Distributed Approximation of Fixed-Points in Trust Structures

Abstract

Recently, Carbone, Nielsen and Sassone introduced the trust-structure framework; a semantic model for trust-management in global-scale distributed systems. The framework is based on the notion of trust structures; a set of ``trust-levels'' ordered by two distinct partial orderings. In the model, a unique global trust-state is defined as the least fixed-point of a collection of local policies assigning trust-levels to the entities of the system. However, the framework is a purely denotational model: it gives precise meaning to the global trust-state of a system, but without specifying a way to compute this abstract mathematical object.

This paper complements q the denotational model of trust structures with operational techniques. It is shown how the least fixed-point can be computed using a simple, totally-asynchronous distributed algorithm. Two efficient protocols for approximating the least fixed-point are provided, enabling sound reasoning about the global trust-state without computing the exact fixed-point. Finally, dynamic algorithms are presented for safe reuse of information between computations, in face of dynamic trust-policy updates.