There are two schools of thought around threat modeling. One school advocates the creation of attack trees and data flow diagrams. This requires extensive, cross-functional, security skills and is not a scalable approach. The other school encourages organic insertion of defenses based only on current context without “boiling the ocean”. This lack of systems thinking leaves applications vulnerable as exploits in a weaker component can open the door to critical systems.
Part of the problem is threat modeling today is largely an art. We need to inject more science in this domain and derive a repeatable and auditable approach that maps to risk. Such a model should abstract away the non-scalable elements and still provide a high degree of assurance in today’s faster velocity business context.
This presentation will outline a threat modeling framework that abstracts traditional methods into systems, data, and people components. You will come away with an approach that takes away some of the scalability problems of traditional threat modeling, yet provides sufficient rigor and systems thinking to help manage risk.