An adaptive service strategy based on user rating in P2P

Abstract

In order to deal with free riding in P2P system, incentive mechanism or rating system is presented, and each user rating is computed by itself, which causes that some users exaggerate their ratings. In this paper, three aspects are developed for free riding and for improving service performance in P2P. First of all, it is useful to keep away fake rating that each user rating is calculated by the responders. Secondly, three kinds of user ratings including single user rating and two group ratings are exploited to get better service performance. The last aspect is to provide a service strategy that is adaptive to the responder workload. Finally, the experimental results demonstrate that this strategy can effectively handle the problem of free riding, and especially, group ratings can greatly boost response rate of the whole P2P system. © Springer-Verlag Berlin Heidelberg 2005.