Orthogonal vector based network coding against pollution attacks in n-layer combination networks

It is proved that combination networks can achieve unbounded network coding gain measured by the ratio of network throughput with network coding to that without network coding in peer-to-peer file sharing system. On the other side, network coding allows the polluted packet generated by a malicious node to corrupt the packets transmission of the entire network. One polluted packet will prevent the other nodes from correct decoding. To settle the problem, we introduce a polluted packets detection method based on multiple vectors orthogonal to the universal set of vectors for the transmitted network coding packets. The main idea of this method is to calculate numbers of orthogonal vectors and send them to different relay nodes and receiver nodes which use them to judge whether the newly received packets are in the subspace based on the source packets. This is a distributed algorithm, without the requirement of an extra secure channel. Through the analysis and simulation in n-layer combination networks case, the computational complexity of the detection algorithm is O(n), n is the packet length, and the detection rate can be greater than 90%.