An incentive compatible and distributed reputation mechanism based on context similarity for service oriented systems

Abstract

One major challenge in service oriented systems is how to establish a trusted relationship between different parties. A reputation mechanism is the best-known method to build trust and foster mutual cooperation. Reputation mechanisms require correct and honest feedback to function effectively. An incentive compatible reputation mechanism entices self-interested parties to report the truth by explicit rewards which can offset the potential profits that could be gained by lying. The existing incentive compatible mechanisms are centralized which gather feedback merely through a specific service offered by a service provider, while they ignore some useful information on other services that are offered by the same service provider. In this paper, we introduce a novel distributed reputation mechanism based on a number of special reputation centers (SRCs) for service oriented environments. Each SRC gathers reputation information for a predetermined service offered by different service providers. The proposed model is collusion resistant and offers incentives to SRCs for making an attempt at gathering reputation information thoroughly and report it honestly. We then also evaluate the integrity and effectiveness of the proposed mechanism by carrying out experiments under different settings of the electronic marketplace. Examination of the results confirms that the new approach outperforms other incentive compatible reputation solutions reported in the literature.