An MCP-based Solution for Managing Slices in Private 5G Networks
Resumo
The management of private 5G networks remains largely manual and depends on domain expertise. AI-driven agents can reduce this complexity by translating natural-language intents into concrete, executable actions, while the Model Context Protocol (MCP) offers a standardized way to expose network functions as formally defined tools. This paper examines an MCP-enabled AI agent for slice and subscriber management on open-source 5G Core (5GC) platforms. Its main contributions are an MCP server developed from scratch for 5G management, designed to integrate multiple 5GC platforms, and an empirical study of open-source Large Language Models (LLMs) in both CPU-only and GPU-accelerated settings under different prompt-engineering strategies. In CPU-only evaluations, the Granite 4.0 3B-H model provides the highest tool-selection accuracy, whereas the Qwen 2.5 7B model presents a lower latency and higher generation throughput. GPU acceleration improves performance, reducing latency by up to 130×. In general, the findings demonstrate that MCP-enabled AI agents offer a practical solution for managing private 5G networks.
Referências
Boateng, G. O., Sami, H., et al. (2026). A survey on large language models for communication, network, and service management: Application insights, challenges, and future directions. 28:527–566.
Brodimas, D., Birbas, A., et al. (2025). Intent-based infrastructure and service orchestration using agentic-ai. IEEE Open Journal of the Communications Society, 6:7150–7168.
Chakraborty, S., Chitta, N., and Sundaresan, R. (2024). Automation of network configuration generation using large language models. In 2024 20th International Conference on Network and Service Management (CNSM), pages 1–7.
Cheng, H., D’Oro, S., et al. (2024). Oranslice: An open source 5g network slicing platform for o-ran. In Proceedings of the 30th Annual International Conference on Mobile Computing and Networking, ACM MobiCom ’24, page 2297–2302, New York, NY, USA. Association for Computing Machinery.
Dao, T. and Gu, A. (2024). Transformers are ssms: Generalized models and efficient algorithms through structured state space duality. [link].
Eswaran, S. and Honnavalli, P. (2022). Private 5G networks: a survey on enabling technologies, deployment models, use cases and research directions: Private 5G networks: a survey on enabling technologies, deployment models, use cases and research directions. Telecommun. Syst., 82(1):3–26.
ETSI (2026). 5G; System architecture for the 5G System (5GS). Technical Specification ETSI TS 123 501, European Telecommunications Standards Institute (ETSI). 3GPP TS 23.501 version 19.6.0 Release 19. Accessed: 2026-01-31.
Grattafiori, A., Dubey, A., et al. (2024). The llama 3 herd of models. [link].
Hong, J., Tu, N. V., and Hong, J. W.-K. (2025). A comprehensive survey on llm-based network management and operations. International Journal of Network Management, 35(6):e70029. e70029 NEM-25-0670.
Jiang, A. Q., Sablayrolles, A., et al. (2023). Mistral 7b. [link].
Luo, J., Zhang, W., et al. (2025). Large language model agent: A survey on methodology, applications and challenges. [link].
Mangipudi, G. M. and Eswaran, S. (2025). Network slicing in 5g: A survey on the concepts, use cases, testbeds and open challenges. In 2025 1st International Conference on Emerging Trends in Information Systems and Informatics (ICETISI), pages 1–7.
Peterson, L. and Sunay, O. (2020). 5G Mobile Networks: A Systems Approach. Morgan & Claypool Press, 1 edition.
Phyu, H. P., Naboulsi, D., and Stanica, R. (2023). Machine learning in network slicing—a survey. IEEE Access, 11:39123–39153.
Qwen, :, Yang, A., Yang, B., et al. (2025). Qwen2.5 technical report. [link].
Saon, G. and others, A. D. (2025). Granite-speech: open-source speech-aware llms with strong english asr capabilities. arXiv.
Sapkota, R., Roumeliotis, K. I., and Karkee, M. (2026). Ai agents vs. agentic ai: A conceptual taxonomy, applications and challenges. Information Fusion, 126:103599.
The Linux Foundation and CAMARA Project (2025). In Concert: Bridging AI Systems Network Infrastructure through MCP: How to Build Network-Aware Intelligent Applications. Technical report. 12 pages.
Usman, Y., Oladipupo, H., et al. (2025). Ai, ml, and llm integration in 5g/6g networks: A comprehensive survey of architectures, challenges, and future directions. IEEE Access, 13:168914–168950.
Wang, C., Zhang, X., et al. (2025). Towards llm-based failure localization in production-scale networks. In Proceedings of the ACM SIGCOMM 2025 Conference, SIGCOMM ’25, page 496–511, New York, NY, USA. Association for Computing Machinery.
Wijethilaka, S. and Liyanage, M. (2021). Survey on network slicing for internet of things realization in 5g networks. IEEE Communications Surveys & Tutorials, 23(2):957–994.
