Søren S. Thomsen

PRINCE: a low-latency block cipher for pervasive computing applications

This paper presents a block cipher that is optimized with respect to latency when implemented in hardware. Such ciphers are desirable for many future pervasive applications with real-time security needs. Our cipher, named PRINCE, allows encryption of data within one clock cycle with a very competitive chip area compared to known solutions. The fully unrolled fashion in which such algorithms need to be implemented calls for innovative design choices. The number of rounds must be moderate and rounds must have short delays in hardware. At the same time, the traditional need that a cipher has to be iterative with very similar round functions disappears, an observation that increases the design space for the algorithm. An important further requirement is that realizing decryption and encryption results in minimum additional costs. PRINCE is designed in such a way that the overhead for decryption on top of encryption is negligible. More precisely for our cipher it holds that decryption for one key corresponds to encryption with a related key. This property we refer to as α-reflection is of independent interest and we prove its soundness against generic attacks.

The Rebound Attack: Cryptanalysis of Reduced Whirlpool and Grøstl

In this work, we propose the rebound attack, a new tool for the cryptanalysis of hash functions. The idea of the rebound attack is to use the available degrees of freedom in a collision attack to efficiently bypass the low probability parts of a differential trail. The rebound attack consists of an inbound phase with a match-in-the-middle part to exploit the available degrees of freedom, and a subsequent probabilistic outbound phase. Especially on AES based hash functions, the rebound attack leads to new attacks for a surprisingly high number of rounds. We use the rebound attack to construct collisions for 4.5 rounds of the 512-bit hash function Whirlpool with a complexity of 2120 compression function evaluations and negligible memory requirements. The attack can be extended to a near-collision on 7.5 rounds of the compression function of Whirlpool and 8.5 rounds of the similar hash function Maelstrom. Additionally, we apply the rebound attack to the SHA-3 submission Grostl, which leads to an attack on 6 rounds of the Grostl-256 compression function with a complexity of 2120 and memory requirements of about 264.

The Grindahl hash functions

In this paper we propose the Grindahl hash functions, which are based on components of the Rijndael algorithm. To make collision search sufficiently difficult, this design has the important feature that no low-weight characteristics form collisions, and at the same time it limits access to the state. We propose two concrete hash functions, Grindahl- 256 and Grindahl-512 with claimed security levels with respect to collision, preimage and second preimage attacks of 2128 and 2256, respectively. Both proposals have lower memory requirements than other hash functions at comparable speeds and security levels.

Rebound attacks on the reduced grøstl hash function

Grostl is one of 14 second round candidates of the NIST SHA-3 competition. Cryptanalytic results on the wide-pipe compression function of Grostl-256 have already been published. However, little is known about the hash function, arguably a much more interesting cryptanalytic setting. Also, Grostl-512 has not been analyzed yet. In this paper, we show the first cryptanalytic attacks on reduced-round versions of the Grostl hash functions. These results are obtained by several extensions of the rebound attack. We present a collision attack on 4/10 rounds of the Grostl-256 hash function and 5/14 rounds of the Grostl-512 hash functions. Additionally, we give the best collision attack for reduced-round (7/10 and 7/14) versions of the compression function of Grostl-256 and Grostl-512.

Cryptanalysis of PRESENT-like ciphers with secret S-boxes

At Eurocrypt 2001, Biryukov and Shamir investigated the security of AES-like ciphers where the substitutions and affine transformations are all key-dependent and successfully cryptanalysed two and a half rounds. This paper considers PRESENT-like ciphers in a similar manner. We focus on the settings where the S-boxes are key dependent, and repeated for every round. We break one particular variant which was proposed in 2009 with practical complexity in a chosen plaintext/chosen ciphertext scenario. Extrapolating these results suggests that up to 28 rounds of such ciphers can be broken. Furthermore, we outline how our attack strategy can be applied to an extreme case where the S-boxes are chosen uniformly at random for each round and where the bit permutation is secret as well.

Slender-Set Differential Cryptanalysis

This paper considers PRESENT-like ciphers with key-dependent S-boxes. We focus on the setting where the same selection of S-boxes is used in every round. One particular variant with 16 rounds, proposed in 2009, is broken in practice in a chosen plaintext/chosen ciphertext scenario. Extrapolating these results suggests that up to 28 rounds of such ciphers can be broken. Furthermore, we outline how our attack strategy can be applied to an extreme case where the S-boxes are chosen uniformly at random for each round, and where the bit permutation is key-dependent as well.

Cryptanalysis of MD2

This paper considers the hash function MD2 which was developed by Ron Rivest in 1989. Despite its age, MD2 has withstood cryptanalytic attacks until recently. This paper contains the state-of-the-art cryptanalytic results on MD2, in particular collision and preimage attacks on the full hash function, the latter having complexity 273, which should be compared to a brute-force attack of complexity 2128.

Practical near-collisions on the compression function of BMW

Blue Midnight Wish (BMW) is one of the fastest SHA-3 candidates in the second round of the competition. In this paper we study the compression function of BMW and we obtain practical partial collisions in the case of BMW-256: we show a pair of inputs so that 300 pre-specified bits of the outputs collide (out of 512 bits). Our attack requires about 232 evaluations of the compression function. The attack can also be considered as a near-collision attack: we give an input pair with only 122 active bits in the output, while generic algorithm would require 255 operations for the same result. A similar attack can be developed for BMW-512, which will gives message pairs with around 600 colliding bits for a cost of 264. This analysis does not affect the security of the iterated hash function, but it shows that the compression function is far from ideal. We also describe some tools for the analysis of systems of additions and rotations, which are used in our attack, and which can be useful for the analysis of other systems.

Pseudo-cryptanalysis of the original blue midnight wish

The hash function Blue Midnight Wish (BMW) is a candidate in the SHA-3 competition organized by the U.S. National Institute of Standards and Technology (NIST). BMW was selected for the second round of the competition, but the algorithm was tweaked in a number of ways. In this paper we describe cryptanalysis on the original version of BMW, as submitted to the SHA-3 competition in October 2008. The attacks described are (near-)collision, preimage and second preimage attacks on the BMW compression function. These attacks can also be described as pseudo-attacks on the full hash function, i.e., as attacks in which the adversary is allowed to choose the initial value of the hash function. The complexities of the attacks are about 214 for the near-collision attack, about 23n/8+1 for the pseudo-collision attack, and about 23n/4+1 for the pseudo-(second) preimage attack, where n is the output length of the hash function. Memory requirements are negligible. Moreover, the attacks are not (or only moderately) affected by the choice of security parameter for BMW.

Deterministic differential properties of the compression function of BMW

In this paper, we give some determinstic differential properties for the compression function of SHA-3 candidate Blue Midnight Wish (tweaked version for round 2). The computational complexity is about 20 compression function calls. This applies to security parameters 0/16, 1/15, and 2/14. The efficient differentials can be used to find pseudo-preimages of the compression function with marginal gain over brute force. However, none of these attacks threaten the security of the BMW hash functions.