Multi-domain Trust Management in variable-threat environments - a user-centric model

Trust Management (TM) systems are trust infrastructures that support authorization for security-critical actions in decentralized environments. In this paper we present a user-centric view to address trust management as it impacts the unanticipated user and/or user behavior for multi-domain applications. This protection can be tuned to deal with users who may be responsible for an elevated threat level, and builds upon a resource-centric architecture. Our model is suitable for variable-threat environments and allows for temporary adjustments of trust levels. The expectation is to enable a Trust Management Agent to determine appropriateness of the unanticipated user or behavior, and reverse restrictions without compromising actions that took place during such periods - we term this, rollback-access. We argue that a rollback-access capability is an essential feature for security-critical applications, and is appropriate for today's military and intelligence community coalitions as they execute their particular missions in the Global War on Terrorism.