This GitHub repository provides the implementation of the lattice-based post-quantum commitment scheme introduced by Martínez and Morillo in https://eprint.iacr.org/2019/1486.
Our implementation is very flexible and its parameters can be adjusted to obtain a trade-off between speed and memory usage, analyzing how suitable for practical use are the underlying lattice-based techniques.
Disclaimer: this implementation is only an academic proof of concept and is NOT ready for production.
To build the code and/or the dependencies make sure build-essential, m4 and libgmp-dev are installed in your system.
This code depends on the following libraries (the specific tested versions are mentioned in parenthesis):
Once the dependencies are installed the code can be built running make.
In order to be able to generate new secure sets of parameters sagemath has to be available and https://github.com/malb/lattice-estimator has to be added as submodule:
git submodule add https://github.com/malb/lattice-estimator latticeEstimator
Then to generate a new set of parameters with λ = 100, n = 512, q = 16381 and d = 2 just run the following line.
sage generate_params.sage 100 512 16381 2
And the full set [λ, n, q, d, k, σ, B, δOL, δM] will be computed and added to generatedParams.csv.
The destination file can be customized including it as an additional parameter with -o or --output.
sage generate_params.sage 100 512 16381 2 --output generatedParams2.csv
The general syntax for benchmarking the protocols is the following.
python3 run_tests.py algorithm parametersWhere algorithm can be keygen, commitment, verifier, opening, linear or multiplicative.
Additionally algorithm can be commit to sequentially benchmark all commitment related protocols (keygen, commitment and verifier), NIZKP to benchmark all the Non-Interactive Zero-Knowledge Proofs (opening, linear and multiplicative) or all to run all of them.
Finally parameters would indicate the line number (minus two) of the set of parameters from generatedParameters.csv that is going to be benchmarked or -1 if all the sets of parameters should be benchmarked sequentially.
Benchmark results indicating both running time in milliseconds and processor cycles are saved to specific files in the ./benchmarks folder.
The number of iterations is defined by a variable it in the run_tests.py file.
If one wants to get the parameter sets from a different file then both run_tests.py and ./src/param.h have to be edited modifying variables filename and FILEPARAMETERS respectively.