This is the source code for the distinguish attack presented in "Curse of Re-encryption" paper.

In this repository, we targetted
non-protected SHAKE (SHA3) and AES in pqm4 as non-protected software
https://github.com/mupq/pqm4

SASEBO AES hardware as non-protected hardware
http://www.aoki.ecei.tohoku.ac.jp/crypto/

a bit-sliced masked AES software presented in SAC 2016 by Schwabe and Stoffelen as masked software
https://github.com/Ko-/aes-armcortexm

a mased AES hardware based on threshod implementation presented in COSADE 2017 by Ueno, Homma, and Aoki as masked hardware
https://github.com/homma-lab/curse_of_re-encryption (published in this repository)

Quick Start Guide

1. Clone this repository to get the source code for the experiment.

   git clone https://github.com/ECSIS-lab/curse_of_re-encryption.git

2. Install the modules for the use of our source code dl.py

   pip install numpy tensorflow scikit-learn

3. Let training datasets of fixed and random trace be fixed.npy and random.npy, respectively, and put them at ./wave/imple/train with dl.py.
   For example, if you clonde this repository, it is /curse_of_re-encryption/distinguish_attack/wave/imple/train
   As dl.py supports the following implementations as imple (after you acquired traces for each implementation), please put it as

Directry name (imple)	Target implementation
aes_nonprotect_hw	Non-protected AES hardware
keccak_nonprotect_sw	Non-protected keccak software
aes_masked_hw	Masked AES hardware
aes_masked_sw	Masked AES software

4. As well, let test datasets be fixed.npy and random.npy, and put them at ./wave/imple/test.

5. Execute dl.py.

   python dl.py