基于.资源分配的云计算数据中心的节能算法的研究

《基于.资源分配的云计算数据中心的节能算法的研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、万方数据ABSTRACTCloudcomputinghasbeenwidelyusedindifferentvendersduetoitsadvantagesofferedbythepublic，privateandhybridcloud，despiteofmanybenefitstothecloudprovidersandcloudusers。therearestillseveralchallengesfacedinmaintainingandmanagingthecloud．Oneofthekeychallengesisthei

2、nefficiencyofenergyconsumptionincloudcomputingdatacenters，datacentersconsumeSOmuchpowerandsomeofthesepowersarenotevenusedforresourcescomputation．OneofthemaincausesofenergyinefficiencyincloudcomputingdatacentersiStheinemcientofresourcesallocation．Thereareseveralfactorsi

3、nfluencingtheimproperresourcesallocationincloudcomputingenvironments，SOtheresourcesallocationremainsasaproblem．Thuseffectiveandefhcientmethodsneededtobeproposedtoovercomethisissue．Inrecentyears．virtualmachinemigrationmethodhasbeenwidelyusedastheWayofutilizingtheresourc

4、estotheeffectively．andtheresearchesofeffectivevirtualmachines(VMs)migrationpoliciesandalgorithmsofresourcesallocationhavebeenthehottopics．Thisthesisfocusesontwomainparts：thefirstpartiStoimplementaproposedvirtualmachinesselectionpolicyforVMmigration．whichiSknownasthe1’】

5、1resholdMaximumUtilization(ThrMaxU)policy．Inthispolicy，wesetfixedupperandlowerhostutilizationthreshold，dynamicallyinaspecifiedtimewecheckifupperandlowerthresholdareviolated．ⅥmentheutilizationofahostiSbelowthelowerutilizationthreshold．a11VMsinthathostshouldbemigratedand

6、thehostisshutdown．Incasewhentheutilizationofsomehostsareabovetheupperutilizationthreshold，wearrangethehostsviolatedtheupperutilizationindecreasingorder，startingwiththehostwhichexceededutilizationportionmostandVMSwiththehigherCPUutilizationamongofthehostsexperiencingvio

7、lationofupperorlowerutilizationthresholdareselectedfirst．WewilladoptthisselectionpolicyandintegrateitwiththeexistingallocationpolicycalledModifiedPowerFitDecreasing(MBFD)toevaluatethepowerandperformancetrade—off．WeuseCloudSimtoolkittosimulateOurworkandcomparetheperform

8、anceofourproposedpolicywithotherbenchmarksintermofpowerconsumption，percentageofSLAviolation，numberofhostsshutdownandn