Research On An Improved BitTorrent Content Distribution System Based On Network Coding

At present, the content distribution system as a content localization and distribution tool through collaboration has become a new type of web application occupying the largest number of Internet network traffic. BitTorrent is a very popular content distribution system based on P2P technology in today's Internet, it can quickly and effectively distributes large file in the Internet network nodes, rather than the source server will bring the excessive load. However, BitTorrent system has problems such as strong dependancy on the seed nodes, the piece selection algorithm achiving only local optimization and lower utilization of network bandwidth and so on.As a new technology in the communication field, Network coding has been widespread concerned and researched. How to apply network coding in P2P content distribution system to solve their existing problems, as well as its ability for P2P content distribution system to improve performance is the focus of this paper.1. An intensive study of the BitTorrent protocol and BitTorrent content distribution system working principle is made, based on this, the open source code of client Enhanced CTorrent is analyzed. The BitTorrent system deficiencies and shortcomings are issued, and a detailed analysis of its causes is put forward.2. An analysis of network coding theory is made in detail. The encoding and decoding principle of random linear network coding is discussed and sparse encoding strategy is researched preliminarily.3. On the basis of intensive understanding of network coding and Enhanced CTorrent client, an improved BitTorrent content distribution system based on network coding is designed and implemented, with solving the existing problems in the system and improving its download performance.4. The codecs performance of random network coding using sparse coding strategy is test. Setting up the private BitTorrent network, the content distribution system is integrating test and an analysis of its download performance is made to prove the superiority of network coding technology.