Research On The Reputation Model Based On Double Trust Information In P2P Network

P2P technology makes the whole network world full of unselfish spirit and air of freedom. It crowds into various applications of internet with a flood speed, develops rapidly and becomes continuously perfect. But because of its anonymity and openness, P2P network has become the hothouse of aggressively spreading of incredible resources such as virus and mendacious resources, so the trust and content identification have become the key factor which obstructs the development of P2P technology. So it is necessary to build a trust mechanism to restrain the spread of malicious resources and encourage veracious transactions among nodes.On the basis of introducing the background, concept, current situation, topology and application scope of P2P technology, this paper summarizes the main security problems during the development of P2P technology. Through the detailed analyzing of the advantages and disadvantages of several typical mature reputation models, an improved distributed reputation model which is based on Double Trust Information is put forward in this paper.Aiming at the security problems such as"mutation"of the nodes'behavior, cheating across categories, the same measurement of the ability of nodes to provide service and send evaluation, slander and collusion, the corresponding solutions are as follows.①In the process of trust calculation, the proposed model takes the time sensitivity into consideration and gives a higher weight to the newest trade evaluation to ighlight its importance. So it can reduce the influence on the reputation model due to the mutation of nodes'behaviors.②Considering the trading differences, the resources on nodes are classified and the trust value of each kind of resources will be computed separately. So the behavior of cheating across the categories can be avoided.③The proposed model distinguishes the ability of nodes which provide services and give evaluations. It uses the service trust to indicate the reliability of service and the feedback trust to indicate the reliability of the evaluation. During the trust calculating process, the global trust value of service and feedback will be computed separately for each type of resources on each node. It can inhibit the problems which the single trust model can't resolve. For example, the nodes with high service trust send fake evaluation or the nodes with high feedback trust supply the malicious service.④This paper introduces a punishing mechanism in order to inhibit the malicious behaviors of malicious nodes such as slander and collusion and encourage the reliable trade among all the nodes.The simulation results show that the proposed model can reduce the global reputation value and trading probability of malicious nodes. It brings down the unfairness of irreproachable nodes.