| With the continuous development of network technology, the secure transmission of information is becoming increasingly important, relevant research has also been a growing concern of people. Establish reasonable trust model is a good way to solve information secure transmission between network peers. However, the traditional trust model can not solve the collusion of malicious peers attack. For these reasons, this paper proposes a peer attacks against collusion trust model. On the basis of the traditional trust model, this trust model is aimed at dynamic behavior and peers collusion attacks better.This paper first introduces the development of the analysis and the formation of the trust model, and presents the main attributes of trust. Then proposed an algorithm based on the recommended trust. The algorithm is introduced separately from overall model overview, updating trust and calculating trust. And we give formulas of trust.For the dynamic behavior of peers in the update phase of trust we propose increases the time frame, and thus update the short trust and history trust.Then analyzed the characteristic behaviors of collusion peers and highlighted the features of similarity and concealment .A method is found to detect collusion peers by the similarity of collusion peers. A fuzzy detector of collusion is proposed to detect the existence of colluding peers by measuring similarity of peers' behavior. And also a"similar" formal definition is built to detect whether the peers are collusion peers.When we find the collusion peers, the recommended trust is ignored to reduce the disadvantage. So the trust model proposed before is more consummate.Finally, the paper analyzes the function of system, designing and implementing each module. Then we make deployments towards system according to the experimental network, to verify the function of each module and advantage. In order to meet more demands of network security, the future research direction is proposed. |
PDF Full Text Request