Keynote CargoDay

• Richard Randriatoamanana (CNRS)
• Tristan Le Toullec (LPO)
• Franck CHARRON (CNRS)

Room: Amphithéâtre S

La nouvelle infrastructure de virtualisation de l'Ecole Centrale de Nantes (VMware vSphere 6)

• Patrick Guyomarc'h (Ecole Centrale de Nantes)

Room: Amphithéâtre S L'ECN s'est dotée en 2014 d'une nouvelle infrastructure globale de virtualisation. Cette infrastructure, totalement résiliente, est répartie sur 2 bâtiments du Campus et s'appuie sur les 2 coeurs de réseau Cisco Nexus 5548. Elle implémente les technologies Cisco UCS Blade, Netapp FAS (7-Mode et Cluster Mode) en Metrocluster. Sur le plan strictement "Virtualisation", nous avons adopté VMware vCenter/vSphere (v6) et exploitons ses fonctionnalités HA/DRS/vMotion net Storage vMotion. Quant au volet "Sauvegarde", il est entièrement assuré par le logiciel Veeam (v9). L'exposé consistera, - dans un premier temps, à présenter le socle matériel (Nexus, UCS Blade, Metrocluster) sur lequel repose la fiabilité de la solution. - et dans un second temps, nous nous attarderons sur les aspects plus logiciels (VMware, Veeam, gestion des vLAN, gestion des Volumes NFS...)

Revising OpenStack to operate the next generation of Cloud Computing platforms

• Adrien LEBRE (INRIA EMN)

Room: Amphithéâtre S Academic and industry experts are now advo cating for going from large-centralized Cloud Computing infrastructures to smaller ones massively distributed at the edge of the network. Among the obstacles to the adoption of this model is the development of a convenient and powerful IaaS system capable of managing a significant number of remote data-centers in a unified way. In this talk, we introduce the premises of such a system. The novelty of our work is that instead of developing a system from scratch, we revised the OpenStack solution in order to operate such an infrastructure in a distributed manner leveraging P2P mechanisms. After describing how we revised the Nova service by leveraging a distributed key/value store instead of the centralized SQL backend, we will present experiments that validated the correct behavior of our prototype, while having promising performance using several clusters composed of servers of the Grid’5000 testbed. We believe that such a strategy is promising and paves the way to a first large-scale and WAN- wide IaaS manager.

IaaS OpenStack LOPS

• Tristan Le Toullec (LPO)
• Thibault LE BARS (LOPS/UBO)

Room: Amphithéâtre S Mise en place d'une Infrastructure As A Service au LOPS

GoDocker: batch job submission using Docker containers

• Olivier SALLOU (IRISA)

Room: Amphithéâtre S GoDocker is a job scheduler like Sun Grid Engine, Slurmm, ... but makes use of Docker containers to execute the jobs, providing job isolation and using pre-built images with needed software.

ProxMox VE IUEM

• Christophe MARTIN (CNRS / LDO)

Room: Amphithéâtre S Exploitation d'un cluster ProxMox VE avec stockage iSCSI/CLVM

Feedbacks Room: Amphithéâtre S