Entropy: a Consolidation Manager for Clusters

Fabien Hermenier 1 Xavier Lorca 2 Jean-Marc Menaud 3 Gilles Muller 4 Julia L. Lawall 5
1 OASIS - Active objects, semantics, Internet and security
CRISAM - Inria Sophia Antipolis - Méditerranée , COMRED - COMmunications, Réseaux, systèmes Embarqués et Distribués
2 TASC - Theory, Algorithms and Systems for Constraints
LINA - Laboratoire d'Informatique de Nantes Atlantique, Département informatique - EMN, Inria Rennes – Bretagne Atlantique
3 ASCOLA - Aspect and composition languages
LINA - Laboratoire d'Informatique de Nantes Atlantique, Département informatique - EMN, Inria Rennes – Bretagne Atlantique
4 Regal - Large-Scale Distributed Systems and Applications
LIP6 - Laboratoire d'Informatique de Paris 6, Inria Paris-Rocquencourt
Abstract : Clusters provide powerful computing environments, but in prac-tice much of this power goes to waste, due to the static allocation of tasks to nodes, regardless of their changing computational re-quirements. Dynamic consolidation is an approach that migrates tasks within a cluster as their computational requirements change, both to reduce the number of nodes that need to be active and to eliminate temporary overload situations. Previous dynamic consol-idation strategies have relied on task placement heuristics that use only local optimization and typically do not take migration over-head into account. However, heuristics based on only local opti-mization may miss the globally optimal solution, resulting in un-necessary resource usage, and the overhead for migration may nul-lify the benefits of consolidation. In this paper, we propose the Entropy resource manager for ho-mogeneous clusters, which performs dynamic consolidation based on constraint programming and takes migration overhead into ac-count. The use of constraint programming allows Entropy to find mappings of tasks to nodes that are better than those found by heuristics based on local optimizations, and that are frequently globally optimal in the number of nodes. Because migration over-head is taken into account, Entropy chooses migrations that can be implemented efficiently, incurring a low performance overhead.
Type de document :
Communication dans un congrès
VEE 2009 - 5th International Conference on Virtual Execution Environments, Mar 2009, Washington, DC, United States. ACM, ACM SIGPLAN/SIGOPS, pp.41-50, 2009, <10.1145/1508293.1508300>
Liste complète des métadonnées


https://hal.archives-ouvertes.fr/hal-01102354
Contributeur : Xavier Lorca <>
Soumis le : lundi 12 janvier 2015 - 16:35:43
Dernière modification le : mardi 13 décembre 2016 - 15:44:43
Document(s) archivé(s) le : lundi 13 avril 2015 - 11:01:03

Fichier

hermenier-vee09.pdf
Fichiers éditeurs autorisés sur une archive ouverte

Identifiants

Citation

Fabien Hermenier, Xavier Lorca, Jean-Marc Menaud, Gilles Muller, Julia L. Lawall. Entropy: a Consolidation Manager for Clusters. VEE 2009 - 5th International Conference on Virtual Execution Environments, Mar 2009, Washington, DC, United States. ACM, ACM SIGPLAN/SIGOPS, pp.41-50, 2009, <10.1145/1508293.1508300>. <hal-01102354>

Partager

Métriques

Consultations de
la notice

571

Téléchargements du document

301