Construction And Simulation Of P2P Trust Model Against Latent Attack Behavior

With the rapid development of computer network technology,in order to solve the traditional network structure in the bottleneck of computing power and storage space,the P2P network emerges.P2P is the direct exchange of computer resources and services through the direct exchange of personal computers,the sharing of resources and services include:storage space,processing ability,information content,and human use of these nodes behavior.Peer to peer network using personal computer and a network of connectivity,allowing a personal computer to contribute to the entire network to its respective ability.Since the emergence of P2P,the community’s views on the P2P is basically get both praise and blame,on the one hand,as a new network technology,P2P has been highly valued,known as "network revolution",and plans to actively carry out research of P2P technology and the technology used in resource sharing,instant communication,distributed computing,contraction engine and many other fields.On the other hand,P2P has many security issues such as copyright issues and traffic flooding problems,network security,file pollution problem,anonymous conmmunication,routing security problems.Therefore,the security problem is an important problem faced by P2P networks.At present for the security of transactions in the P2P network,and puts forward many trust management model,compared to the classical model is EigenTrust and PeerTrust,but existing trust models in resisting attacks just focus on in the nodes of once behavior,and can only deal with some special situations of malicious attacks,while ignoring the multiple intermittent attacks,This kind of behavior is similar to the APT attack,and it has the characteristics of pertinence,persistence,polymorphism and concealment,which is very harmful to the P2P network.Trust model proposed in this paper is on the basis of classical EigenTrust and PeerTrust,the introduction of the concept of the importance of resources to curb malicious nodes deliberately accumulated trust value of the speed.Recorded each transaction information and trust value,and through each node of the historical volatility calculated fluctuation intensity and fluctuation frequency to identify whether the node has latent attack characteristics,and the node global trust value is used as punishment mechanism.Finally,the simulation experiment will compared with the classical model and the improved model.The model can effectively identify potential attack nodes,making the calculated trust value more accurately.At the same time,it can turn malicious nodes to normal nodes and to make the model more safe and reliable.