AI-Driven Framework for Scalable Management of Network Slices
Luis Blanco, Sławomir Kukliński, Engin Zeydan, Farhad Rezazadeh, Ashima Chawla, Lanfranco Zanzi, Francesco Devoti, Robert Kołakowski, Vasiliki Vlahodimitropoulou, Ioannis P. Chochliouros, Anne-Marie Bosneag, Sihem Cherrared, Luis A. Garrido, Sergio Barrachina‐Muñoz, Josep Mangues
Abstract
This article describes a scalable solution for orchestrating and managing a massive number of network slices that leverages Artificial Intelligence (AI) techniques to design robust and sustainable networks. To achieve this goal, the proposed approach decomposes the management and orchestration (M&O) plane using separation of concerns and uses AI techniques to automate M&O operations. The M&O automation is achieved through the use of multiple, distributed and AI-driven control loops. The control loops have different goals and may work on the node level, slice level, inter-slice level or orchestration domain level. We also present a case study of using the proposed distributed intelligent components to scale, optimize and improve the network infrastructure. Finally, we briefly describe some challenges and future directions for scalable M&O on the road to 6G.