Research On Content Distribution Based On Network Coding

With the growth of demand for big data, content distribution becomes one of the most important applications on the Internet. P2P content distribution network attracts more and more attention for its good flexibility, availability and scalability by many users and researchers. At the same time, peer-to-peer nodes and blocking data transportation have brought peer dynamics and scheduling problem.Ever since proposed in2000, Network Coding has attracted a lot of attention in the areas of network communications and information processing, etc. Unlike traditional routing scheme in which nodes just store and forward the data, the essence of network coding is to allow coding in the intermediate nodes. Theory and practice show that network coding has good performance in improving throughput and reducing energy consumption.The introduction of network coding into P2P content distribution network helps solve dynamic and scheduling problem, improves throughput and reduces distribution time, intermediate peers produce encoded blocks and then forward to other peers. The introduction of computing in intermediate nodes, however, raises the computational complexity, which becomes the obstacle to further improve overall system performance.We combine the idea of chunked network coding and sparse network coding and propose a "Partial Segment Network Coding" scheme. In PSNC, intermediate nodes just code partial instead of all data blocks. This helps improve coding efficiency and reduce computational complexity. Meanwhile, we optimize some details in p2p content distribution system based on PSNC scheme, improve segment scheduling problem solutions and introduce hierarchical nodes and slow start mechanism, which further improve the overall system performance.