Secure Network Coding For Wiretapping Attack And Nodes Selfish Behavior

In recent years, network coding is a significant breakthrough in the field of information theory; it is also the hot pot technology of network transmission. With the development of network coding, security issues can not be ignored, these security issues limits the scope of application of network coding, hindered its application and promotion in practice. Existing security studies mainly divided into polluting and wiretapping. The recent secure network coding studies are aimed at these two security issues. So far, there are few results from the study of secure network coding.Existing secure network coding against wiretapping attacks are mainly from coding perspective. Coding complexity exchange with transmission security, therefore, often ignore the importance of the network topology to secure transmission. And for the node selfish behavior, traditional solutions have certain limitations, are not entirely suitable for network coding environment. This paper is based on network coding theory; the security model is weakly-secure model. To solves the problem of eavesdropping and node selfish behavior. On this basis, depending on the aggressive behavior, we design a variety of effective security network coding methods.This paper first studies secure network coding algorithm against wiretapping attacks based on route selection, for the eavesdropping attacks in network coding environment, proposed a secure network coding algorithm against wiretapping attacks based on route selection. From the perspective of the network topology structure, first studies unicast network. In the case of eavesdropping link(untrusted link position) location are known, Analysis the messages carries by eavesdropping links. Under the premise of maximum flow unchanged, remove the less eavesdropping links or normal links, to ensure that the Eavesdropper can not get complete source informations and the destination node can get complete informations. Then we re-structure system transfer matrix and get the secure network coding. The algorithm is then applied by unicast network extended to the multicast network applications. On the basis of this algorithm, an improved alogrithm is proposed; further improve the effectiveness of the algorithm. then we compare the advantages and disadvantages of secure coding method and secure topology method.The simulation experiments shows the initial method can effectively resist the network wiretapping attacks, and improved algorithms for performance has been further improved.Then we consider the wiretapping attacks from the perspective of network segmentation.For both cases, the eavesdropping nodes are known and unknown, two unicast security algorithms are proposed. In the case of eavesdropping node locations are known,Security algorithm core idea is: under the premise of maximum flow does not change, find the appropriate network segmentation method, the network is divided up into three sub-networks, and ensure that at least one sub-network eavesdropper has not been tapped before, so as to achieve the weak security target. In the case of eavesdropping node locations are unknown,Security algorithm core idea is: Under the premise of guaranteed n ?2m ?l( n is the number of eavesdropping nodes, m is the number of sub-networks are segmented, l is the number of sub-network which maximum flow is 1), we can find the right security algorithm. Then, the algorithms steps are given; by simulation experiments confirmed the validity of the two algorithms.Finally, node selfish behaviors in network coding environment were studied. We discussed the particularity of node slefish behavior in network coding environment, for the presence of a certain number of selfish nodes in the network, calculate the transmission failfure probability and maxflow(minimal cut) expected value,then recoding by the expected value.By means of this algorithm data transmission to source is most probably succeed. computer simulations show that although the proposed protocol causes a certain amount of redundancy waste, but it can avoid data transmission failed by selfish behaviors.