Credible Reputation Systems for P2P e-Communities

Reputation mechanisms for distributed e-Communities are vital tools for facilitating trust decisions regarding transactions between entities. Motivated by the current challenges in the area of P2P reputation systems regarding their desgign, credibilty enhancement and objective evaluation, in this thesis we worked towards (1) creating a framework for the development and evaluation of secure reputation systems, and (2) designing and evaluating a credible reputation system for P2P communities with incentives for honest recommendations. We have thus created a conceptual model and a credibility framework for the design of credible reputation systems. We also proposed an evaluation framework for reputation systems for their objective evaluation and comparison. We then developed a credible reputation system (CREPARS) which consists of (a) credibility-enhanced reputation estimation algorithms and processes and (b) a novel recommendation exchange mechanism which is based on recommendation trustworthiness of entities and uses a PKI-based payment scheme. For the evaluation of the proposed reputation system we used credibility analysis and simulation in various attack scenarios and in comparison with other well-known reputation systems. The results have shown that the proposed reputation mechanisms exhibit resilience to various attacks and offer incentives for honest recommendations, leading to increased efficiency.