A P2P File-Sharing Model Based On Global Trust

With the rapid development of internet and P2P technology, new applications based on P2P technology have been emerged one after another, particularly, file sharing service based on P2P has become the most widely used application based on P2P.Lacking of efficient regulation and incentive mechanism for file sharing service based on P2P due to its openness and non-centricity is one of the greatest challenges faced by the application. One way to solve that challenge is to design reasonable, efficient global-trust model.This essay is aimed to develop a P2P file sharing model based on global-trust by defining model framework and modules, and in the newly-proposed model, global flow contribution, a new measurement for P2P file sharing regulation and incitement, has been designed on the basis of global trust measurement. Global trust measurement is deemed as a global measurement for quality of file sharing service at each node, and it represents the contribution of specific node to upload flow of all nodes. P2P file sharing regulation and incentive mechanism can be easily realized in the model defined in this essay by designing service discrimination mechanism taking global trust and global flow contribution as standards. The file sharing model designed in this essay supports decentralized algorithm and storage of global trust, and decentralized algorithm has been introduced in great detail. Given the possible loss of storage trust due to dropping out of nodes in practical application, user indexing mechanism and trust redundancy storage mechanism have been designed in the essay, in order to enhance the robustness of global trust storage.With imitation experiment, the global trust model designed here is proved to be efficient and feasible in terms of sharing service quality reliability guarantee and service discrimination mechanism efficiency. At the end, redundancy storage mechanism is proved to be more robust and more suitable for practical application compared to single-node storage mechanism after estimation to loss probability of trust for both mechanisms.