An Implementation of Multi-Input Functional Encryption

Prerequisites

We rely on

• FLINT for pretty much everything
• DGS for sampling from discrete Gausisans over the Integers

We ship these libraries as submodules.

Installation

git clone --recursive https://github.com/GaloisInc/gghlite-flint.git
cd gghlite-flint
mkdir dgs/m4
autoreconf -i
export CFLAGS="-O3"
./configure
make -j

Usage

GGH Params

Prints parameter choices to stdout.

GGH Instance

Instantiates a GGHLite instance

NIKE

A non-interactive key exchange (but we skip running the hash function in the end). For example, calling

./applications/nike -l 80 -N 7

will run 7-partite NIKE with security level λ=80.

Subdirectories

• applications NIKE, parameter estimation, benchmarks
• dgs discrete Gaussians over the integers
• dgsl discrete Gaussians over arbitrary regular lattices and ideal lattices modulo $x^{2^k}+1$
• flint a current snapshot of the FLINT library
• gghlite the actual GGHLite implementation
• oz efficient functions for computing in $\mathbb{Z}[x]/(x^{2^k}+1)$ and $\mathbb{Z}_q[x]/(x^{2^k}+1)$
• tests correctness tests