iPET: Privacy Enhancing Traffic Perturbations for Secure IoT Communications

iPET is a privacy enhancing traffic perturbation technique that counters ML-based fingerprinting attacks. iPET uses adversarial deep learning, specifically, Generative Adversarial Networks (GANs), to generate these perturbations. Unlike conventional GANs, a key idea of iPET is to deliberately introduce stochasticity in the model.

Our paper describes the workings of iPET in detail. Link to paper.

We make our source code available in this repository.

The code is organized as follows:

1. The IoT_Device_Fingerprinting directory contains the code for the attacker fingerprinting models (described in Section 2).
2. The ipet directory contains the core algorithm of our proposal (described in Section 4).
3. baselines contains code for traffic obfuscation techniques used in our evaluations (Section 5.1.2).

Each component contains relevant README files.

To generate iPET perturbations, the user must first train the discriminator as described here.

Dependencies

We have tested our code on Ubuntu 20.04 and Python 3.8 (Anaconda Distribution). Additional dependencies can be installed from environment.yml.

$ conda env create --file environment.yml
$ conda activate ipet

Contact

For any queries, please feel free to raise issues or contact the authors.

References

Akshaye Shenoi, Prasanna Karthik, Kanav Sabharwal, Jialin Li, and Dinil Mon Divakaran. “iPET: Privacy Enhancing Traffic Perturbations for Secure IoT Communications,” in PETS (Privacy Enhancing Technologies Symposium), 2023

@article{ipet2023,
author = {Akshaye Shenoi and Prasanna Karthik and Kanav Sabharwal and Jialin Li and Dinil Mon Divakaran},
title = {{iPET: Privacy Enhancing Traffic Perturbations for Secure IoT Communications}},
year = {2023},
journal = {Privacy Enhancing Technologies Symposium}
}