ML-Driven Adaptive Routing and Performance in Software-Defined Networks (SDN)

Software-Defined Networks (SDN) provide centralized control for programmable routing, yet traditional algorithms like OSPF and ECMP struggle with dynamic traffic patterns, congestion hotspots, and QoS demands in large-scale deployments. This paper conducts a systematic review of machine learning (ML) techniques— including supervised classifiers, reinforcement learning (RL) agents, and graph neural networks (GNNs)— applied to SDN routing and performance optimization, highlighting their roles in traffic classification (up to 99.81% accuracy), predictive KPI forecasting, and adaptive path selection.

We propose the Hybrid Causal-RL-GNN (HCRG) framework, which fuses Graph Attention Networks (GAT) for topology-aware state encoding with a causality-enhanced Soft Actor-Critic (SAC) agent to quantify action impacts and maximize a composite reward function balancing latency, packet loss, and throughput. Trained offline on Mininet-emulated NSFNET and Fat-Tree topologies with Ryu controllers, HCRG deploys via OpenFlow for real-time flow rule installation, incorporating hyperparameters like learning rate 0.001 and discount factor 0.99 over 20,000 episodes.

Extensive evaluations under normal, congested, and failure scenarios demonstrate HCRG's superiority: 28% latency reduction (22 ms vs. 45 ms baselines), 22% throughput increase (2.2 Gbps), and 35% loss mitigation (1.6%), outperforming ROAR, RouteNet, and ECMP by 15-35% while maintaining <5 ms inference latency at scale. This work advances autonomous SDN traffic engineering, with implications for 5G/6G and edge computing, paving the way for federated extensions in multi-domain environments.

Software Defined Networks (SDN), Machine Learning (ML), Reinforcement Learning (RL), Graph Neural Networks (GNNs), Hybrid Causal-RL-GNN (HCRG

1. Gill, S. S., et al. (2025). Optimized SDN-based routing protocol to improve performance in wireless multimedia sensor networks. AIP Conference Proceedings, 3211(1), 030041. [Google Scholar] [Crossref]

2. Javed, A. R., et al. (2024). Hybrid machine learning based network performance improvement in SDN and VANET [Doctoral dissertation, Loughborough University]. Loughborough University Repository. [Google Scholar] [Crossref]

3. Khan, M. A., et al. (2024). A comprehensive survey on machine learning and SDN integration. [Google Scholar] [Crossref]

4. Alsamhi, S. H., et al. (2025). An intelligent FedAvg-BWO optimizer to enhance federated learning in 6GVANET. PMC, PMC12657367. [Google Scholar] [Crossref]

5. Ferrag, M. A., et al. (2021). Deep learning for cybersecurity in SDN. IEEE Access, 9, Article 9493245. [Google Scholar] [Crossref]

6. Zhang, Y., et al. (2024). Graph neural networks for programmable data center networks. Scientific Reports, 14, Article 70983. [Google Scholar] [Crossref]

7. Mijalkovic, R., et al. (2023). Reinforcement learning for traffic engineering in SDN. International Journal of Computer Applications, 45(1), 897. Montazerolghaem, A., et al. (2021). Machine learning in SDN/NFV. Journal of Network and Computer Applications, 194, 103136. Rashad, S., et al. (2024). AI-driven anomaly detection in SDN. Modern Journal of Computers, Cybernetics & Research Reviews. Article 1270. [Google Scholar] [Crossref]

8. Mahdavinejad, M. S., et al. (2021). Multi-agent RL for SDN load balancing. International Journal of Advanced Computer Science and Applications, 12(5), 41. [Google Scholar] [Crossref]

9. Bannour, F., et al. (2023). Algorithms for network analytics in SDN. BIMSA Lab Technical Report, 3964. [Google Scholar] [Crossref]

10. Y. He et al., "Reinforcement learning-based SDN routing scheme with causal inference and GNN," Frontiers Comput. Neurosci., vol. 18, Art. no. 1393025, Apr. 2024. doi: 10.3389/fncom.2024.1393025. [Google Scholar] [Crossref]

11. [Online]. Available: https://www.frontiersin.org/journals/computationalneuroscience/articles/10.3389/fncom.2024.1393025/full [Google Scholar] [Crossref]

12. Y. Li et al., "Traffic engineering in hybrid software defined network via reinforcement learning," J. Netw. Comput. Appl., vol. 194, Art. no. 103136, Aug. 2021. doi: 10.1016/j.jnca.2021.103136. [Online]. Available: https://www.sciencedirect.com/science/article/abs/pii/S1084804521001363 [Google Scholar] [Crossref]

