Skip to content

zkFHE/zkOpenFHE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

zkOpenFHE - an OpenFHE fork with (zero-knowledge) proofs of computation

Fully Homomorphic Encryption (FHE) is a powerful cryptographic primitive that enables performing computations over encrypted data without having access to the secret key. However, FHE only achieves a relatively weak security notion (IND-CPA or IND-CPA-D), which is often insufficient for real-world deployments.

The goal of zkOpenFHE is to be a drop-in replacement for the excellent OpenFHE library, with the additional ability to prove the correct evaluation of an FHE circuit using a (zero-knowledge) succinct non-interactive argument of knowldege ((zk)SNARK).

To achieve this, we mirror's OpenFHE's interfaces, and augment them under the hood with automatic constraint generation and extended witness computation, using libsnark.

Usage

OpenFHE zkOpenFHE
c = cryptoContext->Encrypt(secretKey, x);


c_rot = cryptoContext->EvalRotate(c, 1);

c2 = cryptoContext->EvalMultNoRelin(c, c_rot);
proofSystem = LibsnarkProofSystem(cryptoContext);

c = cryptoContext->Encrypt(secretKey, x);
proofSystem.PublicInput(c);

c_rot = proofSystem->EvalRotate(c, 1);

c2 = proofSystem->EvalMultNoRelin(c, c_rot);

A ProofSystem has three modes:

  • PROOFSYSTEM_MODE_EVALUATION: evaluate the FHE circuit, just like OpenFHE
  • PROOFSYSTEM_MODE_CONSTRAINT_GENERATION: generate the constraints corresponding to the FHE circuit and automatically optimize them
  • PROOFSYSTEM_MODE_WITNESS_GENERATION: generate the extended witness for the FHE circuit, which is needed for proving correct computation

Links and Resources

For a more complete overview of our project, have a look at zkFHE.github.io, where you'll find out more about what we're trying to achieve

Installation

To install zkOpenFHE, refer to OpenFHE's General Installation Information: readthedocs for more information

Or refer to the following for your specific operating system:

  • Linux

  • MacOS. Note that the libsnark backend will not build on Apple Silicon, but a fix should be available soon.

  • Windows

Benchmarks and Examples

Contributing

If you'd like to contribute, please reach out! We're also very grateful if you report issues, big or small, or if you'd like to contribute some examples.