This repository contains the implementation of OblivSketch, an oblivious network measurement service. This service leverages Intel SGX to protect the sensitive network statistics while providing an efficient network measurement service. It also utilises oblivious data structures and algorithms to avoid memory access side-channels against Trusted Execution Environments (TEE). The detailed construction and security proof are presented in our NDSS'21 paper [1].
- Git
- Intel SGX (The SGX simulation mode can be used to run our code, but it does not offer the hardware security feature, and the performance gain is not significant because it does not have the paging issue)
- Ubuntu 18.04
- g++-7 (7.5.0 in ubuntu 18.04)
- cmake 3.17
- openssl 1.1.1 (SGXSSL is needed)
git clone https://github.com/MonashCybersecurityLab/measurement.git
cd measurement
mkdir build
cd build
# use cmake to build the code
cmake ..
cmake --build . --target [strawman|strawman_HW|oblivious|oblivious_HW|cleanup]
The above commands build the strawman and oblivious designs discusssed in the paper. Note that *_HW requires SGX-enabled hardware to execute.
After compiling the project, users can run app {path_to_traffic}
to start the test program. The test program generates
the sketch from the traffic dump and executes all measurement tasks.
- Submit an issue
- Email the authors: [email protected], [email protected]
- Note 1: This is a test program relying on the pre-processed traffic. We note that some features are implemented but
require further scrutinisation, so they are currently in the
test
branch. We will try to clean it up by the end of Feb. - Note 2: The code for OVS integration is deeply coupled with the OVS. I will simplify it and provide the patch file, so everyone can easily update the OVS code for the integration purpose.
[1] Shangqi Lai, Xingliang Yuan, Joseph K. Liu, Xun Yi, Qi Li, Dongxi Liu, and Surya Nepal. 2021. OblivSketch: Oblivious Network Measurement as a Cloud Service. In the Network and Distributed System Security Symposium (NDSS). DOI: https://dx.doi.org/10.14722/ndss.2021.24330