Skip to Main Content
Automated trust negotiation is an approach to establish trust in pervasive computing environment, where the involved participators belong to different security domains and need to establish trust before interactions can take place. However, during the trust negotiation, onepsilas private information can be easily exploited to infer through observing onepsilas behavior, even if access control policies are strictly enforced. In order to prevent such inference, we propose C-ATN, an ATN model based on the credibility of credentials. We propose the access threshold to take place of the access control policies. Compared to the previously proposed approaches, malicious observer can dig out nothing unless the corresponding policies are satisfied. After we demonstrate that finding the safest reply is an NP-Complete problem, we still provide polynomial-time arithmetic to find out the near-optimal solution, and in some cases it is the optimal solution.