Leakage-resilient constructions have attracted significant attention over the last couple of years. In practice, pseudorandom functions are among the most important such primitives, because they are stateless and do not require a secure initialization as, e.g. stream ciphers. However, their deployment in actual applications is still limited by security and efficiency concerns. This paper contributes to solve these issues in two directions. On the one hand, we highlight that the condition of bounded data complexity, that is guaranteed by previous leakage-resilient constructions, may not be enough to obtain practical security. We show experimentally that, if implemented in an 8-bit microcontroller, such constructions can actually be broken. On the other hand, we present tweaks for tree-based leakage-resilient PRFs that improve their efficiency and their security, by taking advantage of parallel implementations. Our security analyses are based on worst-case attacks in a noise-free setting and suggest that under reasonable assumptions, the side-channel resistance of our construction grows super-exponentially with a security parameter that corresponds to the degree of parallelism of the implementation. In addition, it exhibits that standard DPA attacks are not the most relevant tool for evaluating such leakage-resilient constructions and may lead to overestimated security. As a consequence, we investigate more sophisticated tools based on lattice reduction, which turn out to be powerful in the physical cryptanalysis of these primitives. Eventually, we put forward that the AES is not perfectly suited for integration in a leakage-resilient design. This observation raises interesting challenges for developing block ciphers with better properties regarding leakage-resilience. © 2012 International Association for Cryptologic Research.
CITATION STYLE
Medwed, M., Standaert, F. X., & Joux, A. (2012). Towards super-exponential side-channel security with efficient leakage-resilient PRFs. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7428 LNCS, pp. 193–212). https://doi.org/10.1007/978-3-642-33027-8_12
Mendeley helps you to discover research relevant for your work.