaming media technology has become the research focus of academia and industry. Its theory and practice is so valuable that it could open up a new idea for mass data storage, distributed computing and search engines at next generation. However, Because of the volatility and no center property of P2P streaming media systems themselves, data scheduling algorithm and node management algorithm become fatal factors affecting system performance. Therefore, it is meaningful to research how the data scheduling algorithm and node management algorithm of the P2P streaming system improve the quality of service system.Firstly, aiming at the problem that a large number of nodes in the P2P network behind NAT devices could not communicate directly with each other, the principle of the NAT traversal technology is analyzed and a NAT traversal methods basing on UDP is proposed. Secondly, some major data scheduling algorithms of P2P streaming systems are researched. In order to solving problem that the current algorithms optimized the scheduling strategy mainly on single factor but ignored another key factors that have influence on the system performance, a kind of data scheduling algorithm based on both priority of data blocks and capacity of nodes is proposed and the pseudo-code of the algorithm and experimental comparative analysis are given. And, the shortcomings of the traditional centralized node management strategy are analyzed in this paper and a kind of hybrid management algorithm is proposed. The process of the algorithm is expounded and the design of the core code about this algorithm is given. At the same time, the feasibility of the algorithm is proved on the theoretical points. At last, the General framework plan of the P2P stream media system is given. The management of nodes, the design of database in the web server module and compilation of the core code are finished. Encoding process of streaming code is analyzed. Real time code for stream media is finished. The simulation results indicate the data scheduling algorithm not only could reduce the start-up delay of the system and server’s load but also improve the system’s stability because of giving full consideration to the neighbor node bandwidth, online time and the relative distance between nodes. Theoretical analysis shows that the node management algorithm is feasible and correct, and the system is affected slightly by the fluctuations of the network. Soft test shows that the database design, the compilation of the core code and the builder in real time video are correct.