On the circulant matrix MDS testing and the search for circulant MDS matrices

被引:0
作者
Malakhov, Stanislav S. [1 ]
机构
[1] HSE Univ, Moscow, Russia
来源
CRYPTOGRAPHY AND COMMUNICATIONS-DISCRETE-STRUCTURES BOOLEAN FUNCTIONS AND SEQUENCES | 2024年
关键词
Circulant matrix; Double circulant code; MDS code; MDS matrix; CODES;
D O I
10.1007/s12095-024-00746-7
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
MDS matrices are used in symmetric cryptography to hinder differential and linear cryptanalysis. This article proposes and examines a new deterministic method that accelerates circulant matrix MDS testing and the search for circulant MDS matrices. The method is to ascertain the MDS property via computing the determinants of only those submatrices that lie in a suitable subset of square submatrices constructed in advance. It is shown that for 8x8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ 8 \times 8 $$\end{document} circulant matrices, this new method reduces thirteenfold the MDS confirmation time and searches for MDS matrices 8 times faster compared to the general method employing all square submatrices. The article also proves that the constructed set can be arranged in a manner that comprises all the submatrices needed for the Laplace expansion of the determinant of any submatrix within the subset. Experiments show that the Laplace expansion allows a further two to seven times speed-up of the MDS testing. Via proposed techniques, several circulant MDS matrices were found including 8x8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varvec{8} \times \varvec{8}$$\end{document} matrices over GF(28)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\textbf {GF}}\varvec{(2<^>8)} $$\end{document} and 16x16\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \varvec{16} \times \varvec{16} $$\end{document} matrices over GF(222),GF(223)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\textbf {GF}}\varvec{(2<^>{22}),} {\textbf {GF}}\varvec{(2<^>{23})} $$\end{document} and GF(224)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\textbf {GF}}\varvec{(2<^>{24})} $$\end{document} with many multiplicative identity element entries, a few different elements of the low Hamming weight and efficient inverses. Besides that, empirical probability mass functions were found for the random variables representing the least dimension of singular submatrices of 16x16\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \varvec{16} \times \varvec{16} $$\end{document} circulant matrices of two chosen forms over GF(2m),m is an element of{8,& ctdot;,24}\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ {\textbf {GF}}\varvec{(2<^>m), m \in \{8, \dots , 24\}} $$\end{document}.
引用
收藏
页码:87 / 119
页数:33
相关论文
共 26 条
  • [1] [Анашкин А.В. Anashkin A.V.], 2017, [Математические вопросы криптографии, Matematicheskie voprosy kriptografii], V8, P5, DOI 10.4213/mvk238
  • [2] Augot D, 2013, IEEE INT SYMP INFO, P1551, DOI 10.1109/ISIT.2013.6620487
  • [3] Barreto PSLM, 2000, 1 OP NESSIE WORKSH L, V13, P14
  • [4] Some classes of the MDS matrices over a finite field
    Belov A.V.
    Los A.B.
    Rozhkov M.I.
    [J]. Lobachevskii Journal of Mathematics, 2017, 38 (5) : 880 - 883
  • [5] Belov AV., 2017, Commun. Appl. Math. Comp, V31, P143
  • [6] Fast algorithms to generate necklaces, unlabeled necklaces, and irreducible polynomials over GF(2)1
    Cattell, K
    Ruskey, F
    Sawada, J
    Serra, M
    Miers, CR
    [J]. JOURNAL OF ALGORITHMS-COGNITION INFORMATICS AND LOGIC, 2000, 37 (02): : 267 - 282
  • [7] A note on semi-orthogonal (G-matrix) and semi-involutory MDS matrices
    Chatterjee, Tapas
    Laha, Ayantika
    [J]. FINITE FIELDS AND THEIR APPLICATIONS, 2023, 92
  • [8] Couselo E., 2000, Discrete Mathematics and Applications, V10, P433, DOI 10.1515/dma.2000.10.5.433
  • [9] Couselo E., 1998, Discrete Math. Appl., V8, P217
  • [10] Daemen J, 1997, LECT NOTES COMPUT SC, V1267, P149