One of our recent works, “Path-Based Graph Neural Network for Robust and Resilient Routing in Distributed Traffic Engineering”, was accepted by the IEEE Journal on Selected Areas in Communications (JSAC) on 10/29/2024.
In this work, we leverage the emerging Graph Neural Network (GNN) architecture and Supervised Learning (SL) technique to improve the performance of distributed TE in unexpected network scenarios with low training/inference overhead. PathGNN employs a novel path-link bipartite graph modeling approach to capture rich network information as generalizable knowledge, and performs efficient message exchanges among routers to quickly infer good local routing strategies. Extensive evaluation results reveal that PathGNN can generate robust and resilient routing to achieve promising network performance under dynamic traffic scenarios and multiple link failures.
Project Overview:
(1) PathGNN’s path-link bipartite graph modeling to extract generalizable knowledge from networks
(2) Offline training procedure of PathGNN with Supervised Learning
(3) Online distributed deployment of PathGNN models with message exchanges among routers
(4) Evaluation results in the Sprintlink network with different traffic variations and random link failures
Abstract:
Distributed Traffic Engineering (TE) aims to optimize network performance by generating individual routing strategies at each router without a global view of the network. A major challenge for these TE solutions is handling performance degradation caused by unexpected traffic fluctuations and unpredictable link failures. Recently, Machine Learning (ML) techniques have introduced new opportunities to enhance distributed TE.
In this paper, we propose Path-Based Graph Neural Network (PathGNN), which leverages the emerging GNN architecture to quickly infer robust and resilient routing strategies in a distributed manner to accommodate unexpected network conditions. PathGNN adopts a novel path-link bipartite graph modeling approach to capture the dynamics of link resources shared by routing paths. It then performs efficient GNN message exchanges among routers to make adaptive local routing decisions for better load balancing. Additionally, PathGNN leverages Supervised Learning (SL) to directly learn from optimal routing strategies through efficient offline training.
Evaluation results on four real-world network topologies demonstrate PathGNN's strong generalization capability. Compared to state-of-the-art distributed TE solutions, PathGNN improves the load balancing performance by at least 24.4% with lower end-to-end delay under dynamic traffic scenarios, and also boosts performance by up to 35.3% under multiple link failures.
Publications:
[JSAC 25] Minghao Ye, Junjie Zhang, Zehua Guo, and H. Jonathan Chao, “Path-Based Graph Neural Network for Robust and Resilient Routing in Distributed Traffic Engineering,” IEEE Journal on Selected Areas in Communications (JSAC), 2025. (Impact factor: 13.8) [Paper URL] [PDF]