The Study Of The Trust Based Access Control Model

With the progress of the society, the development of the network technology has demonstrated its strong momentum. Many users can get all kinds of the resource from the network. The problem of the performance bottleneck of the traditional Server/Client structure restricts the users'making full use of the resource. The concept of the server and client in the P2P structure has greatly weakened, which makes it become one of the most important applications. But the structure also provides the opportunity to attack the network for the malicious peers, which causes lots of security problems to the network.The characteristics of the P2P network decides the access control of the traditional Server/Client structure can't be simply applied to the network, because it is difficult to configure the traditional security approach. In this paper, we analyze the existing model, and combine the trust mechanism with the role based access control. In view of the malicious behavior of the network, we propose a trust-based access control model. The paper makes further research and discusses about the problems of the trust value, the division of the trust value interval and the policy of the access control, which mainly contains as follow:1. P2P network is dynamic and anonymous. And it is difficult to guarantee the quality of service of the peers in the network. In this paper, we analyze that the method of getting the direct trust value has some shortcomings. The punishment mechanism is introduced to the calculation of the direct trust value, and then the direct trust value can be get. Meanwhile, it is ready for getting the recommendation trust value.2. Because the P2P network is a complex system, we can get the direct trust value of two peers by the direct transactions. But when two peers don't have direct transactions, it is essential to ask other peers'recommendation. In this paper, we combine the evaluation method of direct trust value with the Fuzzy Petri Nets in order to obtain the recommendation trust value.3. The peer can join in the P2P network at any time, the role can be assigned according to its performance. The peer of different trust value will gain different role. The role of the peer is derived by its trust value interval range. In the paper the fuzzy clustering can be used to judge the factor that influences the trust value interval. And it can divide the trust value interval. The peer can get the authority according to the role.We can analyze from verification that the number of successful transactions must be greater than the number of unsuccessful transactions, otherwise the peer is not credible. When the unsuccessful transaction appears, the punishment mechanism makes the trust value drop sharply so that it restrains the malicious behavior. And when the peer provides good service, the trust value can rise. Fuzzy clustering can divide the trust interval.