Reaearch On Trust Model And Search Technology In P2P Network

P2P (Peer-to-Peer) network is a new kind of network that had obtained wildly application. P2P changes the mode of tradition network named C/S (Client/Server), demands logical reciprocal between peers, every peer in network can play two role both as client that acquires services and as server that provides services, files can transport directly between two peers without the help of center server.Peers can add in and off network as one pleases own to that P2P network's essential attribute named anonymity and independent, some peers upload fake resource, computer virus deliberately under the situation without management in order to disturb system operation. The measure that pull trust model in P2P network can reduce the number of malicious peers and make peers'behavior standard, improve the availability of P2P network. How to search the needing resource fast and accurate from share files is a research focus for P2P technology.Author analysis features and defect about these model existed, propose a trust model based on super node named TSN(Trust Super Node). This model improves these algorithms about joining of node, trust degree of peers between groups, Peers'similarity and response delay are considered to trust calculation. TSN model weights for direct trust and indirect trust are dynamically adjusted. Strategies are considered for model's robustness. The simulation shows that propose model has advantages in successful request and system security over the exiting P2P trust models.Author analysis features and defect about these searching algorithm existed, propose a new algorithm named SAT(Search Algorithm based on Trust Model), this algorithm improves the query algorithm combining forward and response algorithm under the TSN trust model, the share resource cache and no-echo forward mechanism are considered to reduce search delay and make response message from the high trust value which can provide real resource. The simulation shows that propose have advantages in successful request and search delay over the Random-walk algorithm.