Slice Of Network Coding-based P2p Content Distribution System Design And Implementation

The development of P2P technology has promoted the change of entire computing model of Internet network from centralized to distributed, and the traditional content distribution has evolved into a mixed-mode of P2P and CDN. So it is no longer difficult to achieve the content distribution with large-scale and large-capacity. Currently the BitTorrent, eMule and Thunder are the most popular PPCDS (P2P content distribution system), but there are still many areas for us to improve it, for instance, the inhomogeneous distribution of resources may bring about the phenomenon of dead file; The block exchange led to the reduction of the utilization of the most resources; network flow has not maximized, and the transportation of the duplication of data stream is also exist.The emergence of network coding theory raises a revolution in information transmission, attracting the majority of scholars and research. Avalanche, the PPCDS based on random liner network coding, for the first time introduced the network coding technology into the field of content distribution, and it greatly enhance the performance of the system, but the problem of high complexity in the encoding and decoding has affected the promotion of the application. To improve the decoding efficiency, in this paper we design a new decoding method, and make it applied to PPCDS. The main work of this paper is as follows:1) Through comprehensive analysis of the defects of the P2P technology and the Avalanche based on the random linear network coding, I have carried out a compromise for the granularity of the shared resource, and finally proposed an idea of sub-chip network coding, meanwhile its encoding and decoding processes are described in detail. In the decoding process of finding the inverse of the coefficient matrix, we abandoned the traditional Gaussian elimination method, and used the triangular decomposition of the matrix instead.2) Based on the study of the CoolStreaming mode, we have presented the architecture of PPCDS based on sub-chip network coding, and design and implement the various modules, in which we have brought up some key algorithms, such as slice selection for the request.3) We design a simple simulation system, optimize the key parameters of the system with an average download time as a standard, comparing with the traditional PPCDS and the PPCDS based on random liner network coding, and simply analyze the results of experiment, finally, we get a conclusion that our system is faster in download file and more robust in efficiency.