Abstract. Almost any cryptographic scheme can be described by tweakable polynomials over GF (2), which contain both secret variables (e.g., key bits) and public variables (e.g., plaintext bits or IV bits). The cryptanalyst is allowed to tweak the polynomials by choosing arbitrary values for the public variables, and his goal is to solve the resultant system of polynomial equations in terms of their common secret variables. In this paper we develop a new technique (called a cube attack) for solving such tweakable polynomials, which is a major improvement over several previously published attacks of the same type. For example, on the stream cipher Trivium with a reduced number of initialization rounds, the best previous attack (due to Fischer, Khazaei, and Meier) requires a barely practical complexity of 2 55 to attack 672 initialization rounds, whereas a cube attack can find the complete key of the same variant in 2 19 bit operations (which take less than a second on a single PC). Trivium with 735 initialization rounds (which could not be attacked by any previous technique) can now be broken with 2 30 bit operations, and by extrapolating our experimentally verified complexities for various sizes, we have reasons to believe that cube attacks will remain faster than exhaustive search even for 1024 initialization rounds. Whereas previous attacks were heuristic, had to be adapted to each cryptosystem, had no general complexity bounds,

Abstract. RSA-FDH and many other schemes secure in the Random-Oracle Model (ROM) require a hash function with output size larger than standard sizes. We show that the random-oracle instantiations proposed in the literature for such cases are weaker than a random oracle, including the proposals by Bellare and Rogaway from 1993 and 1996, and the ones implicit in IEEE P1363 and PKCS standards: for instance, we obtain a practical preimage attack on BR93 for 1024-bit digests (with complexity less than 2 30). Next, we study the security impact of hash function defects for ROM signatures. As an extreme case, we note that any hash collision would suffice to disclose the master key in the ID-based cryptosystem by Boneh et al. from FOCS ’07, and the secret key in the Rabin-Williams signature for which Bernstein proved tight security at EUROCRYPT ’08. We also remark that collisions can be found as a precomputation for any instantiation of the ROM, and this violates the security definition of the scheme in the standard model. Hence, this gives an example of a natural scheme that is proven secure in the ROM but that in insecure for any instantiation by a single function. Interestingly, for both of these schemes, a slight modification can prevent these attacks, while preserving the ROM security result. We give evidence that in the case of RSA and Rabin/Rabin-Williams, an appropriate PSS padding is more robust than all other paddings known. 1

Abstract. At the ECRYPT Hash Workshop 2007, Finiasz, Gaborit, and Sendrier proposed an improved version of a previous provably secure syndrome-based hash function. The main innovation of the new design is the use of a quasi-cyclic code in order to have a shorter description and to lower the memory usage. In this paper, we look at the security implications of using a quasi-cyclic code. We show that this very rich structure can be used to build a highly efficient attack: with most parameters, our collision attack is faster than the compression function! Key words: hash function, provable security, cryptanalysis, quasi-cyclic code, syndrome decoding. 1