AI-Generated Network Configurations for Software-Defined Networking (SDN)

Software-Defined Networking (SDN) has significantly improved network management by separating the control plane from the data plane and providing centralized, programmable control of network resources. However, network configuration still requires considerable technical knowledge and manual effort, particularly when network requirements change frequently. This research examined the potential of Artificial Intelligence (AI), Generative AI, and Large Language Models (LLMs) to address these challenges by automatically generating network configurations from high-level, natural-language requirements. The literature review indicates that recent developments in LLM-based network management, Intent-Based Networking (IBN), Retrieval-Augmented Generation (RAG), multi-agent systems, and Infrastructure-as-Code (IaC) have created new possibilities for intelligent SDN configuration. Existing research demonstrates that AI can interpret user intent and generate machine-readable network configurations, reducing manual configuration effort and improving the efficiency of network management. However, the reviewed studies also identify important limitations, including configuration errors and hallucinations, insufficient topology awareness, vendor-specific implementations, limited formal verification, security concerns, and the need for human intervention before configurations are deployed. Based on these findings, this research proposes an AI-driven approach for generating SDN network configurations from natural-language network requirements. The proposed approach aims to combine AI-based intent interpretation with configuration generation and validation so that generated configurations can be evaluated for accuracy, correctness, consistency, and applicability before deployment. The research also aims to compare AI-generated configurations with manually created configurations in terms of configuration time, accuracy, consistency, and error rate. The proposed research has the potential to reduce repetitive work for network administrators and provide a more accessible method of managing  SDN  environments.  Nevertheless, AI-generated configurations should not be considered reliable for unrestricted autonomous deployment without appropriate validation and security mechanisms. Future development should therefore focus on formal verification,  real-time  network  monitoring, vendor-independent configuration generation, explainable AI, secure AI execution, digital twins, and continuous feedback mechanisms.   In conclusion, AI-generated network configurations represent a promising direction for the future of SDN management. By combining SDN programmability with the       natural-language understanding and configuration-generation capabilities of modern AI systems, network management can move from predominantly manual configuration toward more intelligent, automated, adaptive, and user-centric network operations. The successful implementation of such systems will depend not only on the accuracy of AI-generated configurations but also on their verification, security, explainability, and ability to operate reliably in dynamic and multi-vendor network environments.

1.  SDN foundation D. Kreutz, F. M. V. Ramos, P. E. Veríssimo, C. E. Rothenberg, S. Azodolmolky, and S. Uhlig, “Software-Defined Networking: A Comprehensive Survey,” Proceedings of the IEEE, vol. 103, no. 1, pp. 14–76, Jan. 2015.   2.  IBN survey A.      Leivadeas and M. Falkner, “A Survey on Intent-Based Networking,” IEEE Communications Surveys & Tutorials, vol. 25, no. 1, pp. 625–655, 2023, doi: 10.1109/COMST.2022.3215919.   3.  Intent-driven autonomous networking K. Mehmood, K. Kralevska, and D. Palma, “Intent-driven autonomous network and service management in future cellular networks: A structured literature review,” Computer Networks, vol. 220, p. 109477, 2023.   4.  LLM-based intent configuration N. Tu, S. Nam, and J.-K. Hong, “Intent-Based Network Configuration Using Large Language Models,” International Journal of Network Management, vol. 35, no. 1, p. e2313, 2025, doi: 10.1002/nem.2313.   5.  Agentic AI + IBN + IaC D. Brodimas, A. N. Birbas, D. Kapolos, and S. Denazis, “Intent-Based Infrastructure and Service Orchestration Using Agentic-AI,” IEEE Open Journal of the Communications Society, vol. 6, pp. 7150–7168, 2025, doi: 10.1109/OJCOMS.2025.3600706.   6.  Generative AI + IBN D. Adanza, L. Gifre, P. Alemany, C. Natalino, P. Monti, R. Muñoz, and R. Vilalta, “Leveraging generative AI for intent-based networking operations in network slices,” Computer Networks, vol. 272, p. 111647, 2025, doi: 10.1016/j.comnet.2025.111647.   7.  Verifiable LLM configuration C. Lin, L. Zhu, H. Liu, D. Zhang, J. Yu, and L. Liao, “DeepConfig: A verifiable configuration generation framework for MAN overlays using LLMs,” Computer Networks, vol. 280, p. 112177, 2026, doi: 10.1016/j.comnet.2026.112177.   8.  LLM network-management survey F. Liu, B. Farkiani, and P. Crowley, “Large Language Models for computer networking operations and management: A survey on applications, key techniques, and opportunities,” Computer Networks, vol. 271, p. 111614, 2025, doi: 10.1016/j.comnet.2025.111614.   9.  LLM-based network management survey J. Hong, N. V. Tu, and J.-W. K. Hong, “A Comprehensive Survey on LLM-Based Network Management and Operations,” International Journal of Network Management, vol. 35, no. 6, p. e70029, 2025, doi: 10.1002/nem.70029.   10.  NetConfEval C. Wang, M. Scazzariello, A. Farshin, S. Ferlin, D. Kostić, and M. Chiesa, “NetConfEval: Can LLMs Facilitate Network Configuration?,” Proceedings of the ACM on Networking, vol. 2, CoNEXT2, Art. no. 7, 2024, doi: 10.1145/3656296.