BEAUMONT, Olivier; BELAID, Ikbel; EYRAUD-DUBOIS, Lionel; LORENZO-DEL-CASTILLO, Juan-Angel

doi:10.1007/978-3-662-48096-0_12

Métadonnées

Licence d’utilisation du document

BEAUMONT, Olivier
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Reformulations based algorithms for Combinatorial Optimization [Realopt]

BELAID, Ikbel
Institut de Mathématiques de Bordeaux [IMB]
Reformulations based algorithms for Combinatorial Optimization [Realopt]

EYRAUD-DUBOIS, Lionel
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Reformulations based algorithms for Combinatorial Optimization [Realopt]

Langue

Communication dans un congrès

Ce document a été publié dans

Euro-Par 2015, 2015, Vienna. 2015-07-25

Résumé en anglais

In the context of service hosting in large-scale datacenters, we consider the problem faced by a provider for allocating services to machines. Based on an analysis of a public Google trace correspond-ing to the use of a production cluster over a long period, we propose a model where long-running services experience demand variations with a periodic (daily) pattern and we prove that services following this model acknowledge for most of the overall CPU demand. This leads to an allo-cation problem where the classical Bin-Packing issue is augmented with the possibility to co-locate jobs whose peaks occur at different times of the day, which is bound to be more efficient than the usual approach that consist in over-provisioning for the maximum demand. In this paper, we provide a mathematical framework to analyze the packing of services exhibiting daily patterns and whose peaks occur at different times. We propose a sophisticated SOCP (Second Order Cone Program) formula-tion for this problem and we analyze how this modified packing constraint changes the behavior of standard packing heuristics (such as Best-Fit or First-Fit Decreasing). We show that taking periodicity of demand into account allows for a substantial improvement on machine utilization in the context of large-scale, state-of-the-art production datacenters.< Réduire