Dynamic multi-provider cluster autoscaling for the Computing Continuum

The paper presents an innovative architecture for automatic scaling of computational resources across the Cloud-Edge Continuum, overcoming the limitations of traditional Kubernetes autoscaling systems, which are generally confined to a single provider.

The solution leverages Liqo, an open-source framework for Kubernetes cluster federation, and introduces an autoscaling mechanism capable of dynamically identifying and utilizing resources distributed across independent providers.

Through allocation policies that account for resource availability and costs, the approach enables more flexible and efficient management of federated infrastructures.

Preliminary experimental results demonstrate improved resource utilization, particularly in environments characterized by dynamic workloads.