The Research Of Reputation-based Trust Model In P2P Network

With the rapid development of the internet, P2P(Peer-to-Peer) network as a distributed computing architecture entered our vision and life, and has attracted more research and attention.The application of the P2P develops vigorously, it also brings some security issues, the biggest security issue is the mutual-trust between peers. Due to the P2P network's open,, dynamic and anonymous properties, there is no sense of trust between peers, and in the network environment lacking a mechanism of punishment and incentive, misbehaviors of malicious peers are emerging in endless, and,the performance of the network is severely affected. So, how to restrain and stamp out the misbehavior of malicious peers, and encourage peers to work well in the network is the content of this paper——P2P Trust Model's main purpose. The decentralized characteristic of the P2P model, making centralized trust model is not suit for P2P network environment, so we must build the P2P trust model in the distributed environment.Based on the solutions of the typical reputation-based trust model TBRM (Trust-Based Reputation Model), this paper presents an improved reputation-based trust model, and designs the solutions for, he trust model including the evaluation and the calculation of the reputation, dynamic adaptability, incentive mechanism, and the distributed implementation strategies.This trust model combines the TBRM's advantages in the aspects of the reputation evaluation and the reputation calculation algorithm. It can distinguish dishonest feedback, and can compute the peer's integrated reputation factually and effectively. Aim at the existing TBRM lacks dynamic adaptability, we introduce the mechanism to suppress the malicious peers'oscillate behavior when provide feedback evaluation, it can make the oscillate peer can not perform evil in the process of providing feedback evaluation, and to most important, it increased the model's dynamic adaptability. We introduce the integral mechanism as incentive mechanism, it can isolate the selfish peers and encourage peers to contribute, thus guarantee higher quality sevice. We design the distributed store strategies of this trust model by Chord, it can guarantee the security of store and access of trust data.The results of the model simulation show that, compared with the TBRM model, the improved model can suppress the malicious peers'feedback oscillating behavior more effectively. The introduce of the integral mechanism not only can isolate free-riding peers effectively and movivate the peer's enthusiasm in uploading files, but also can encourage them to provide high quality files.