Maximizing average happiness privately & securely.
Suppose there's a group of n
ladies and n
lads, and they want to be matched in pairs.
Each person has some deeply hidden desires, represented by an integer vector.
For each pair, cost of matching them together is a function of their desires (the lower it is, the better).
We want to find a matching that minimizes the average cost, without revealing their secrets - each person should learn only who is their better half.
Oblivious minimum cost maximum flow algorithm running under SPDZ protocol [1,2], based on ideas from [3].
mpc
- mini-framework for MPC computation (SPDZ online phase, fundamental circuits etc)mpc_flow
- implementation of oblivious minimum cost flow and matching algorithms for use in MPCdealer
- tool that precomputes stuff for SPDZ protocolmatcher
- the secret matching application
- Rust 1.58 - to compile the projects
- Python 3.8 - for convenience scripts
- OpenSSL - for generating self-signed certificates
- Build everything:
cargo build --release
- Create test environment (default is 16 nodes):
./prepare-test-env.py
- Precompute parameters for SPDZ:
./precompute-spdz.py
- Run all test nodes locally:
./run-all-parties.py
You can run test nodes individually using ./run-party.py
; run scripts with --help
for more information.
[1] Multiparty Computation from Somewhat Homomorphic Encryption
[2] Practical Covertly Secure MPC for Dishonest Majority – or: Breaking the SPDZ Limits
[3] Data-oblivious graph algorithms for secure computation and outsourcing
[4] Improved Primitives for Secure Multiparty Integer Computation