Abstract
The Nordic Tier-1 centre for LHC is distributed over several computing centres. It uses ARC as the internal computing grid middleware. ALICE uses its own grid middleware AliEn to distribute jobs and the necessary software application stack. To make use of most of the AliEn infrastructure and software deployment methods for running ALICE grid jobs on ARC, we are investigating different possible virtualisation technologies. For this a testbed and possible framework for bridging different middleware systems is under development. It allows us to test a variety of virtualisation methods and software deployment technologies in the form of different virtual machines.
Highlights
Grid computing is the central idea to allow analysis of the huge amount of data generated by the LHC [1] at CERN
We provided a default network via libvirt to enable network address translation (NAT) between the host and the guest
For CernVM one has to distinguish between the setup/startup time of the first VM which has to fetch the needed files from remote storage and the consecutive jobs that will fetch most of the files from the caches
Summary
Grid computing is the central idea to allow analysis of the huge amount of data generated (more than 10 PB/year) by the LHC [1] at CERN. The AliEn system does not have a transparent interface to the ARC job management, so the internal centres in NDGF appear as separate entities in the ALICE grid. Virtualisation helps to run tasks that need a complex software environment and allows to dynamically provide resources to different users, in order to optimise the capacity utilisation of a computing cluster.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have