- Microsoft SEAL 3.2.0
- GMP Library
- CMake
- Go 1.13 or higher
| 0) Install dependencies (Ubuntu): | 0) Install dependencies (CentOS): |
|---|---|
sudo apt-get install build-essential |
sudo yum groupinstall 'Development Tools' |
sudo apt-get install cmake |
sudo yum install cmake |
sudo apt-get install libgmp3-dev |
sudo yum install gmp-devel |
sudo apt-get install golang-go |
sudo yum install golang |
- Globally install Microsoft SEAL 3.2.0:
wget https://github.com/microsoft/SEAL/archive/refs/tags/v3.2.0.tar.gz
tar -xvf v3.2.0.tar.gz
cd SEAL-3.2.0/native/src
cmake .
make
sudo make install
- Compile the SealPIR library in
adveil/C/:
cd C/
cmake .
make -j
- Run the desired experiment! (See next section)
- On the Broker machine:
bash run_targeting.sh
--port 8000 \
--numprocs 1
- On the client machine:
bash run_client.sh --brokerhost localhost --brokerport 8000 --trials 10 --autoclose
or cycle through all experiments at once:
bash clicycle.sh --brokerhost localhost --brokerport 8000 --trials 10 --autoclose
The client triggers the start of the experiment on the server.
The targeting_params.sh script iterates through a parameters and initializes the server with the params.
Each client run starts a new experiment under the specified parameters and saves it to a JSON file.
Therefore, to cycle through all experiments, we re-run the client many times (e.g., 100); this is done automatically using clicycle.sh.
Resulting experiment summary will be saved in the adveil/results directory.
All results as saved as .json files (one per experiment) in the adveil/results`` directory. Use concat.pyto merge multiple files into one.json``` array file.
python concat.py --dir ../results --out ../results/targeting.json
While the code has been tested on MacOS (Big Sur), there is a bug in the cgo interface.
On Linux, uint64_t is cast as C.ulong in the Go code which does not work on Mac (seems to be a bug?)
The compiler will handle switching between two instances of sealpir.go to work around this issue.
Please see AdVeil Accuracy for instructions on reproducing the ANN accuracy experiments.
- Part of the anonymous token code was obtained from the Privacy Pass implementation.
This implementation of AdVeil is intended as a proof-of-concept prototype only! The code was implemented for research purposes and has not been vetted by security experts. As such, no portion of the code should be used in any real-world or production setting!
Copyright © 2021-2023 Sacha Servan-Schreiber
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.