Application of Large Language Models (LLMs) for Network Incident Diagnosis

The article examines the application of large language models for network incident diagnosis within AIOps. Traditional network management methods are shown to become ineffective under exponentially growing telemetry data volumes. Key LLM applications are analyzed: automated log analysis, localization of root causes of failures, natural language network interaction, and multi-agent systems. Technology limitations are considered: LLM hallucinations, excessive agency, vulnerability to attacks, privacy concerns, and computational complexity. It is concluded that LLMs open new opportunities for intelligent automation of network diagnostics.

1. Doyle K. How AIOps enables next-generation networking / K. Doyle // TechTarget [Электронный ресурс]. – URL: https://www.techtarget.com/searchnetworking/opinion/Next-gen-network-management-difficult-without-AIOps (дата обращения: 26.07.2026).

2. Mokdessi G. Cisco AI Assistant-Your Shortcut to Smarter, More Productive IT / G. Mokdessi, M. Sarmento // Cisco Blogs [Электронный ресурс]. – URL: https://blogs.cisco.com/networking/cisco-ai-assistant-your-shortcut-to-smarter-more-productive-it (дата обращения: 26.07.2026).

3. Wagar C. The progression of AIOps in today's network environment / C. Wagar // Nokia [Электронный ресурс]. – URL: https://www.nokia.com/blog/the-progression-of-aiops-in-todays-network-environment/ (дата обращения: 26.07.2026).

4. When AIOps Become "AI Oops": Subverting LLM-driven IT Operations via Telemetry Manipulation / D. Pasquini, E.M. Kornaropoulos, G. Ateniese [et al.] // arXiv [Электронный ресурс]. – URL: https://arxiv.org/abs/2508.06394 (дата обращения: 26.07.2026).

5. A Network Arena for Benchmarking AI Agents on Network Troubleshooting / Zh. Wang, A. Cornacchia, A. Sacco [et al.] // arXiv [Электронный ресурс]. – URL: https://arxiv.org/abs/2512.16381 (дата обращения: 26.07.2026).

6. Graph Structure-Enhanced Large Language Model for Optical Network Fault Diagnosis: An Explainable Alarm Root Cause Localization Approach / Y. Wang, Y. Pang, Y. Liu [et al.] // IEEE Internet of Things Journal. – 2025. – Vol. 12, Iss. 15. – P. 31493–31510.

7. Towards LLM-Based Failure Localization in Production-Scale Networks / Ch. Wang, X. Zhang, R. Lu [et al.] // Proceedings of the ACM SIGCOMM 2025 Conference. – New York: ACM, 2025. – P. 496–511.

8. Huang X. LogRules: Enhancing Log Analysis Capability of Large Language Models through Rules / X. Huang, T. Zhang, W. Zhao // Findings of NAACL 2025. – Association for Computational Linguistics, 2025. – P. 452–470.

9. RT-LogAAS: A Real-time Log Anomaly Analysis System based on Large Language Models for Net-Cloud / J. Wang, X. Cai, G. Yang [et al.] // CISAI '25: Proceedings of the 2025 8th International Conference on Computer Information Science and Artificial Intelligence. – New York: ACM, 2025. – P. 1658–1664.

10. ALPHA: LLM-Enabled Active Learning for Human-Free Network Anomaly Detection / X. Luo, Sh.M.N. Jha, A. Sinha [et al.] // 2025 IEEE International Performance, Computing, and Communications Conference (IPCCC). – IEEE, 2025. – URL: https://ieeexplore.ieee.org/document/11304694 (дата обращения: 26.07.2026).

11. MicroRCA-Agent: Microservice Root Cause Analysis Method Based on Large Language Model Agents / P. Tang, Sh. Tang, H. Pu [et al.] // arXiv [Электронный ресурс]. – URL: https://arxiv.org/abs/2509.15635 (дата обращения: 27.07.2026).

12. ThousandEyes [Электронный ресурс]. – URL: https://www.thousandeyes.com (дата обращения: 27.07.2026).

13. Marvis Virtual Network Assistant Overview // Juniper Networks [Электронный ресурс]. – URL: https://www.juniper.net/documentation/us/en/software/mist/mist-aiops/topics/concept/marvis-vna.html (дата обращения: 27.07.2026).

14. Hipolito M. HPE unveils agentic AI upgrades to Juniper for self-driving IT / M. Hipolito // IT Brief UK [Электронный ресурс]. – URL: https://itbrief.co.uk/story/hpe-unveils-agentic-ai-upgrades-to-juniper-for-self-driving-it (дата обращения: 27.07.2026).

15. Delivering Nokia Enhanced AIOps with the Right Foundations // Tech Field Day [Электронный ресурс]. – URL: https://techfieldday.com/video/delivering-nokia-enhanced-aiops-with-the-right-foundations (дата обращения: 27.07.2026).

16. Bushaus D. Project ONE aims to deliver agentic ODA / D. Bushaus // TM Forum [Электронный ресурс]. – URL: https://inform.tmforum.org/features-and-opinion/project-one-aims-to-deliver-agentic-oda (дата обращения: 27.07.2026).

17. LLM Excessive Agency | OWASP LLM06:2025 Explained // A10 Networks [Электронный ресурс]. – URL: https://www.a10networks.com/glossary/llm-excessive-agency/ (дата обращения: 27.07.2026).

18. Using lightweight LLMs to cut incident response times and reduce hallucinations // Help Net Security [Электронный ресурс]. – URL: https://www.helpnetsecurity.com/2025/08/21/lightweight-llm-incident-response/ (дата обращения: 27.07.2026).