Elastic Capacity Management for Slurm & Kubernetes Clusters Sharing AI Factory Infrastructure

This screencast shows how OpenNebula, thanks to its Elastic Capacity Management capabilities, enables Slurm and Kubernetes clusters to share AI infrastructure resources dynamically across isolated tenants. Using GPU nodes with NVIDIA InfiniBand as part of a testbed provided by CESGA (Spain), this demo shows elastic (re)allocation of compute, GPUs, and networking resources. It also demonstrates through InfiniBand benchmarking that PCI passthrough brings near-zero virtualization overhead to AI Factory environments.

Learn more about OpenNebula for AI Factory management: https://opennebula.io/ai-factory/

#aifactory #aiinfrastructure