A Scalable Evidence Based Self-Managing Framework for Trust Management

Centrally managed, traditional security systems put limits on collaborative activities among huge number of entities in current open networks (such as Grids). This requires new approaches to handling security in large distributed systems and the need for new research especially in areas concerned with the provision of security through collaboration. This paper presents the design of a large-scale, self-managing Trust Management Framework (TMF) that makes efficient use of apparently invisible evidences that are scattered across potentially global networks. The TMF's design dictates a layered architecture for capturing evidence at the data layer of a network, transforming it into formed reputations in the information layer and utilizing these reputations to determine trustworthiness of an entity in the knowledge layer of the network. In essence, the main focus of the proposed work is to automate the acquisition of scattered evidence and the formulation, evolution and dissemination of reputations in a scalable way in order to make improved security decisions.