This research project focuses on the development of cryptographic mathematical models and constructions that address realistic security requirements at the implementation level. This is a fundamental problem as cryptographic security formalisms are often criticized for lack of relevance given the wide range of attacks available at the implementation level. Indeed, traditional cryptographic attacks are restricted in the way private data can be accessed; hence, the security of systems relying on such constructs is contingent on external non-cryptographic means for enforcing the necessary tamper resilience. Unfortunately, this physical tamper resistance is either too expensive or unreliable. The research extends models of cryptographic attacks to include various forms of private data tampering and access and brings the theory of cryptographic constructions closer to security concerns in practice. In particular, the tamper proofing of a wide set of cryptographic primitives is considered in an extended adversarial setting, such as digital signatures, public key encryption, secure function evaluation, as well as arbitrary cryptographic functions. This research thus explores the boundaries of what is achievable algorithmically and practically through cryptographic means.
