An axiomatic basis of trust in distributed systems

  • Authors:
  • P. Venkat Rangan

  • Affiliations:
  • EECS, University of California-Berkeley, Berkeley, Califomia

  • Venue:
  • SP'88 Proceedings of the 1988 IEEE conference on Security and privacy
  • Year:
  • 1988

Quantified Score

Hi-index 0.00

Visualization

Abstract

Computing systems are evolving into large distributed systems interconnecting organizations and even countries. The relationships among these organizations are characterized by the need for competition and cooperation without a common trusted agent. This paper presents an axiomatic theory of trust in distributed systems. The paper introduces basic notions about developing a logic or a theory, and shows that modal logics of belief, with their Kripke style possible-worlds semantics, are appropriate for basing a theory of trust on. We review a modal logic of belief, and construct a model of the distributed system so that the logic is sound and complete with respect to the model. Any sentences in the logic may then be added to the logic as axioms, and these axiomatic sentences are considered as trust specifications. We present methods to map a formal trust specification on to mechanisms for its implementation in the distributed system. Trust and security are closely related in distributed systems, Security specification and verification methods can be integrated into our theory of trust. We analyze the trusts required in public key based secure communication. A theory of trust has many applications such as, design of authentication servers based on trust relationships, and design of mechanisms that are tolerant to failures in trusts.