RIGEL: Real-time Optical Anomaly Diagnosis with Stateful In-Network Inference based on Distributed On-switch GNNs

2026-07-31

AI Summary

This work addresses the limitations of existing optical network anomaly diagnosis methods, which rely on centralized processing and incur high communication overhead between the control and data planes, thereby failing to meet real-time requirements. To overcome these challenges, we propose RIGEL—the first distributed graph neural network (GNN) system leveraging Tofino programmable switches—enabling stateful, real-time anomaly detection directly in the data plane through hardware-software co-design. We introduce a novel deployment of state-preserving, distributed GNN inference on programmable switches, integrating GraphSAGE with an autoencoder architecture, and develop a general model adaptation technique tailored to hardware constraints. Experimental evaluation on a real-world packet-optical testbed demonstrates that RIGEL achieves high-accuracy, low-latency detection and localization of optical-layer anomalies, significantly outperforming state-of-the-art approaches.

Abstract

The recent booming of data-intensive applications has complicated optical network management, making real-time optical anomaly diagnosis a must-have feature. However, existing approaches are mostly based on centralized data analytics and thus can hardly avoid the latency and overhead due to message exchanges between data and control planes. In this work, we propose and prototype RIGEL, which, to the best of our knowledge, is the first real-time optical anomaly diagnosis system that realizes stateful distributed in-network inference through collaborative graph neural networks (GNNs) on Tofino switches. The system is designed to be fully in-network, and a software-hardware co-design is proposed to preprocess high-dimensional spectral data for being suitable for hardware-based in-network inference. Next, we first develop an effective model to combine an autoencoder with a GraphSAGE-based GNN, and then propose a generalizable method to adapt the model to Tofino switch. The effectiveness of RIGEL is showcased in a realistic packet-over-optical network testbed, verifying that it achieves highly accurate diagnosis to detect and locate optical anomalies timely and highlighting its benefits over the state-of-the-art methods.

Problem

• optical anomaly diagnosis
• real-time
• in-network inference
• distributed GNNs
• network management

Innovation

• in-network inference
• graph neural networks
• optical anomaly diagnosis
• stateful distributed computing
• hardware-software co-design