Science of Security

TRUST researchers are developing a science base for security, with hopes to ultimately leverage these views in revising course content and embodying this theory in tools for system developers.

Much computer security today is primarily reactive, deploying defenses for known attacks; it needs to become proactive, which is possible only if we can build systems in a principled way. A science of security would provide, for example, mental tools for understanding (a) how to expose trust assumptions intrinsic in a system design and how different defense mechanisms relocate trust assumptions in a system; (b) how to characterize security properties in a way that gives insight into enforcement mechanisms and verification approaches; (c) what classes of security properties can various classes of defenses support, and (d) what classes of attacks can various classes of defenses resist. The expectation is that this science can become a basis for an engineering discipline.