Intent-Driven Deployment and Live Migration Using LLM
With the rise of large language models (LLMs), network management will require significantly less human intervention. This demo shows service deployment from a user request using simple natural language. We leverage LLM to decompose and translate the user request into TM Forum Intent representation. This intent is a blueprint for generating Kubernetes deployment manifest files, ultimately deploying the service to the designated location. This approach significantly reduces manual configuration and streamlines service deployment.
Authors
- John S. Vardakas (ORCID: https://orcid.org/0000-0003-1656-0507)
- Christos Verikoukis (ORCID: https://orcid.org/0000-0001-8774-1052)
- Golshan Famitafreshi (ORCID: https://orcid.org/0000-0003-0634-0041)
- João Fonseca (ORCID: https://orcid.org/0000-0003-4490-2746)
- Gereziher Adhane (ORCID: https://orcid.org/0000-0002-2616-5757)
- Kostas Ramantas (ORCID: https://orcid.org/0000-0002-1304-784X)
- Konstantinos Togias
- Swastika Roy (ORCID: https://orcid.org/0000-0002-8895-6562)
- Jhofre Ojeda
Institutions
- Industrial Systems Institute (GR)
- Iquadrat (Spain) (ES)
Publication Details
- Published
- 2024-10-21
- DOI
- https://doi.org/10.1109/camad62243.2024.10942727
- Primary Topic
- Mobile Agent-Based Network Management
- Type
- article
- Field-Weighted Citation Impact
- 0.00