| Peer-to-Peer Network is a new kind of network which is designed to fully use single point's ability with it's self-organization and distribution. It does well in distributed computing, distributed queries, distributed storage and file sharing. It makes massive information spreading freely in network without servers. Because the system is open, independent and reciprocity, some bad behaviors which like forging resources and free-rider are spreading unchecked in network. For the system are running in high quality we must prevent cheating behavior, free-rider and some other bad behaviors.Among dealing with those bad behaviors, the model based on feedback mechanism of reputation is well developed. This kind of mode also has some problems such as low efficiency and lack of confidence. A new model based on reputation and Ant Colony System well solved those problems. This model does well in following aspects. First, usual secure methods is hardly used in P2P Network for it's high dynamic. Ant Colony System does well in dealing with dynamic situation. We make use of improved ACS in querying recommended paths in P2P Network. In the model ant choose the path with more pheromone leading to the server providing required resources. After the client downloaded the resources, it evaluate the server with it's work and the evaluation will change the pheromone of the path leading to the server. That will make the better server providing more services. Second, to get comprehensive information of the server we need to know the server's another aspect which is reputation. Reputation is other point's evaluation to this point. The point's reputation represents other's historical evaluation to it, so it is this point's historical behaviors.Finally, we make the model out. Compared the efficiency and security of this model with other model which base on ACS, it performs very well in simulation experiment. It prevent malicious nodes and polluted file spread in the network to make it running in high quality. |
PDF Full Text Request