A Trust Mechanidm For P2P E-commerce Systems

According to the statistics by CNNIC and DCCI, the users of network are growing rapidly in recent years, as well as the application of the internet, especially for e-commerce. E-commerce has brought us convience as well as problems like trust issues. The expectation of the development of e-commerce in China is illustrated at the begaining of this paper, which is followed by the instancy trust issues, to show study on e-commerce trust mechanism is signality and essential.There are two main types of trust mechanism, one is cenralised reputation and the other is distributed reputation system. This paper is focused on distributed reputation system. Peer-to-peer(P2P) networks represent an environment well suited for distriuted reputation management.The basic trust mechanism used in this paper is one of the distributed reputation systems. The study of trust mechanism is based on P2P e-commerce system.In this paper, a trust evaluation mechanism of P2P e-commerce system is described detailedly. In order to show the importance of the trust mechanism to P2P e-commerce systems, the special natures of P2P e-commerce systems and how trust mechanisms impact it are briefly introduced first. Then is the review of related research, with a conclusion that most of them are not comprehensive enough especially on the factors selecting, and that they are not strong enough to resist variety of malicious acts.Based on analysis and comparison of the existing work, and verified by simulation, a reputation-based trust mechanism for P2P e-commerce systems is chosen because of its better performance. However, the mechanism still has some shortages. By studying further, a model which is more accurate and with stronger ability to anti attacks from malicious peers is presented in this paper. In the mechanism, a peer has two kinds of reputations, namely local reputations and global reputations. The local reputation of a peer relative to another peer is calculated in terms of the reference peer's rating of the transaction between the two peers, whereas the global reputation is computed based on all peers'rating of the transaction between them. Local reputation is further divided into local reputation of sellers and local reputation of buyers, and the global reputation is also been further divided into global reputation of sellers and global reputation of buyers. To compute the trust value of sellers and buyers, the concept of belief factor is introduced to integrate the local reputation with the global reputation. So the model can avoid peers cover up malicious sales by honest purchases. To improve the accuracy and anti-attack capability of the model, a filtering method based on the iterative dichotomy for aggregate and more comprehensive trust factors are introduced.Finally, the effectiveness and resistibility of the proposed trust model are evaluated by theoretical analysis and simulation experiments. The experimental results show that the proposed trust evaluation model outperforms existing models, it can convergence the trust value to a fixed value sooner and also more accurate and effective to resist malicious attacks.