13. A. R. Roy et al., "Routenet: Using graph neural networks for SDN performance prediction," Int. J. Comput. [Google Scholar] [Crossref]

14. Appl., vol. 45, no. 1, p. 897, Jul. 2023. [Online]. Available: https://www.ijcert.org/index.php/ijcert/article/view/897 [Google Scholar] [Crossref]

15. I. Ampratwum et al., "Deep reinforcement learning and graph neural networks for WDM network optimization," in Proc. IEEE Conf. Commun., 2025, pp. 1-6. [Online]. Available: https://ieeexplore.ieee.org/iel8/8782664/10834807/11054280.pdf [Google Scholar] [Crossref]

16. V. Sujatha et al., "Optimizing software-defined networking (SDN) with machine learning: A comprehensive review," J. Qusay Young Sci., vol. X, no. Y, pp. 620-635, 2025. [Online]. Available: https://jqcsm.qu.edu.iq/index.php/journalcm/article/download/2193/1057 [Google Scholar] [Crossref]

17. S. Scott-Hayward et al., "A survey of machine learning techniques applied to software-defined networking (SDN): Research issues and challenges," IEEE Commun. Surveys Tuts., 2016 (repub. 2025). [Online]. Available: https://www.scribd.com/document/741205216 [Google Scholar] [Crossref]

18. J. M. Moynihan et al., "Using SDN and reinforcement learning for traffic engineering," in Proc. ICACCE, [Google Scholar] [Crossref]

19. Cape Town, South Africa, 2016, pp. 1-6. [Online]. Available: [Google Scholar] [Crossref]

20. https://pubs.cs.uct.ac.za/id/eprint/1162/1/ICACCE_2016_paper_102.pdfpubs.uct [Google Scholar] [Crossref]

21. R. R. Roy et al., "RouteNet: Leveraging graph neural networks for network modeling and prediction," arXiv:1910.01508, Oct. 2019. [Online]. Available: https://arxiv.org/abs/1910.01508arxiv [Google Scholar] [Crossref]

22. H. Kheddar et al., "Reinforcement-learning-based intrusion detection in SDN," IEEE Trans. Ind. Informat., vol. 20, no. 3, pp. 1-12, 2024. doi: 10.1109/TII.2024.10729241. [Online]. Available: https://ieeexplore.ieee.org/iel8/9739/5451756/10729241.pdfieeexplore.ieee [Google Scholar] [Crossref]

23. V. Sujatha et al., "Optimized efficient predefined time adaptive neural network for SDN stream traffic classification," Expert Syst. Appl., vol. 252, Art. no. 124207, 2025. doi: 10.1016/j.eswa.2025.124207. [Online]. Available: https://www.sciencedirect.com/science/article/abs/pii/S0957417425017075sciencedirect [Google Scholar] [Crossref]

24. Z. Su et al., "Deep reinforcement learning for SDN routing optimization," IEEE Trans. Netw. Service Manag., vol. 20, no. 2, pp. 1234-1245, 2023. [Google Scholar] [Crossref]

25. M. Xiao and Z. Zhang, "GNN-enhanced RL for QoS-aware SDN traffic engineering," Comput. Netw., vol. 235, Art. no. 109345, 2023. [Google Scholar] [Crossref]

26. J. Xie et al., "Machine learning for dynamic SDN control," IEEE Commun. Mag., vol. 57, no. 10, pp. 8894, 2019. [Google Scholar] [Crossref]

27. Y. Teng et al., "Hybrid causal-RL models for SDN performance," J. Syst. Archit., vol. 145, Art. no. 102789, 2023. [Google Scholar] [Crossref]

28. F. Bannour et al., "Algorithms and architectures for network analytics in SDN," BIMSA Lab Tech. Rep., no. 3964, 2023.ML_drivenSDN.docx [Google Scholar] [Crossref]

• "Towards GDP to GEP-Centric Model: A Proposed GEP Index Framework and Its Application in Haryana"
• Physicochemical Analysis of Petroleum Products in Selected Depots in Calabar Metropolis Cross River State and The Effects on Motor Engine
• Impact of Employee Welfare Measures on Workforce Well-Being in SBI: Insights from the Public Banking Sector
• Problem of Small-Scale Farmers in Agricutlure Sector in Tirunelveli Taluk
• Globalization, Economic Development, and Ecological Footprint in Tunisia: A QARDL Approach