Kubernetes is a popular container orchestration system that has seen massive adoption in both industry and academic IT departments. Long available in the Clouds, it has recently also become the main interface for several deployed and upcoming large-scale scientific compute facilities, like the Prototype National Research Platform (PNRP). The aim of this webinar is to provide science users that currently rely on traditional batch systems, like SLURM and HTCondor, a solid foundation that will allow them to effectively use Kubernetes-managed resources. The main focus is enabling the porting of batch-oriented scientific workflows, but we will also cover additional Kubernetes concepts not typically present in batch systems that could potentially be useful for science users.
-----
Instructor
Igor Sfiligoi
Lead Scientific Software Developer and Researcher, SDSC
Over the years, Igor Sfiligoi has been engaged in the computing community across a wide breadth of roles, participating among others as researcher, operator, developer, architect, and manager. He has been working in both academia and industry, which helps him appreciate the variety of problems computers help to solve. He has been working with both systems that support tightly coupled and worldwide-scale pleasantly parallel codes and realizes that each approach has its own advantages and challenges. In his current role at SDSC he is putting his extensive knowledge to use across many scientific domains, spanning computational biology, fusion research, high energy physics and astrophysics.