Using Recommendations for Managing Trust in Distributed Systems

Each time we carry out vital communication in any distributed computer system such as the Internet, we face an inherent risk. This risk arises because we can never be completely certain about the trustworthiness of entities that mediate our on-line interactions. To minimise this risk, users must be given the chance to assess trust on the network, and be given an opportunity to pick an option with the least level of perceived risk. In this paper, we explain why traditional network security mechanisms are incomplete in their function to manage trust, and provide a general model based on recommendations.