A Cognitive Serverless Framework for the Cloud-Edge Continuum
Title: A Cognitive Serverless Framework for the Cloud-Edge Continuum
DNr: Berzelius-2024-93
Project Type: LiU Berzelius
Principal Investigator: Zhou Zhou <zhou.zhou@umu.se>
Affiliation: Umeå universitet
Duration: 2024-03-04 – 2024-10-01
Classification: 10201
Homepage: https://cognit.sovereignedge.eu/


COGNIT is an AI-enabled Adaptive Serverless Framework for the Cognitive Cloud-Edge Continuum that enables the seamless, transparent, and trustworthy integration of data processing resources from providers and on-premises data centers in the cloud-edge continuum, and their automatic and intelligent adaption to optimize where and how data is processed according to application requirements , changes in application demands and behaviour, and the operation of the infrastructure in terms of the main environmental sustainability metrics. Specifically, a cloud-edge continuum infrastructure would be built first, then the virtual machine monitoring data would be generated from this infrastructure. The generated data would be used to train the machine learning based models. The ML models could predict the future workload of a certain virtual machine. Finally, the model would be integrated to cloud-edge continuum infrastructure to implement real-time prediction and model online fine-tuning.