Unified Cyberinfrastructure with Kubernetes
Presenters
Event Type
Tutorial
Cloud and Distributed Computing
Containers
Datacenter
TUT
TimeMonday, 9 November 20202:30pm - 6:30pm EDT
LocationTrack 8
DescriptionKubernetes is emerging as the leading container orchestration solution. It can be deployed in the cloud and on-premise and supports federating several independent clusters into much larger logical resource pools.
One of Kubernetes advantages is its ability to effectively and securely co-schedule service containers alongside user containers. Service applications run with high privileges, with functionality and performance typical of bare metal deployments. User applications are given only basic privileges and limited resources, allowing for both a safe operating environment and fair resource sharing.
In this tutorial the attendees will learn how Kubernetes was used to create a nationwide cyberinfrastructure that is serving hundreds of scientific groups through NFS-funded projects like OSG, PRP, TNRP and CHASE-CI. The program includes a Kubernetes architectural overview, hands on sessions operating on the PRP production Kubernetes cluster and examples of how OSG and PRP are leveraging the Kubernetes cluster for facilitating scientific computing.
One of Kubernetes advantages is its ability to effectively and securely co-schedule service containers alongside user containers. Service applications run with high privileges, with functionality and performance typical of bare metal deployments. User applications are given only basic privileges and limited resources, allowing for both a safe operating environment and fair resource sharing.
In this tutorial the attendees will learn how Kubernetes was used to create a nationwide cyberinfrastructure that is serving hundreds of scientific groups through NFS-funded projects like OSG, PRP, TNRP and CHASE-CI. The program includes a Kubernetes architectural overview, hands on sessions operating on the PRP production Kubernetes cluster and examples of how OSG and PRP are leveraging the Kubernetes cluster for facilitating scientific computing.
Links
