DeepNC: A Fast GNN-based Pre-Verification Surrogate for TSN Configuration

2026-07-27

AI Summary

This study addresses the severe computational bottleneck in Time-Sensitive Networking (TSN) configuration space exploration, where repetitive Network Calculus (NC)-based schedulability analyses consume over 90% of the total time, hindering large-scale deployment. To overcome this limitation, the authors propose DeepNC, the first framework to structurally embed network calculus theory into a graph neural network (GNN) to construct a high-fidelity surrogate model for end-to-end delay prediction. DeepNC further incorporates a pre-verification filtering mechanism that drastically reduces the number of formal verification calls while preserving their correctness. Experimental results demonstrate that DeepNC improves the coefficient of determination (R²) by 55.8% and reduces mean absolute percentage error (MAPE) by 65.3% compared to existing learning-based approaches. Moreover, it decreases formal verification invocations by 93.25%, accelerating the overall configuration process by more than two orders of magnitude.

Abstract

Time-Sensitive Networking (TSN) is critical to deterministic communication in safety-critical domains, with formal verification such as Network Calculus (NC) serving as the cornerstone for schedulability guarantees. However, during automated configuration-space exploration, repeated schedulability analysis consumes over 90% of the total configuration time, becoming the primary bottleneck for large-scale TSN configurations. To address this challenge, we propose DeepNC, a novel pre-verification surrogate module that pioneers the structural fusion of NC principles into a Graph Neural Network (GNN) for TSN configuration-space exploration. Rather than replacing formal verification, DeepNC acts as a high-speed pre-verification filter, reserving computationally expensive formal verification only for promising candidates. Extensive evaluations demonstrate that DeepNC significantly improves worst-case delay prediction accuracy over state-of-the-art learning-based methods, increasing the average $R^2$ by 55.8% and reducing the average MAPE by 65.3%. More importantly, its high-fidelity regression substantially reduces the number of formal verification calls during configuration-space exploration by 93.25%, while accelerating NC-based verification by more than two orders of magnitude.

Problem

• Time-Sensitive Networking
• Network Calculus
• schedulability analysis
• configuration-space exploration
• verification bottleneck

Innovation

• Graph Neural Network
• Network Calculus
• Time-Sensitive Networking
• Pre-verification Surrogate
• Schedulability Analysis