Interest-driven Reasoning Based on Commitment Alternating Temporal Logic in the P2P Reputation System

Abstract

P2P files sharing system is popular in network applications. The trust evaluation, designed for restricting the malicious peers in the system, is often based on reputation. When peers choose the trust nodes to download files in terms of the trust evaluation, they have established trust among themselves. Besides the trust, the utility (or interest) gained by peers is the other important factor during the process. In order to find out how the system to evolve based on real utility and trust, we introduce a logic language of game theory, CATL, into the P2P reputation system of files sharing. We propose some deductive rules of strategy reasoning in terms of the utility of the peers to look for the trend of the node selection. The new logic tool works well and we analyze two cases based on it. One case, namely, “advertisement effect”, blocks the newcomer and induces peers to collude with advocating reputation each other. The other case of “small-file-trend” constitutes a threat to the reputation system in Sybil Attack.