A Deep Detection Scheme against Pollution Attacks in Wireless Inter-flow Network Coding

Wireless inter-flow network coding can improve the performance of multi-source unicast communication in multihop wireless networks. However, it encounters with serious security threat of pollution attack. In this paper, by combining signature and arbitration in wireless inter-flow network coding, we propose a deep detection scheme to locate malicious nodes after detecting pollution attacks. Firstly, we design a digital signature scheme which is uncorrelated with field size to detect pollution attacks and locate malicious nodes within one hop. Secondly, we design an arbitration scheme based on trusted node to solve the hard problem of locating malicious nodes beyond one hop, which results from the coding operation of inter-flow network coding. Analysis results indicate that the detection scheme is able to defend against all four types of pollution attacks thoroughly with better performance.