The Research On The Security Of Multi-Source Random Linear Network Coding

As network coding can improve the throughput and reliability of network transmission, also, fully use the network resource, the research of network coding has received great interests and become one of the most attractive fields in network information theory. Recently, network coding has been effectively applied in traditional computer network, wireless sensor network and P2P network. However, the applications based on network coding are vulnerable to malicious pollution attack and entropy attack, which, to some extent, limits the applied range of network coding.First, we introduce basic theory network coding comprehensively, and illustrate the random linear network coding in detail, also, we analyze the principles of pollution attack and entropy attack, and the damage caused by them. Second, we introduce a multi-source random linear network coding model through extending the single source random linear network coding model, and propose a new scheme against pollution attack and entropy attack in multi-source random linear encoding network. For resisting pollution attack, we design a new security sub scheme based on message vector subspace. In this scheme, verification nodes verify message through chencking whether all the original message that constitute the message is belong to a certain subspace. For resisting entropy attack, we design a dynamic identity ring signature sub scheme based on message generation. Considering stale message would be filtered by the sub scheme based on message vector subspace, so, we only sign for prior distributed information in order to filter stale prior distributed message at verification nodes. At last, we analyze the security and performance of our scheme. The analysis and experiment results show our scheme can improve the verification cost and encoding throughput in comparison with related schemes.