Paper 2017/188

Division Cryptanalysis of Block Ciphers with a Binary Diffusion Layer

Wenying Zhang and Vincent Rijmen

Abstract

In this paper, we propose an accurate security evaluation methodology for block ciphers with a binary diffusion layers against division cryptanalysis. We illustrate the division property by the independence of variables, and exploit a one-to-one mapping between division trails and invertible sub-matrices. We give a new way to model the propagation of division property of linear diffusion layers by the smallest amount of inequalities which are generated from linear combinations of row vectors of the diffusion matrix. The solutions of these inequalities are exactly the division trails of linear transformation. Hence the description is compact and optimal. As applications of our methodology, we first present a 10-round integral distinguisher for Skinny, proposed at CRYPTO 2016 which is of one round more than that found by using the previous method. For Midori, proposed at ASIACRYPT 2015, the designers have obtained a 3.5-round integral characteristic. Surprisingly, we find 7-round integral distinguishers both for Midori64 and Midori128. Most importantly, we obtain the longest integral distinguishers for block ciphers with a binary diffusion layer. It seems that any more improvement of this kind of integral distinguishers using the division property is impossible. Therefore, the technique can be used to prove security against division cryptanalysis, and we can hopefully expect it to become a useful technique for designers.

Metadata
Available format(s)
PDF
Publication info
Preprint. MINOR revision.
Contact author(s)
wzhang @ esat kuleuven be
History
2017-02-28: revised
2017-02-28: received
See all versions
Short URL
https://ia.cr/2017/188
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2017/188,
      author = {Wenying Zhang and Vincent Rijmen},
      title = {Division Cryptanalysis of Block Ciphers with a Binary Diffusion Layer},
      howpublished = {Cryptology {ePrint} Archive, Paper 2017/188},
      year = {2017},
      url = {https://eprint.iacr.org/2017/188}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.