Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 66: Grid-Computing

T 66.4: Vortrag

Dienstag, 28. März 2017, 17:35–17:50, JUR 372

On-demand provisioning of HEP compute resources on shared HPC centers — Frank Fischer, •Günther Erli, Manuel Giffels, Thomas Hauth, and Günter Quast — Karlsruher Institut für Technologie

Rather than solely relying on dedicated HEP computing centers, it is nowadays more reasonable and flexible to utilize remote computing capacity via virtualization techniques.

Since last year the remote HPC center (NEMO Cluster, Freiburg University) was scaled by a factor of 20. This contribution reports on challenges, experiences and recent developments with scaling dynamic deployment from a small prototype system to a TOP500 cluster.

The remote and local systems are tied together with the ROCED scheduler [1] such that, from the user perspective, local and remote resources form a uniform, virtual computing cluster with a single point-of-entry. This environment serves the needs of researchers participating in the CMS, AMS and Belle II experiments at KIT. A job volume of over 2 million CPU hours per month (in average) is managed with this system.

[1] O. Oberst et al. Dynamic Extension of a Virtualized Cluster by using Cloud Resources, J. Phys.: Conference Ser. 396(3)032081, 2012