Distributed approximation of fixed-points in trust structures

Abstract

We consider distributed algorithms for solving a range of problems in a framework for trust in large-scale distributed systems. The framework is based on the notion of trust structures; a set of 'trust-levels' with two distinct partial orderings. In the trust model, a global trust-state is defined as the least fixed-point of a collection of local policies of nodes in the network. We show that it is possible to compute the global trust-state using a simple, robust and totally asynchronous distributed-algorithm. We also consider a distributed notion of proof-carrying-requests as a means of approximating the least fixed-point, enabling sound reasoning about the global trust-state without computing the exact fixed-point. Our proof-carrying-request model is different than the notion of proof-of-compliance from traditional trust-management; in particular, all proofs are efficiently verifiable or easily rejected, but, in the worst case, may require as much communication as computing the actual trust-state itself. © 2005 IEEE.