Traffic-MoE: A Sparse Foundation Model for Network Traffic Analysis

TL;DR

Traffic-MoE introduces a sparse, expert-routing model for network traffic analysis that significantly improves detection accuracy, throughput, and efficiency, enabling real-time deployment in security environments.

Contribution

It presents a novel sparse foundation model with dynamic routing for efficient, high-performance network traffic analysis, addressing computational challenges of large pre-trained models.

Findings

Up to 12.38% better detection performance

91.62% increase in throughput

47.81% reduction in inference latency

Abstract

While pre-trained large models have achieved state-of-the-art performance in network traffic analysis, their prohibitive computational costs hinder deployment in real-time, throughput-sensitive network defense environments. This work bridges the gap between advanced representation learning and practical network protection by introducing Traffic-MoE, a sparse foundation model optimized for high-efficiency real-time inference. By dynamically routing traffic tokens to a small subset of specialized experts, Traffic-MoE effectively decouples model capacity from computational overhead. Extensive evaluations across three security-oriented tasks demonstrate that Traffic-MoE achieves up to a 12.38% improvement in detection performance compared to leading dense competitors. Crucially, it delivers a 91.62% increase in throughput, reduces inference latency by 47.81%, and cuts peak GPU memory consumption by 38.72%. Beyond efficiency, Traffic-MoE exhibits superior robustness against adversarial traffic shaping and maintains high detection efficacy in few-shot scenarios, establishing a new paradigm for scalable and resilient network traffic analysis.