An Architectural Approach to Composing Reputation-Based Trustworthy Services

In SOA, Reputation-Based Trust (RBT) mechanism is applied to achieve trust management. RBT enables services to assess the trust level of other services based on the reputation accumulated from user recommendations. A key challenge to apply RBT is to prevent the strategic behavior of users when they provide recommendations -- they might give unfair ratings to benefit themselves. In this paper, we propose a novel architectural approach to integrating auction mechanisms into the trust framework to prevent benefits from untruthful incentives. In this architecture we define an auction-based trust negotiation protocol and realize it in the trust framework. The contribution of our architecture is that it scales and produces accurate results to achieve protection against untruthful incentives, especially when a majority of ratings are unfair, without the potential increase in a computation overhead. An example on a travel agent scenario is devised to collect empirical evidence.