P2p Network Security Model Based On Reputation Mechanisms

P2P network provides a kind of environment for open and free files-exchanging and for resources-sharing. With the promotion of the technology of P2P, the internet model of information storing will develop from the present model of "centralization" to the one of "marginalization". Meanwhile, the nodes of P2P network from various organizations and clients can join or leave arbitrarily and freely, and in the network environment lacking a system of inspiration and punishment, no natural trust among the nodes results in more complaint, greed and cheating. Consequently, ensuring P2P network's security becomes one of the biggest challenges, while the effective way to solve such a security problem is to set up a fine distributional trust system.Trust relation among the nodes is the most basic and nuclear security needs, and the key for safely applicating of P2P network. As a result, it is necessary to take into account how to build up a well-operational and effective trust system and model suited for P2P network environments. As a consequence, by analyzing the pioneers' experience and previous researches, and by fully thinking of a series of evaluation factors, this research paper discusses "the trust evaluation model based on trust level", with the aim of encouraging network resources sharing, of improving the security of P2P network, of decreasing the risk of trading , and of prompting advantageous development of network. This paper, through studying the quantification, integrity and transmission of trust, design an array of mathematic models involving the trust evaluation system. The import of related factors achieves the goal of setting up a system of urging eximious nodes and of punishing and isolating malicious or mediocre and lazy ones. The using of improved strategies in choosing service nodes, to some extent, can balance the network load, and construct the model to control safely visiting among the nodes of P2P network. The import of time factors makes nodes concern about trust in the long run, and to pay more attention to recent behaviors. The mechanism of good friends and black list saves the time of trust computing and nodes choosing, decreases the network load and delay, and improves the model's efficiency. Finally, the model involves simulated test and capability analyzing, which tentatively testifies the rationalization and effectiveness of this model.