Computer-aided decision-making with trust relations and trust domains (cryptographic applications)*

We propose generic declarative definitions of individual and collective trust relations between interacting agents and agent collections, and trust domains of trust-related agents in distributed systems. Our definitions yield (1) (in)compatibility, implicational and transitivity results for trust relationships, including a Datalog-implementability result for their logical structure; (2) computational complexity results for deciding potential and actual trust relationships and membership in trust domains; (3) a positive (negative) compositionality result for strong (weak) trust domains; (4) a computational design pattern for building up strong trust domains; and (5) a negative scalability result for trust domains in general. We instantiate our generic trust concepts in five major cryptographic applications of trust, namely: Access Control, Trusted Third Parties, the Web of Trust, Public-Key Infrastructures and Identity-Based Cryptography. We also show that accountability induces trust. Our defining principle for weak and strong trust (domains) is (common) belief in and (common) knowledge of agent correctness, respectively.