shuizhongmose / OSSE

Shang Z, Oya S, Peter A, et al. Obfuscated access and search patterns in searchable encryption[J]. arXiv preprint arXiv:2102.09651, 2021

An Implementation of OSSE

This is repository contains the prototype implementation of OSSE. The files in this repository refer to the scheme as DP-SSE (Differentially Private SSE) since this was its original name.

This implementation is a research poc built for micro benchmarking purpose, and is not intended to be used in production-level code as it has not been carefully verified for potential performance issues or security flaws.

Authors: Zhiwei Shang, Simon Oya, Andreas Peter, Florian Kerschbaum.

Paper: Obfuscated Access and Search Patterns in Searchable Encryption, 28th Network and Distributed System Security Symposium (NDSS), 2021.

Contact Zhiwei for questions about the code: [e-mail].

Prerequisites

Make sure you have the following installed.

• Python 3.x.x
• GMP 5.x
• PBC
• OpenSSL

Installation

    git clone --recursive https://github.com/z6shang/dp_sse.git
    cd fhipe/charm
    git checkout master && git pull
    cd ..
    # It should be noted that you might need to modify the original Makefile in the following way if you are using conda environments
    # Modify 
    # $(CHARM): FORCE
    #    cd charm && ./configure.sh  --install=. && make
    # Into
    # $(CHARM): FORCE
    #	 cd charm && ./configure.sh  --python=your/path/to/python --install=. && make
    sudo make install

Running the benchmarking

Make sure that you are in the dir dp_sse/code.

	python3 dp_sse_bench.py

Modules

This library ships with the following modules:

• DPSSE: In dp_sse/code/dp_sse.py, implements differential private SSE.
• DPSSE for benchmarking simulation In dp_sse/code/dp_sse_bench.py, implements/simulates the dpsse in benchmarking.
• Datasets: In dp_sse/db/*.json, there are the preprocessed dataset used to build dpsse and dpsse benchmarking.

Submodules

We rely on the following two submodules:

1. FHIPE for ipe support. (Note that it is not required in benchmarking.)
   1. FLINT for the C backend
   2. Charm for the pairings implementation