Resources management on the Grid'5000 testbed
Abstract
Due to its origins in the High Performance Computing community, the Grid'5000 testbed has a fairly peculiar approach to resources management: all resources are tracked by a flexible HPC resources and tasks manager that is in charge of managing their lifecycle and allocating experiments (jobs). In this talk, we will discuss pros, cons, and lessons learned about this approach, as well as possible improvements that could be ported to other resources management approaches.
Format | Presentation |
---|---|
Origin | Files produced by the author(s) |