On Maximum Elastic Scheduling in Cloud-Based Data Center Networks for Virtual Machines with the Hose Model

在线阅读 下载PDF 导出详情
摘要 Withthegrowingpopularityofcloud-baseddatacenternetworks(DCNs),taskresourceallocationhasbecomemoreandmoreimportanttotheefficientuseofresourceinDCNs.Thispaperconsidersprovisioningthemaximumadmissibleload(MAL)ofvirtualmachines(VMs)inphysicalmachines(PMs)withunderlyingtree-structuredDCNsusingthehosemodelforcommunication.Thelimitationofstaticloaddistributionisthatitassignstaskstonodesinaonce-and-for-allmanner,andthusrequiresaprioriknowledgeofprogrambehavior.Toavoidloadredistributionduringruntimewhentheloadgrows,weintroducemaximumelasticityscheduling,whichhasthemaximumgrowthpotentialsubjecttothenodeandlinkcapacities.Thispaperaimstofindtheschedulewiththemaximumelasticityacrossnodesandlinks.Wefirstproposeadistributedlinearsolutionbasedonmessagepassing,andwediscussseveralpropertiesandextensionsofthemodel.Basedontheassumptionsandconclusions,weextendittothemultiplepathscasewithafattreeDCN,anddiscusstheoptimalsolutionforcomputingtheMALwithbothcomputationandcommunicationconstraints.Afterthat,wepresenttheprovisionschemewiththemaximumelasticityfortheVMs,whichcomeswithprovableoptimalityguaranteeforafixedflowschedulingstrategyinafattreeDCN.Weconducttheevaluationsonourtestbedandpresentvarioussimulationresultsbycomparingtheproposedmaximumelasticschedulingschemeswithothermethods.Extensivesimulationsvalidatetheeffectivenessoftheproposedpolicies,andtheresultsareshownfromdifferentperspectivestoprovidesolutionsbasedonourresearch.
机构地区 不详
出版日期 2019年01月11日(中国期刊网平台首次上网日期,不代表论文的发表时间)
  • 相关文献