Show simple item record

hal.structure.identifierLaboratoire d'informatique de l'École polytechnique [Palaiseau] [LIX]
hal.structure.identifierCisco Systems France
dc.contributor.authorDesmouceaux, Yoann*
hal.structure.identifierLaboratoire d'analyse et modélisation de systèmes pour l'aide à la décision [LAMSADE]
dc.contributor.authorToubaline, Sónia*
hal.structure.identifierLaboratoire d'informatique de l'École polytechnique [Palaiseau] [LIX]
dc.contributor.authorClausen, Thomas
HAL ID: 176822
ORCID: 0000-0002-7400-8887
*
dc.date.accessioned2018-11-09T15:29:05Z
dc.date.available2018-11-09T15:29:05Z
dc.date.issued2018
dc.identifier.issn1573-7705
dc.identifier.urihttps://basepub.dauphine.fr/handle/123456789/18212
dc.language.isoenen
dc.subjectData center networkingen
dc.subjectVM migrationen
dc.subjectApplication-aware allocationen
dc.subjectMILPen
dc.subjectMulti-objective optimizationen
dc.subjectPareto optimalityen
dc.subject.ddc005en
dc.titleFlow-Aware Workload Migration in Data Centersen
dc.typeArticle accepté pour publication ou publié
dc.description.abstractenIn data centers, subject to workloads with heterogeneous (and sometimes short) lifetimes, workload migration is a way of attaining a more efficient utilization of the underlying physical machines. To not introduce performance degradation, such workload migration must take into account not only machine resources, and per-task resource requirements, but also application dependencies in terms of network communication. This paper presents a workload migration model capturing all of these constraints. A linear programming framework is developed allowing accurate representation of per-task resources requirements and inter-task network demands. Using this, a multi-objective problem is formulated to compute a re-allocation of tasks that (1) maximizes the total inter-task throughput, while (2) minimizing the cost incurred by migration and (3) allocating the maximum number of new tasks. A baseline algorithm, solving this multi-objective problem using the ε-constraint method is proposed, in order to generate the set of Pareto-optimal solutions. As this algorithm is compute-intensive for large topologies, a heuristic, which computes an approximation of the Pareto front, is then developed, and evaluated on different topologies and with different machine load factors. These evaluations show that the heuristic can provide close-to-optimal solutions, while reducing the solving time by one to two order of magnitudes.en
dc.relation.isversionofjnlnameJournal of Network and Systems Management
dc.relation.isversionofjnlvol26en
dc.relation.isversionofjnlissue4en
dc.relation.isversionofjnldate2018-10
dc.relation.isversionofjnlpages1034-1057en
dc.relation.isversionofdoi10.1007/s10922-018-9452-5en
dc.subject.ddclabelProgrammation, logiciels, organisation des donnéesen
dc.relation.forthcomingnonen
dc.relation.forthcomingprintnonen
dc.description.ssrncandidatenonen
dc.description.halcandidateouien
dc.description.readershiprechercheen
dc.description.audienceInternationalen
dc.relation.Isversionofjnlpeerreviewednonen
dc.relation.Isversionofjnlpeerreviewednonen
dc.date.updated2018-11-09T15:20:48Z
hal.identifierhal-01917762*
hal.version1*
hal.update.actionupdateMetadata*
hal.update.actionupdateFiles*
hal.author.functionaut
hal.author.functionaut
hal.author.functionaut


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record