Artifacts for "DRAWN APART : A Device Identification Technique based on Remote GPU Fingerprinting", by Tomer Laor, Naif Mehanna, Vitaly Dyadyuk, Antonin Durey, Pierre Laperdrix, Clémentine Maurice, Yossi Oren, Romain Rouvoy, Walter Rudametkin, and Yuval Yarom
You will find below three distinct demo pages, each testing a different timing measurement method.
Note that both the GPU and Offscreen methods are only available by default at the moment on Chromium-based browsers (see progress on the support of OffscreenCanvas for Firefox and Safari).
Our classification over real data can be viewed using the interactive Python notebook at the following link: https://colab.research.google.com/github/DrawnApart/DrawnApart/blob/master/drawnapart_demo.ipynb
The raw data is contained in the different tgz files present at the root of this repository.
The revisited FPStalker code with DrawnApart is also available at the root of this repository, under the name "fpstalker_drawnapart.ipynb". For privacy-related reasons, we use a bogus dataset for the execution.
It can be executed by visiting the interactive notebook at the following link:
https://colab.research.google.com/github/DrawnApart/DrawnApart/blob/master/fpstalker_drawnapart.ipynb
To assess the robustness of our Drawn Apart method, we filmed two experiments where we swapped hardware between two computers.
Please see the paper's official webpage