Enhancing block cipher diffusion: A mathematical approach to key-dependent mix column transformations
Year: 2025
Type: article
Abstract: In this paper, we propose a novel key-dependent MixColumn transformation to enhance the diffusion properties of block ciphers, specifically within the AES framework. The proposed method introduces dynamic variation into the MixColumn step by extracting a pseudo-random value from the round key, which is used to perform cyclic permutations on the MixColumn coefficients. This leads to the construction of a round-dependent transformation matrix, mathematically defined and proven to maintain invertibility over GF(2⁸). We provide a formal mathematical representation of the proposed model and analyze its diffusion characteristics, linear independence, and resistance to cryptanalytic attacks. Our approach eliminates static transformation patterns and introduces key-driven unpredictability, offering a promising direction for adaptive cryptographic primitives without increasing algorithmic complexity or requiring hardware modifications.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4410279858