P2P Content-Propagation Mechanism Tailored by Network Coding
Xi Wei, Dongyang Long · 2009
The recent studies have shown that network coding can benefit peer-to-peer (P2P) content distribution network (CDN) since it eases the packet scheduling and balances the network load by utilizing peers' resources to code. But better coding performance costs increasing computation complexity and buffering space, which may result in higher delay and decoding failure. The situation will be worse if the distribution environment is unsuitable for network coding. This paper contributes to the integration of optimal coding algorithm, adaptive topology and dissemination mechanism. We introduce a network coding approach, multi-generation mixing (MGM), into P2P CDN and establish deterministic guidelines to optimize the approach. Then, we propose PCMNC, a P2P content-propagation mechanism tailored by network coding, which deploys P2P CDN based on improved combination networks which benefit network coding. Furthermore, we optimize the dissemination mechanism to accelerate the uniform distribution of encoding packets. We theoretically analyze the effectiveness of PCMNC and verify it by implementing simulation experiments on NS-2.