GO-Docker: Batch scheduling with containers

Olivier Sallou 1 Cyril Monjeaud 1
1 Plateforme bioinformatique GenOuest [Rennes]
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, UR1 - Université de Rennes 1, Plateforme Génomique Santé Biogenouest®, Inria Rennes – Bretagne Atlantique
Abstract : Lightweight virtualization technologies gained attention by offering performance and effective scalability across cloud and physical architecture. GO-Docker is a new open source batch scheduling tool that provides container support (Docker). It is based on proven technologies and tools to provide job isolation and custom images for user jobs. Its architecture scales to handle large configurations and provides end-user easy access with a Web UI, CLI tools and API access for external programs integration. Containers provide job isolation, preventing resources overlap, and easier management for the cluster administrators. For the end-user, it provides a choice of operating systems, pre-built configurations and possible root access to the container. Its plugin architecture eases the integration of new scheduling algorithms or other execution/control mechanisms. The software targets multi-user systems with a central authentication (ldap, ...) and shared storage (home directory, shared data, etc.) and manages Docker access for users, leveraging security concerns with container access.
Liste complète des métadonnées

Cited literature [2 references]  Display  Hide  Download

https://hal.inria.fr/hal-01213323
Contributor : Olivier Sallou <>
Submitted on : Monday, October 26, 2015 - 7:29:03 AM
Last modification on : Thursday, November 15, 2018 - 11:58:39 AM
Document(s) archivé(s) le : Wednesday, January 27, 2016 - 10:51:36 AM

Files

poster_godocker.pdf
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

Citation

Olivier Sallou, Cyril Monjeaud. GO-Docker: Batch scheduling with containers. IEEE Cluster 2015, Sep 2015, CHICAGO, United States. ⟨http://www.mcs.anl.gov/ieeecluster2015/⟩. ⟨10.1109/CLUSTER.2015.89⟩. ⟨hal-01213323⟩

Share

Metrics

Record views

1241

Files downloads

555