Net2d-LLM: Translating Structured Network Intents into CLI using LLMs with Execution in a Network Digital Twin

• Jerônimo Menezes UNIPAMPA
• Leonardo Bitzki UNIPAMPA
• Diego Kreutz UNIPAMPA

DOI: https://doi.org/10.5753/errc.2025.17772

Resumo

This paper evaluates the use of Large Language Models (LLMs) to translate structured network intents, expressed as Desired State Models (DSMs), into vendor–specific CLI configurations. Net2d–LLM implements a deterministic, single–pass pipeline in which DSMs are rendered into constrained prompts and translated into CLI commands executed within a Network Digital Twin (NDT). The evaluation compares multiple LLMs under identical conditions, using latency, token usage, efficiency, and configuration consistency as assessment dimensions. The results show that general–purpose LLMs, when guided by structured prompts, can reliably generate syntactically valid and functionally correct configurations, demonstrating reproducibility and confirming the feasibility of LLM–driven automation in multivendor network environments.

Referências

Angi, A., Sacco, A., and Marchetto, G. (2025). LLNeT: An intent-driven approach to instructing softwarized network devices using a small language model. IEEE TNSM.

Fuad, A., Ahmed, A. H., Riegler, M. A., and Čičić, T. (2024). An intent-based networks framework based on large language models. In IEEE NetSoft, pages 7–12.

Hollósi, G., Ficzere, D., and Varga, P. Generative AI for low-level NETCONF configuration in network management based on YANG models. In IEEE CNSM, pages 1–7.

Hong, J., Tu, N., and Hong, J. A comprehensive survey on LLM-based network management and operations. 35(6):e70029.

Jeong, E.-D., Kim, H.-G., Nam, S., Yoo, J.-H., and Hong, J. W.-K. S-witch: Switch configuration assistant with LLM and prompt engineering. In IEEE NOMS, pages 1–7.

Lin, J., Dzeparoska, K., Tizghadam, A., and Leon-Garcia, A. AppleSeed: Intent-based multi-domain infrastructure management via few-shot learning. In IEEE NetSoft.

Lira, O. G., Caicedo, O. M., and da Fonseca, N. L. S. Large language models for zero touch network configuration management.

Liu, F., Farkiani, B., and Crowley, P. A survey on large language models for network operations & management: Applications, techniques, and opportunities.

Mekrache, A. and Ksentini, A. (2024). Llm-enabled intent-driven service configuration for next generation networks. In IEEE NetSoft, pages 253–257.

Mekrache, A., Ksentini, A., and Verikoukis, C. (2024). Intent-based management of next-generation networks: An llm-centric approach. Ieee Network, 38(5):29–36.

Mondal, R., Tang, A., Beckett, R., Millstein, T., and Varghese, G. What do LLMs need to synthesize correct router configurations? In 22nd ACM HotNets, pages 189–195.

Tageldien, M., Selim, B., and Sboui, L. (2025). Large language models in intent-based networking: a comprehensive survey across the intent lifecycle. In ITC-Egypt. IEEE.

Tu, N., Nam, S., and Hong, J. W.-K. (2025). Intent-based network configuration using large language models. International Journal of Network Management, 35(1):e2313.

Wang, C., Scazzariello, M., Farshin, A., Kostic, D., and Chiesa, M. Making network configuration human friendly.

Wei, Y., Xie, X., Hu, T., Zuo, Y., Chen, X., Chi, K., and Cui, Y. (2025a). Inta: Intent-based translation for network configuration with llm agents. In IEEE 33rd ICNP, pages 1–16.

Wei, Y., Xie, X., Zuo, Y., Hu, T., Chen, X., Chi, K., and Cui, Y. (2025b). Leveraging llm agents for translating network configurations. arXiv preprint arXiv:2501.08760.