An architectural approach to composing reputation-based distributed services

Abstract

Reputation-Based Trust (RBT) model with embedded incentive mechanisms provides an accurate quantitative measurement for services choosing their partners based on fair ratings accumulated from users. These mechanisms stimulate services to offer ratings truthfully, otherwise they lose their gains or even receive penalties. However, leveraging such mechanisms in distributed environments is a challenging task by its centralized nature. In this paper, we propose a new architecture development that combines relevant architectural components to make trust systems highly scalable with the auction mechanisms' capability to prevent lie. In this architecture we define an auction-based trust negotiation protocol that guides the interactions of distributed services and realize it in the distributed trust framework. Our architecture scales efficiently for increasing numbers of services interacting with the system, while still achieving protection against untruthful incentives even when a majority of ratings are unfair. An example of a supply chain is devised with empirical evidence collected. © 2010 Springer-Verlag Berlin Heidelberg.