integral

Integral Distinguisher

Find integral distinguisher via monomial prediction

Since the MDS matrix model is quite heavy, the model takes quite a while to solve and it is therefore hard to find the best integral distinguishers.

Requirements

Nix

run nix develop

Manual

• Requires python, uv, minizinc, pyminizinc
• Run uv sync and source .venv/bin/activate

HW method

Uses HW method proposed in "Finding Bit-Based Division Property for Ciphers". For each monomial trail at the MDS Matrix we check if the masked sub matrix is invertable (=correct monomial trail).

The tool checks each bit of the output for a balanced property. This is done for all inputs with one constant bit. Therefore we need to check if all 32*32 combinations are key dependent in order to rule out integral distinguisher for a given round number.

usage: mon_int_cp.py [-h] [-r ROUNDS] [-b BREAK]

Find integral distinguisher with monomial prediction.

options:
  -h, --help            show this help message and exit
  -r ROUNDS, --rounds ROUNDS
  -b BREAK, --break BREAK