Network Tomography with Path-Centric Graph Neural Network
Yuntong Hu, Junxiang Wang, Liang Zhao
TL;DR
This paper tackles network tomography under limited observability by introducing DeepNT, a path-centric graph neural network that jointly predicts end-to-end path performance metrics and infers underlying network topology. The method samples candidate paths, builds path-centric embeddings, enforces metric-specific triangle inequalities, and completes topology through convex, differentiable constraints with sparsity, enabling gradient-based optimization and convergence guarantees. The approach demonstrates superior accuracy across additive, multiplicative, min/max, and boolean metrics on real-world and synthetic networks, and robustly reconstructs topology when partial observations are available. Overall, DeepNT provides a scalable and principled framework for network tomography that integrates data with partial prior knowledge to achieve accurate inferences without requiring full topology or handcrafted metrics.
Abstract
Network tomography is a crucial problem in network monitoring, where the observable path performance metric values are used to infer the unobserved ones, making it essential for tasks such as route selection, fault diagnosis, and traffic control. However, most existing methods either assume complete knowledge of network topology and metric formulas-an unrealistic expectation in many real-world scenarios with limited observability-or rely entirely on black-box end-to-end models. To tackle this, in this paper, we argue that a good network tomography requires synergizing the knowledge from both data and appropriate inductive bias from (partial) prior knowledge. To see this, we propose Deep Network Tomography (DeepNT), a novel framework that leverages a path-centric graph neural network to predict path performance metrics without relying on predefined hand-crafted metrics, assumptions, or the real network topology. The path-centric graph neural network learns the path embedding by inferring and aggregating the embeddings of the sequence of nodes that compose this path. Training path-centric graph neural networks requires learning the neural netowrk parameters and network topology under discrete constraints induced by the observed path performance metrics, which motivates us to design a learning objective that imposes connectivity and sparsity constraints on topology and path performance triangle inequality on path performance. Extensive experiments on real-world and synthetic datasets demonstrate the superiority of DeepNT in predicting performance metrics and inferring graph topology compared to state-of-the-art methods.