Keynote: CERN Experiences - Ricardo Rocha & Clenimar Filemon
summary
CERN shows a working setup that bursts its OpenStack-backed fleet (about 320,000 cores and 200+ clusters) into public clouds to cover peak demand, claiming the batch system handles ~80% of compute and ~400,000 concurrent jobs. A host cluster runs a Kubernetes Federation control plane and deploys a daemonset to turn nodes into HTCondor compute hosts; clusters join with kubefed commands. They integrate datasets and solve cloud networking quirks across GKE, AKS, Amazon and Oracle cloud and demo a workflow spanning three clusters.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:31About Cern
- 4:18Antimatter Factory
- 4:34Ams Experiment
- 9:09Kubernetes and Federation
- 10:39Cases Regarding Federation at Cern
- 11:15Batch System
related talks
15mThe CERN Cloud Infrastructure: 5 years in 15 minutes
16mKeynote: Awards Ceremony
18mKeynote: Mind the Gap: Bridging Cloud Native Innovation with Real-World Use
41mBeyond Particle Physics: The Impact of Ceph and OpenStack on CERN's Multi-D...
16mA novel Cloud High Performance Computing infrastructure for user
14mKeynote: Science at Light Speed: Cloud Native Infrastructure for Astronomy Workloads
34mFishfingers and Custard: An OpenStack and K8s Story
47mCloud Native and Sustainable, Reproducible Scientific Computing