PSPC · Provable Security for Physical Cryptography

Modern cryptographic security definitions do not capture real world adversaries who can attack the algorithm's physical implementation, as they do not take into account so called side-channel attacks where the adversary learns information about the internal state of the cryptosystem during execution, for example by measuring the running time or the power consumption of a smart-card.Current research on side-channels security resembles a cat and mouse game. New attacks are discovered, and then heuristic countermeasures are proposed to prevent this particular new attacks. This is fundamentally different from the ""provable security"" approach followed by modern cryptography, where one requires that a cryptosystem is proven secure against all adversaries in a broad and well-defined attack scenario. Clearly, this situation is unsatisfactory: what is provable security good for, if ultimately the security of a cryptosystem hinges on some ad-hoc side-channel countermeasure?Despite this, until recently the theory community did not give much attention to this problem as it was believed that side-channels are a practical problem, and theory can only be of limited use to prevent them. But recently results indicate that this view is much too pessimistic.On a high level, the goal of this project is to bring research on side-channels from the realm of engineering and security research to modern cryptography. One aspect of this proposal it to further investigate the framework of leakage-resilience which adapts the methodology of provable security to the physical world. If a cryptosystem is leakage-resilient, then this implies that its implementation is secure against every side-channel attack, making only some mild (basically minimal) assumptions on the underlying hardware.""