You are looking at a specific version 20130817:054233 of this paper. See the latest version.

Paper 2013/381

Breaking the Even-Mansour Hash Function: Collision and Preimage Attacks on JH and Grøstl

Bingke Ma and Bao Li and Ronglin Hao

Abstract

The Even-Mansour structure and the chopMD mode are two widely-used strategies in hash function designs. They are adopted by many hash functions including two SHA-3 finalists, the JH hash function and the Grøstl hash function. The Even-Mansour structure combining the chopMD mode is supposed to enhance the security of hash functions against collision and preimage attacks, while our results show that it is not possible to achieve this goal with an unbalanced compression function. In this paper, we show generic attacks on the Even-Mansour hash functions including both collision and preimage attacks. Our attacks show the structure flaws of the Even-Mansour hash functions. All these attacks can be applied to specific hash functions based on the Even-Mansour structure. We achieve the first collision and (2nd-)preimage attacks on full JH and Grøstl respectively. For the JH hash function, we achieve collision and (2nd-)preimage attacks on the full JH compression function with a time gain $2^{10.22}$. After a simple modification of the padding rules, we obtain full round collision and (2nd-)preimage attacks on the modified JH hash function with a time gain $2^{10.22}$. For the Grøstl hash function, we obtain both collision and (2nd-)preimage attacks on the full Grøstl hash function with a limited time gain $2^{0.58}$.

Metadata
Available format(s)
-- withdrawn --
Publication info
Preprint. MINOR revision.Not published
Keywords
Even-Mansour hash functionchopMD modepreimagecollisionJHGrøstl
Contact author(s)
bkma @ is ac cn
History
2013-08-17: withdrawn
2013-06-17: received
See all versions
Short URL
https://ia.cr/2013/381
License
Creative Commons Attribution
CC BY
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.