PKC-Sec: Security Analysis of Classical and Post-Quantum Public Key Cryptography Assumptions

点击翻译按钮获取中文摘要

Public key cryptography (PKC) depends on the existence of computational problems that are incredibly hard - but not impossible - to solve. Classical examples that were fundamental to the origins of PKC in the 1970s (and indeed were prominent centuries earlier) are the integer factorisation problem and the discrete logarithm problem (DLP). While there are no known efficient, i.e., polynomial-time algorithms for solving these problems that run on classical computers, thanks to Shor's astounding breakthrough ideas in 1994, both can be solved efficiently on a quantum computer of sufficient size. Intense research in the areas of quantum computation, quantum information theory and quantum algorithms ensued, and replacement post-quantum (PQ) cryptosystems have been studied in earnest for the past 15 years or so, with standardisation efforts in process by both NIST and ETSI. PQ cryptosystems must be secure against both classical and quantum computers and therefore their underlying hardness assumptions must be studied intensely before they can be fully trusted to replace our existing PKC hardness assumptions. Until these standards have been established and cryptographic practice migrates entirely to PQ cryptography, it is also essential that the study of classical hardness assumptions persists, particularly as sporadic and sometimes spectacular progress can occur: for instance, for a special but large family of finite fields the DLP can be solved on a classical computer in quasi-polynomial time, i.e., `very nearly' efficiently, thanks to a series of results due to Dr. Granger and his collaborators, and Joux and his collaborators.In this project we will research and develop algorithms for solving computational problems that are foundational to the security of PKC, both now and in the future. In particular, we will study: the DLP in the aforementioned special family of finite fields, for which an efficient classical algorithm is potentially on the horizon; the security of the Legendre pseudo-random function, which is extremely well suited for multi-party computation and has been proposed for use in the next iteration of Ethereum - the de facto standard blockchain platform - but is not so well-studied; and finally the security of supersingular isogeny-based PQ cryptography, which although a relatively young field offers many very promising applications. Due to their nature, any cryptographic assumptions based on mathematical constructions are potentially weaker than currently believed, and we will deepen our understanding and assess the hardness of these natural and fundamental problems, thus providing security assurances to the cryptography community and more generally all users of cryptography.