Speaker
Description of the work
BiG Grid, the Dutch NGI started with a production IAAS HPC cloud infrastructure based on OpenNebula and KVM. With a self-service portal individuals and projects can have access to a virtualised infrastructure they can configure at will.
Offering users and projects such a development opportunity created a lot of creative projects and approaches.
Offering services such as Matlab, R and Galaxy workflows as a hosted infrastructure that scales on demand emerged as the way to go.
In the past 10 months we've been co-developing and implementing these type of compute services.
In this talk these experiences are presented, as well as issues and challenges.
Wider impact of this work
We think the community can learn from our development approach and can be inspired by the possibilities of having a IAAS / SAAS infrastructure that can be used as a piloting test-bed to quickly develop such dynamically scaling clusters.
We also welcome input for collaboration and discussion.
Link for further information
www.biggrid.nl
www.cloud.sara.nl
Printable Summary
BiG Grid, the Dutch NGI started with a production IAAS HPC cloud infrastructure based on OpenNebula and KVM. With a self-service portal individuals and projects can have access to a virtualised infrastructure they can configure at will.
In this talk experiences with R, Matlab(tm) and Galaxy (cloudman) are presented, where a clear shift from IAAS to SAAS is shown.
