Defense in Depth Formulation and Usage in Dynamic Access Control

Many network systems secure their resources using a defense in depth strategy, which can result in complex policies being distributed on the many access control points of a network. These policies are subject to frequent changes to deal with different factors such as change in security situation or change in resources. Moreover, while we have a vague intuitive understanding of the defense in depth strategy, we certainly lack a rigorous definition for it that would allow us to objectively assess whether a policy distribution on a network satisfies this strategy. In this paper, we propose a definition for defense in depth based on a notion of refinement given in product family algebra. We use this definition to articulate several implementations of the defense in depth strategy taking into account local access policies and global constraints on the resources of the considered network. We also discuss the automation of the calculations needed to derive the appropriate access policies to deploy at the nodes of a network.