2 releases

0.1.1 Sep 6, 2023
0.1.0 Sep 6, 2023

#1694 in Cryptography

Download history 8/week @ 2024-02-22 5/week @ 2024-02-29 9/week @ 2024-03-07 38/week @ 2024-03-14 4/week @ 2024-03-28

51 downloads per month

GPL-3.0-only

35KB
403 lines

smol

Rust Dependency status Documentation

Smol-mpc is a tiny library to learn Secure Multiparty Computation (MPC) basics using the Rust programming language. Smol-mpc allows to the user to experiment with toy examples of MPC protocols and allows him/her to understand the basic concepts behind MPC. We developed Smol-mpc based on the ideas present in TinySMPC with some elements taken from Secure Computation Library. In the case of TinySMPC, we took the basic architecture and adapted it to be more compatible with Rust. Some elements like virtual machines and shares are very similar, but Smol-mpc differs significantly at some points to give the user a better experience using Rust. The main difference with Tiny SMPC is the way in which we consider the virtual machine. In our case is more similar to the way in which ideal functionalities are specified in the theory of MPC, that is, the memory of each virtual machine is addressed using IDs for each variable stored. For Secure Computation Library, we took the ideas used there in the math library to implement the $\mathbb{F}_p$ with $p = 2^{61} - 1$ (which is a Mersenne prime) as our underlying algebraic structure. We also took the idea from Secure Computation Library to implement a pseudo-random generator using AES-CTR.

The documentation for this library and some examples of how to use it can be found here.

Milestones

  • Fully document the source code.
  • Add a user-friendly output functionality for the virtual machine states.
  • Write a protocol for Beaver triple generation.
  • Allow support for other protocols.

Disclaimer

We stress that the work presented here is purely educational and does not intend to show a secure or efficient implementation. The core of the library is to give the user an idea of how protocols work at a very high level. So, the implementation may have security issues and sometimes it may not represent all the details and caveats of a real-world secure and efficient implementation of the techniques covered here.

Dependencies

~1MB
~19K SLoC