Random Linear Network Coding Security Issues

Information transmission and sharing is the most important purpose of the network. Data are transmitted through "store and forward" procedure by the intermediate node in the existing networks. There also emerged C/S and P2P architectures, which have inherent disadvantages of the simple store and forward mechanism. When network coding was first proposed in 2000, it has quickly become a hot research field in the international academic. With network coding the intermediate node was required to transmit data through "store-coding-forward" procedure and the receiver in order to recover the original data by decoding the received encoded data blocks. It is obviously different from traditional routing, but it can greatly improve throughputs, robustness and security of networks. Recently, a number of prominent high-tech companies have been doing related research in network coding.The idea that network coding uses cheap computing power to increase network efficiency, which has important implications on the next generation network architecture development. Nevertheless, there are still many problems need further research, such as the security problems in network coding are different from those in traditional packet routing networks. Network coding is inherent vulnerable to the data pollution attacks. This thesis was mainly focus on this problem.Firstly, we reviewed the fundamental theory of network coding, summarized the latest research works on the security aspects of network coding, analyzed their difference between traditional routing and network coding, and discussed the difficulties in bringing the traditional network security techniques into the network coding systems. Secondly, we discussed two kinds of pollution detection solutions which are based on homomorphic hash function and linear space signature, derived the general formation of the homomorphic hash function and proved their correctness. Finally, comparatively analyzed their computational cost and payload efficiency under different data block size conditions, and designed a new combinatory detection solution for the pollution attack problem.