ITIL V3进行VMWare容量管理的最佳实践方法.docx
- 文档编号:23202827
- 上传时间:2023-05-15
- 格式:DOCX
- 页数:12
- 大小:349.61KB
ITIL V3进行VMWare容量管理的最佳实践方法.docx
《ITIL V3进行VMWare容量管理的最佳实践方法.docx》由会员分享,可在线阅读,更多相关《ITIL V3进行VMWare容量管理的最佳实践方法.docx(12页珍藏版)》请在冰豆网上搜索。
ITILV3进行VMWare容量管理的最佳实践方法
ProvenPractice
ImplementingITILv3CapacityManagementinaVMwareenvironment
Introduction
ForthosefamiliarwiththeITILframeworkyouwillofcourseknowthatbyitsverynatureittriestoremainplatformagnostic.ButbasedonourexperienceworkingwithITILandVMware,anumberofkeyareashavebeenidentifiedasbeingafocuspointwherethespecificsofusingVMwarewillhaveaneffect.
ToprovidesomeguidancetopeoplenotfamiliarwiththeITILv3CapacityManagementterminologyIwillprovideabriefoverviewofeachsectionpriortodiscussingwherethekeyareasofinterestareinrelationtoVMware.
IntendedAudience
Thispaperaimstoprovidesomeinitialguidanceforpeoplethathaven’timplementedorareinthemiddleofimplementingITIL®v3CapacityManagementandhaveaninfrastructurethatcompriseseitherfullyorpartlyofVMwaretechnology.WewilldiscussthekeyareasofITILwhereVMwarewillhaveanimpactandprovideguidancebasedontheauthor’sownexperienceofimplementingITILprocessesandworkinginaVMwareenvironment.
TargetedatCapacityManagementandServiceManagementprofessionals,thiswillalsobeofinterestedtoVCPs.
Outline
1.BusinessCapacityManagement
2.ServiceCapacityManagement
3.ComponentCapacityManagement
4.Summary
1.BusinessCapacityManagement
Themainfocusforthebusinesscapacitysub-processistoensurethatthefuturebusinessrequirementsareunderstoodandthatwehavesufficientcapacitytomeetthoserequirements.ThisisourbridgebetweentheorganizationandtheCapacityManagementprocess.Thisinformationwillbeprovidedfromanumberofsourcesandwilldependonyourorganizationalstructure.InthepurelyITILsensetheyshouldcomefromthefollowingsources:
⏹ServiceStrategy
⏹ServicePortfolio
However,inthereal-worldthisinformationcouldbeprovidedbyarangeofadditionalsourcesthatmayinclude:
⏹ServiceLevelManager
⏹ServiceManager
⏹ITDirector
Wewouldbeexpectinginformationonchangestoexistingservices,anynewservices,expectedgrowthandimportantlypatternsofbusinessactivity.Itisthislatterareathatisofinterest;thepatternsofbusinessactivityareprovidedbytheDemandManagementprocess.
ThisprocesswillworkwithCapacityManagementtodeterminealong-termdemandmanagementstrategytomeettherequirementsofthebusiness.Allshort-termdemandmanagementisundertakenbytheoperationalCapacityManagementprocessandcouldincludetherestrictionofuseraccess,orbybalancingworkloads.ByusingthedynamicnatureoftheVMwareproductswehavebeenabletomoreeffectivelyfitthetechnologyaroundtheneedsofthebusiness.
Byanalyzingthebusinesspatterns(usingwhichevermethodyoucurrentlyuse,perhapsmonitoringkeytransactionratesandcombiningthosewithbusinessforecasts)youcandetermineanypotentialshorttermusagepeaksandrespondwithadditionalcapacity.Thiscanbedoneviaanumberofoptions:
⏹ThecreationofVirtualMachines,ifitsappropriateforthesystemandyouhaveasuitablelicensestructure
⏹ThrottlebacktheexistingresourceallocationonsomeofthekeyVMstoallowfortheincreaseindemand
⏹TheotheroptionandoneuniquetoVMwareistoutilizeDistributedResourceScheduler(DRS)combinedwithVMotion.Whenconfiguredcorrectlyi.e.selectingthecorrectlevelofautomationandmigrationlevelitispossibleforanyshort-termdemandmanagementactivitiestobemanagedautomatically.
PriortoimplementinganyoftheaboveoptionsitisessentialfromthecapacityplanningperspectivetoundertakeamodelinginitiativetodeterminewhatresourcelimitscanbeappliedtoboththeVirtualMachinesandtheunderlyinghardware.FromtheITILperspectivetherealsoneedstobeworkconductedwiththeConfigurationandChangeProcesstodetermineastandardprocessfortrackinganydynamicreallocationofVirtualMachines.
AkeygoalforCapacitymanagementistoprovidevaluetothebusiness;thiscanbedonetogreateffectusingVMwareandconsolidation.Usingconsolidationitispossibletoreducethenumberofphysicalserversandtoutilizethosereducedphysicalserversatamoreoptimalrate.Wehavecompletedanumberofconsultancyprojectsforourclientstoassistintheconsolidationoftheirinfrastructureandcertainlythekeytothisisintheplanningstage.Inordertoundertakethisexerciseamodelingtoolisused.Withinthiscontexttherearetwotypesofmodelingtool.
