Balancing the use of batteries and opportunistic scheduling policies for maximizing renewable energy consumption in a Cloud data center

Yunbo Li 1, 2, 3 Anne-Cécile Orgerie 4, 1 Jean-Marc Menaud 3, 2
1 MYRIADS - Design and Implementation of Autonomous Distributed Systems
Inria Rennes – Bretagne Atlantique , IRISA_D1 - SYSTÈMES LARGE ÉCHELLE
3 ASCOLA - Aspect and Composition Languages
Inria Rennes – Bretagne Atlantique , LS2N - Laboratoire des Sciences du Numérique de Nantes
Abstract : The fast growth of cloud computing considerably increases the energy consumption of cloud infrastructures, especially , data centers. To reduce brown energy consumption and carbon footprint, renewable energy such as solar/wind energy is considered recently to supply new green data centers. As renewable energy is intermittent and fluctuates from time to time, this paper considers two fundamental approaches for improving the usage of renewable energy in a small/medium-sized data center. One approach is based on opportunistic scheduling: more jobs are performed when renewable energy is available. The other approach relies on Energy Storage Devices (ESDs), which store renewable energy surplus at first and then, provide energy to the data center when renewable energy becomes unavailable. In this paper, we explore these two means to maximize the utilization of on-site renewable energy for small data centers. By using real-world job workload and solar energy traces, our experimental results show the energy consumption with varying battery size and solar panel dimensions for opportunistic scheduling or ESD-only solution. The results also demonstrate that opportunistic scheduling can reduce the demand for ESD capacity. Finally, we find an intermediate solution mixing both approaches in order to achieve a balance in all aspects, implying minimizing the renewable energy losses. It also saves brown energy consumption by up to 33% compared to ESD-only solution.
Type de document :
Communication dans un congrès
PDP 2017 - 25th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing , Mar 2017, St Petersburg, Russia. 2017
Liste complète des métadonnées

https://hal.inria.fr/hal-01432752
Contributeur : Anne-Cécile Orgerie <>
Soumis le : jeudi 12 janvier 2017 - 09:42:50
Dernière modification le : mardi 11 juillet 2017 - 09:48:47

Fichier

pdp.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01432752, version 1

Citation

Yunbo Li, Anne-Cécile Orgerie, Jean-Marc Menaud. Balancing the use of batteries and opportunistic scheduling policies for maximizing renewable energy consumption in a Cloud data center. PDP 2017 - 25th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing , Mar 2017, St Petersburg, Russia. 2017. <hal-01432752>

Partager

Métriques

Consultations de
la notice

511

Téléchargements du document

98