A 32-bit Cipher for Cryptographic Mitigations against Software Attacks

This repository contains reference implementations and tools for cryptanalysis.

Implementation of beanie in c and rust as well as a hardware implementation.

A bit-based model to find the minimum amount of differentially/linearly active S-Boxes with different configurations of the cipher.

Also includes a model for clustering that counts the number of characteristics above a certain probability threshold.

Contains program to experimentally calculate the EDP/ELP for random keys/tweaks

Experiments based on https://eprint.iacr.org/2023/288.pdf

Propagates forward and backward with probability one and finds overlapping cells

Done by hand

Sat model to find integral distinguisher via monomial prediction.

@article{beanie,
  title={BEANIE--A 32-bit Cipher for Cryptographic Mitigations against Software Attacks},
  author={Gerhalter, Simon and Hod{\v{z}}i{\'c}, Samir and Medwed, Marcel and Nageler, Marcel and Folwarczny, Artur and Nikov, Ventzi and Hoogerbrugge, Jan and Schneider, Tobias and McConville, Gary and Eichlseder, Maria},
  journal={IACR Transactions on Symmetric Cryptology},
  volume={2025},
  number={4},
  year={2025}
}

• To generate constraints we used the tool https://github.com/hadipourh/sboxanalyzer

• Due to historical reasons in a lot of the analysis the state is indexed row wise and not column wise like in the spec.

• For most tools we use minizinc with the solver or-tools and the python library pyminizinc.