Taming Volatility: Stable and Private QUIC Classification with Federated Learning

2025-09-12

AI Summary

Federated learning (FL) for privacy-preserving network traffic analysis faces dual challenges from realistic non-IID data: statistical heterogeneity and temporal dynamics—particularly the diurnal variations in QUIC traffic—which cause uneven client data availability and training instability. This work is the first to systematically characterize the adverse impact of such temporal fluctuations on FL convergence. We propose a client-side local buffering mechanism that decouples real-time traffic variability from model update scheduling, thereby ensuring training stability under non-IID dynamic conditions. Evaluated on the CESNET-QUIC22 dataset across 14 autonomous clients, our method achieves a 95.2% F1-score under strict privacy constraints—only 2.3 percentage points below the centralized baseline—while significantly improving convergence robustness and practical deployability.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Natural Language Processing: Safety and Robustness
• Application Domains: Mobility, Driving & Flight

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Federated Learning (FL) is a promising approach for privacy-preserving network traffic analysis, but its practical deployment is challenged by the non-IID nature of real-world data. While prior work has addressed statistical heterogeneity, the impact of temporal traffic volatility-the natural daily ebb and flow of network activity-on model stability remains largely unexplored. This volatility can lead to inconsistent data availability at clients, destabilizing the entire training process. In this paper, we systematically address the problem of temporal volatility in federated QUIC classification. We first demonstrate the instability of standard FL in this dynamic setting. We then propose and evaluate a client-side data buffer as a practical mechanism to ensure stable and consistent local training, decoupling it from real-time traffic fluctuations. Using the real-world CESNET-QUIC22 dataset partitioned into 14 autonomous clients, we then demonstrate that this approach enables robust convergence. Our results show that a stable federated system achieves a 95.2% F1 score, a mere 2.3 percentage points below a non-private centralized model. This work establishes a blueprint for building operationally stable FL systems for network management, proving that the challenges of dynamic network environments can be overcome with targeted architectural choices.

Problem

• Addressing temporal volatility in federated QUIC classification
• Stabilizing FL training under non-IID network traffic data
• Ensuring consistent local training despite traffic fluctuations

Innovation

• Client-side data buffer mechanism
• Decoupling training from traffic fluctuations
• Stable federated QUIC classification system