thankuyou / funshade

Function Secret Sharing for Secure Thresholded Distance Metric Evaluation.

Funshade

Functional Secret Sharing for Two-Party Secure Thresholded Distance Evaluation.

• Languages: C (99+), Python (3.8+, for the wrapper)
• Platforms: Linux, Windows, MacOS (any system with a C compiler)
• Dependencies: libsodium (optional, for fast and secure random number generation)
• Code Author: Alberto Ibarrondo
• License: GNU GPLv3 (any code derived from this must be open source)
• Version: 1.1.0

Description

Funshade is a library that implements a protocol to securely compute a distance metric between two vectors, followed by a secure comparison to a threshold. It is extremely lightweight, making use of cheap primitives such as AES and requiring a single intermediate round of communication and just two numbers sent per distance computation in the preprocessing model (besides input sharing and output reconstruction, common to all secret-sharing MPC schemes).

Consider citing the paper if you use it in your research!

Installation

The library is written in plain C to make it portable. It has been tested in Linux and Windows, but it should work in any system with a C compiler. To compile it, you can:

• Use the provided CMakeLists.txt with cmake(mkdir build && cd build && cmake .. & cmake --build .)
• Directly call your compiler with the -msse -msse2 -maes flags for faster hardware-based AES acceleration (and consider -O3 -march=native for further optimizations).

For conveniency and for seamless integration with higher-level languages, we also provide a Python wrapper. Install it with:

• pip install .

As an optional dependency, it uses libsodium for fast and secure random number generation.

Usage

The library is designed to be used as a black-box, with a simple API.

Chech the bottom of fss.h for the available functions, test_fss.c for some uses in C, or test_funshade.py for a step-by-step Python example.