Evaluation of P2P Algorithms for Probabilistic Trust Inference in a Web of Trust

The problem of finding trust paths and estimating the trust one can place in a partner arises in various application areas, including virtual organisations, authentication systems and reputation systems. We study the use of peer-to-peer algorithms for finding trust paths and probabilistically assessing trust values in systems where trust is organised similar to the ‘web of trust’. We do this through discrete event simulation of random as well as scale free trust networks based on flooding as well as selective search algorithms. Our main conclusion is that in many situations these algorithms can be seen as belonging to a single class of algorithms that perform equally, and only differ through (and are sensitive to) parameter choices. We will also see that flooding is the only applicable method if one stresses the requirement for finding all trust paths, and if networks are less connected.