scalib 0.3.0

Side-Channel Analysis Library

Welcome to SCALib

The Side-Channel Analysis Library (SCALib) is a Python package that contains state-of-the-art tools for side-channel evaluation. It focuses on providing efficient implementations of analysis methods widely used by the side-channel community and maintaining a flexible and simple interface.

SCALib contains various features for side-channel analysis. Please read SCALib-workflow for more details:

• Metrics:

  • SNR: Signal-to-noise ratio.

  • Ttest: T-test estimation.

• Modeling leakage distribution:

  • LDAClassifier: Template in linear subspaces.

• Attacks:

  • SASCAGraph: Generalization of Divide & Conquer with Soft Analytical Attacks.

• Postprocessing to analyse attacks results:

  • rankestimation: Histogram based full key rank estimation.

Install

You can install SCALib by using PyPi packages and running:

pip install scalib

Wheels for Windows and Linux are provided. More information about source compilation, checkout develop informations.

Pseudo-Example

Next, we detail a short pseudo example which illustrates the usage of SCALib. For a full running example, please visit this example.

# compute snr
snr = SNR(nc=256,ns=ns,p=1)
snr.fit(traces_p,x_p)

# build model
pois_x = np.argsort(snr.get_snr()[0][:-npoi])
lda = LDAClassifier(nc=256,ns=npoi,p=1)
lda.fit(traces_p[:,pois_x],x_p)

# Describe and generate the SASCAGraph
graph_desc = ´´´
   # Small unprotected Sbox example
   TABLE sbox   # The Sbox
   VAR SINGLE k # The key
   VAR MULTI p  # The plaintext
   VAR MULTI x  # Sbox input
   VAR MULTI y  # Sbox output
   PROPERTY x = k ^ p   # Key addition
   PROPERTY y = sbox[x] # Sbox lookup
   ´´´
graph = SASCAGraph(graph_desc,256,len(traces_a))

# Encode data into the graph
graph.set_table("sbox",aes_sbox)
graph.set_public("p",plaintexts)
graph.set_distribution("x",lda.predict_proba(traces_a))

# Solve graph
graph.run_bp(it=3)

# Get key distribution and derive key guess
k_distri = graph.get_distribution("k")
key_guess = np.argmax(k_distri[0,:])

SCALib-workflow

The current version of SCALib contains modules for all the necessary steps for a profiled side-channel attack. Even if modules of SCALib can be used independently, a typical usage of SCALib for it goes in four steps:

1. Metrics: In this step, standard metrics are evaluated from the measurements. This helps to find point-of-interest (POIs), quantify avaiable information, etc. When applicable, these metrics are implemented with a one-pass algorithm to save the cost of data load / store.

2. Modeling: In this step, models are built to extract information about a variable y from the leakage. Modeling methods work in two phases. The first one is to fit(t,x) with traces t and target values x. This creates the model parameters. In the second step, this model can be used to predict_proba(t) that returns the probability of every possible target values based on the traces t and the model.

3. Attacks: This modules contains attack methodologies. It essentially uses the probabilities from the modeling step in order and recombine them to recover a key. The module SASCAGraph is an extension of Divide & Conquer attacks that leverage soft analytical side-channel attacks. It allows to define PROPERTY that link intermediate varibles within an implementation. By providing the SASCAGraph with the distributions of these variables, it propagates information on all the variables (e.g. secret keys).

4. PostProcessing: Once the attack has been performed, the postprocessing allows to evaluate the efficiency of the attack. Namely by estimating the rank of the correct key with rank_accuracy, it allows to quantify what is the remaining computational power that is needed by the adversary to recover the correct key.

Full example of SCALib is available here.

About us

SCALib has been initiated by Olivier Bronchain during his PhD at Crypto Group, UCLouvain, Belgium. His colleague Gaëtan Cassiers co-authored SCALib. It has already been used by many other researcher at UCLouvain which contributed either directly or by constructive feedbacks.

Contributions and Issues

We are happy to take any suggestion for features would be useful for side-channel evaluators. If you want to contribute to the project, please visit DEVELOP.rst for relevant information. Please contact Olivier Bronchain at [e-mail] for any futher suggestions / questions.

License

This project is licensed under GNU AFFERO GENERAL PUBLIC LICENSE, Version 3. See COPYING for more information.

Publications

SCALib has been used in various publications, let us know if you used it so that we can add it to the list.

1. “Mode-Level vs. Implementation-Level Physical Security in Symmetric Cryptography: A Practical Guide Through the Leakage-Resistance Jungle”, D. Bellizia, O. Bronchain, G. Cassiers, V. Grosso, Chun Guo, C. Momin, O. Pereira, T. Peters, F.-X. Standaert at CRYPTO2020.

2. “Exploring Crypto-Physical Dark Matter and Learning with Physical Rounding Towards Secure and Efficient Fresh Re-Keying”, S. Duval, P. Méaux, C. Momin, F.-X. Standaert in TCHES2021 - Issue 1.

3. “Breaking Masked Implementations with Many Shares on 32-bit Software Platforms or When the Security Order Does Not Matter”. O. Bronchain, F.-X. Standaert in TCHES2021 - Issue 3.

4. “Improved Leakage-Resistant Authenticated Encryption based on Hardware AES Coprocessors”. O. Bronchain, C. Momin, T. Peters, F.-X. Standaert in TCHES2021 - Issue 3.