MohamadMansouri / fault-tolerant-secure-agg

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Secure and Fault-Tolerant Secure Aggregation

Description of the project

FTSA is a protocol to securely aggregate (sum) vectors from multiple clients. The aggregation is performed at a server which we call the aggregator. The protocol has two main interesting properties:

  • Privacy: It preserves the privacy of the clients since it does not leak any information about the users' vectors except their sum.

  • Fault-Tolerance: The aggregator can compute the aggregate of the clients' vectors even if some clients drop from the protocol at any point in time.

A detailed description and evaluation of the protocol are submitted as a technical paper to ACSAC conference. The submitted manuscript can be found @ docs/acsac-submission.pdf

What does this repository contain?

In this repository, you can find:

  • The source code of our protocol (can be found @ ftsa/protocols):

    • The building blocks used to construct the protocol
    • A prototype implementation of our protocol
    • A prototype implementation of SecAgg, the state-of-the-art for fault-tolerant secure aggregation protocols.
  • Documentation of the source code with a detailed description of all the classes and methods. It can be found @ docs/html.

  • Six Jupyter notebooks that allow the user to test our protocol (and SecAgg protocol). They can be found @ ftsa/jupyter_nb.

  • Benchmarks of both our protocol and SecAgg with different parameters. Also, a visualization of the benchmark results. They can be found @ ftsa/benchmarks.

Setup of the project

The project requires a python version higher than 3.8 and a Jupyter notebook client (you can install one by running pip install notebook).

  • First create a python virtual environment and activate it as follows:
virtualenv .venv
source .venv/bin/active
  • Then launch the setup as follows:
python setup.py build
python setup.py install

Running the notebooks

Please refer to here

Running the benchmarks

Please refer to here

About


Languages

Language:Python 72.2%Language:Jupyter Notebook 27.8%