ForaconsolidationexercisewewouldlooktouseatoolsuchasVMwareCapacityPlannerthatwillanalyzethecurrentutilizationofyourserverestateanddetermineyouroptimumconsolidationrequirements
Oncetheconsolidationexerciseiscompletetherequirementsofyourmodellingtoolmaychange;obviouslyhavingourownmodelingtoolwearebiasedastoourchoice,butwhichevertoolyouchooseitisimportantthatitcontainsthefollowingfunctionality:
●Theabilitynotonlytomodeltheresourceconsumptionandtherelativeresponsetimeaswell
●IfyouareconsolidatingserversofdifferingOSsandplatformstheseneedtobesupported
●Toeliminateanyissueswithclockdrift,etc,itisrecommendedthatthedatabecapturedviaVCenter.
Goingthroughanyvirtualization/consolidationexerciseshouldgeneratesomeadditionalprocessmetricsthatcanbeusedtodemonstratetheeffectivenessoftheprocessandtheutilizationofVMware.WithinBusinessCapacityManagementwearelookingtoprovidemetricstodemonstrateourvaluetothebusinessandwhetherwearemeetingtheirrequirements.Thesortofmetricswecouldlookatproducingare:
⏹%reductioninphysicalestate
⏹%reductioninpowerconsumption(usefulasdocumentedevidenceinobtaininganyenvironmentalstandarde.g.ISO14001)
⏹%costreductioninmaintenancecontracts
2.ServiceCapacityManagement
Thissub-processisconcernedwithresourceconsumption,activitypatterns,peaksandtroughsoftheliveoperationalservices.WhilstitistheresponsibilityoftheServiceLevelManager(SLM)topolicetheresponsetimetargetssetwithintheServiceLevelAgreements(SLA),itistheCapacityManagementprocessthatprovidesthedatathatisused.
Asweknow,theresponsetimeofaservice/applicationisakeyperformancemetricanditisusedtodeterminetheenduserexperienceofusingaservicethatconventionalmetricslikeCPUutilization,memoryconsumption,etcwon’tnecessarilyshow.
Datarelatingtotheresponsetimeofaservicecanbedifficulttoobtain,butiscriticaltodetermininghowaserviceisperforming.ThissortofdatawillalsobecomemoreimportantwhenwelookatconsolidatingserversintoaVMwareenvironmentastheresourcelimitsthatareappliedwhenviewingthehardwareasawholewillhaveagreaterimpact.
Figure2showstherelationshipbetweenutilizationandtheresponsetime. Asyoucanseetherelationshipisn’talinearone;astheutilizationstartstogetcloserto100%theresponsetimestartstoincreasegreatly. AsdiscussedpreviouslythisissuewillbecomemoreapparentasthenumberofVirtualMachineshostedincreasesandwillneedperhapsmoremanagementthaniftheserverswerephysical. Thekeytomanagingthisistoensureaccuratecollectionoftheresponsetime. AgainthismayneedtobechangedwhenimplementingVMware,butitwilllargelydependonyourcurrentsetup. Ifyouareusinga“networksniffer”styletooltocaptureresponsetimedataateachstageofausertransactionthismayhavetobeamendedtoreflectachangeinstructureandwillcertainlyneedtobechangedifyouhaveusedVMwaretoconsolidateserversattheserviceleveli.e.allphysicalcomponentsthatculminatetoproduce“servicexyz”arenowononephysicalVMwarehost. Thesamewillofcoursebetrueifyouhaveconsolidatedallphysicalhostsfromadistributedenvironmenttoacentralizeddatacenter.
Ifyoucurrentlyuseaterminalemulationtooltographicallyrecreateanumberofstandarduserstepstocalculatetheresponsetime,thenyoumaynotneedtoalterthecaptureatall,butrememberthatthedatacapturedpriortoimplementationwillthenprovideausefulcomparisonfromtheSLAperspective. ItwillalsoprovideavaluablebusinessmetricastothebenefitsofmovingtoaVMwareenvironmenti.e.“WehaveconsolidatedServiceAtoVMwareandithasreducedtheuserresponsetimeby25%”.
Thecollectionofdatarelatingtoaservicecanbefurthercomplicatedwhencapturingdatafromservers/VirtualMachinesrunningdifferingoperatingsystems. We’vealreadydiscussedhowresponsetimemetricscangiveanindicationoftheoverallperformanceofaservice;buttodeterminetheperformanceoftheunderlyingcomponentsweneedtofallbackonthetraditionalmetricssuchasCPUutilization,memoryconsumptionetc.Thiswouldn’tbeanissueinanon-virtualenvironment(assumingyourtoolsetcancollectfrommultipleplatforms)butwhenyoumovetoVMwareyouneedtoconsiderthisdatacollectioncarefully.
UltimatelyyoueitherneedatoolthatiscapableofcollectingdatafromalloftheplatformsthatyouhavevirtualizedoryouneedtobeabletocollectitfromacentralsourcethathasvisibilityofalltheVirtualMachines.
InourexperiencethisiswherethedataprovidedbytheVCentercanproveinvaluable. FromthiscentralpointyoucanaccesskeymetricsonCPU,memoryconsumption,acrossallplatformsandVirtualMachines.
Whenthisdataiscombinedwithresponsetimedataitwillprovideavaluablecapacityplanningfeedforyouroperationalservices.
InadditiontotheVMwareestateyoumayalsoneedtocapturedatafromadditionalnon-virtualizedplatformsi.e.UNIX,mainframesetc. Ifaservice
- 配套讲稿:
如PPT文件的首页显示word图标,表示该PPT已包含配套word讲稿。双击word图标可打开word文档。
- 特殊限制:
部分文档作品中含有的国旗、国徽等图片,仅作为作品整体效果示例展示,禁止商用。设计者仅对作品中独创性部分享有著作权。
- 关 键 词:
- ITIL V3进行VMWare容量管理的最佳实践方法 V3 进行 VMWare 容量 管理 最佳 实践 方法