A Framework for System Security

Abstract

Actors in our general framework for secure systems can exert four types of control over other actors’ systems, depending on the temporality (prospective vs. retrospective) of the control and on the power relationship (hierarchical vs. peering) between the actors. We make clear distinctions between security, functionality, trust, and distrust by identifying two orthogonal properties: feedback and assessment. We distinguish four types of system requirements using two more orthogonal properties: strictness and activity. We use our terminology to describe specialized types of secure systems such as access control systems, Clark–Wilson systems, and the Collaboration Oriented Architecture recently proposed by The Jericho Forum.