NetDiffus: Network Traffic Generation by Diffusion Models through Time-Series Imaging

TL;DR

NetDiffus leverages diffusion models to generate high-fidelity synthetic network traffic by transforming time-series data into images, outperforming GAN-based methods and enhancing various network analysis tasks.

Contribution

This work introduces NetDiffus, a novel framework that converts network traffic time-series into images for diffusion-based synthetic data generation, surpassing GANs in fidelity and utility.

Findings

66.4% increase in data fidelity over state-of-the-art methods

18.1% improvement in downstream machine learning tasks

Effective across diverse traffic traces for multiple applications

Abstract

Network data analytics are now at the core of almost every networking solution. Nonetheless, limited access to networking data has been an enduring challenge due to many reasons including complexity of modern networks, commercial sensitivity, privacy and regulatory constraints. In this work, we explore how to leverage recent advancements in Diffusion Models (DM) to generate synthetic network traffic data. We develop an end-to-end framework - NetDiffus that first converts one-dimensional time-series network traffic into two-dimensional images, and then synthesizes representative images for the original data. We demonstrate that NetDiffus outperforms the state-of-the-art traffic generation methods based on Generative Adversarial Networks (GANs) by providing 66.4% increase in fidelity of the generated data and 18.1% increase in downstream machine learning tasks. We evaluate NetDiffus on seven diverse traffic traces and show that utilizing synthetic data significantly improves traffic fingerprinting, anomaly detection and traffic classification.