The research paper “On the Integration of AI/ML-based scaling operations in the 5Growth platform” has received the best Fast Track paper Award in the IEEE Conference on Network Function Virtualization and Software Defined Networks (IEEE NFV-SDN’20).
This work has been authored by Mr. Jorge Baranda, Dr. Josep Mangues-Bafalluy, Dr. Engin Zeydan, Mr. Luca Vettori and Dr. Ricardo Martínez, researchers from CTTC, in collaboration with Dr. Xi Li, Dr. Andres Garcia-Saavedra (NEC), Dr. Carla Fabiana Chiasserini, Dr. Claudio Casetti (POLITO), Mr. Konstantin Tomakh, Mr. Oleksii Kolodiazhnyi (MIRANTIS) and Dr. Carlos Jesus Bernardos (UC3M) in the context of the H2020 European 5Growth project .
This paper presents the detailed architecture and first prototype of the 5Growth platform taking AI/ML-based network service autoscaling decisions to enforce the autonomous and automated assurance of vertical service level agreements. This entails the integration of a data engineering pipeline for real-time data gathering and model execution within the Service Orchestrator module of the 5Growth architecture. One interesting point of this data engineering pipeline is that it is based on open-source tools. This prototype has been evaluated in a real experimental infrastructure and the performed evaluation shows that AI/ML-related service handling operations are well below instantiation/termination lifecycle management procedures and that online classification can be performed in the order of hundreds of milliseconds.