Research On Secure Network Coding In Multiple-source Multicast Network

In traditional multicast routing algorithms, the intermediate nodes just forward the received information or data. But in this way it can’t achieve the maximum throughput of the network. The proposal of network coding breaks the traditional way of information processing. The intermediate nodes of network coding can process the received information such as encoding and then forward it. Network coding indicates a new direction for improving the throughout and the security of network, and becomes the focus in network research area.This paper mainly considers the problem about the network security based on network coding. At present the mainly security threats of network are the contamination and eavesdropping adversaries. Eavesdropping adversary is that there are wiretappers in the network who can eavesdrop a subset of links and want to obtain some knowledge about what the source has transmitted. Contamination adversary is that the attackers not just want to obtain some valuable knowledge, but also disrupt normal communication by various means including adding an error message ormodifing the normal transmission of information. Both of these two attacks are based on eavesdropping attack, so it is very important to study the anti-eavesdropping secure network coding.This paper first systematically introduces the basic theory of network coding. Then presents the basic model of secure network coding, and introduces a necessary condition in CSWN. Then proposes the relation between the security of random linear network coding and the size of finite field, and proves the security by simulation. Next presents the model of secure network coding in the multiple-source network, and introduces the secure condition. Based on the secure LIF algorithm for single-source multicast network, presents the algorithm for multiple-source multicast network. Then classifies the links based on the relation between the symbols transmitted on the links and sent by the sources. Next discusses the relation between the security and the number of random code that every source needs to add, and gives the conditon which the number should meet. At last we introduce that sometimes a secure network code for a general wiretap network does not need randomness.