All modern security systems are based on the fundamental assumption that the electronic hardware on which they are implemented is trustworthy. Users require that their computers on which they do online banking are trustworthy, and that the hardware running the braking system of their car, or a medical implant cannot be attacked. Unfortunately, countless real-world examples illustrate that attacks targeting weaknesses at the implementation-level are often rather easy to carry out.
For instance, so-called side-channel attacks break cryptographic implementations by exploiting physical phenomena such as the power consumption of the device or introducing faults into the computation. Other important examples include attacks due to bad randomness, viruses running inside a physical system, or in the extreme case, an adversary that embeds backdoors (so-called hardware trojans) into integrated circuits. Traditionally, most defenses against these attacks work at the system-level, and come without formal security analysis. The goal of our research is to develop new foundations for designing provably secure countermeasures against these emerging attacks. To this end, we combine techniques from modern cryptography with recent advances in hardware engineering and secure programming languages.
Besides developing the mathematical foundations for these new threat models, we often also test the practical feasibility (and resistance) of our new countermeasures with prototype implementations and by attacking implementations in practice.