EMC RecoverPoint产品介绍

EMCRecoverPoint产品介绍假设场景EMCLSIIBMHDSHPSAN应用服务器组困扰：不能支持异构存储；无法恢复逻辑错误；无法恢复到任意时间点；成本高昂；……Agenda什么是RecoverPoint？RecoverPoint的特点RecoverPoint的使用场景什么是CDP？定义： “持续数据保护是一套方法，它可以捕获或跟踪数据的变化，并将其独立存放在生产数据之外，以确保数据可以恢复到过去的任意时间点。持续数据保护系统可以基于块、文件或应用实现，可以为恢复对象提供足够细的恢复粒度，实现几乎无限多的恢复时间点。” －－SNIA(存储网络工业协会)CDPvs.传统备份：RPO不同CDPvs.数据复制：逻辑错误的处理方式CDP的实现模式基于文件基于数据块基于主机端卷管理软件或客户端代理软件Agent；基于传输层；基于存储层；RecoverPoint系列数据保护、复制和灾难恢复解决方案RecoverPoint/SE每站点单个CLARiiONCX、CX3或CX4阵列CDP、CRR和CLR数据保护集成了WAN带宽缩减功能利用CLARiiON拆分器支持Windows、Linux、VMware、Solaris、HP-UX和AIX分层的存储容量：4/8/16/24/32->150TB可升级到完整版RecoverPoint本地保护CDP远程数据CRR并行本地和远程(CLR)保护RecoverPoint多个异构存储器和服务器CDP、CRR和CLR数据保护集成了WAN带宽缩减功能支持基于主机、CLARiiON或网络的拆分器以1TB为增量逐步增加容量，最多达600TB与Brocade和Cisco提供的智能结构集成本地保护远程保护并行本地和远程保护RecoverPoint配置RecoverPoint是一个多节点群集RecoverPoint/SE和RecoverPoint为2至8个节点所有节点都在双活配置中可以跨节点分布负载可对运行中的系统添加节点基于主机、阵列或结构的写操作拆分器主机或CLARiiON拆分器或Brocade及基于Cisco结构的拆分器上的轻量级驱动程序仅截取受保护LUN的写操作将写操作发送到本地应用装置RecoverPoint/SE仅支持Windows和CLARiiON拆分器支持镜像的SAN两个独立的SAN提供了可靠性支持无中断结构升级通过故障隔离提供RecoverPoint配置的高可用性支持Cisco和Brocade（包括M机型）连接结构支持iSCSI和4Gb/s光纤通道RecoverPoint应用装置提供本机4Gb/s光纤支持CLARiiON拆分器支持光纤通道和iSCSI卷主机、连接结构或CLARiiON拆分器存储阵列主机第2层SAN（A/B结构）更新更新更新更新更新更新更新更新更新更新目标阵列管理员设置重要性和RPO策略中午12:00早晨6:00下午6:00中午12:0020分钟RPO6分钟RPO源阵列RecoverPoint监视带宽并优化资源利用率在RecoverPoint/SE外部更改带宽WAN带宽节约管理员指定带宽消耗策略源CG1目标CG1源CG2目标CG2RecoverPoint管理本地和/或远程监视和管理复制设置和修改复制策略与VMwarevCenterSiteRecoveryManager和vCenter集成；支持MicrosoftHyper-V包括在线帮助按应用程序向用户授权包括自动化向导还提供命令行界面(CLI)使用单一界面复制SAN连接的卷WriteSplittersHost-based,fabric-based,andCLARiiON-basedwritesplittersInterceptswritesfromthehostandsplitseachwriteManagedthroughRecoverPoint;installedandupgradedindependentlyofRecoverPointRecoverPoint/SEonlysupportsaWindowshost-basedwritesplitterandtheCLARiiON-basedwritesplitter.SplitterTypeHowDeployedOverheadHost-basedInI/Ostackjustabovethemulti-pathsoftwareAddswritetrafficattheHBA;nootherimpactFabric-basedInintelligentstorageserviceshardwareonaBrocade-orCisco-basedswitchOperatesatwirespeeds;noimpactCLARiiON-basedInFLAREoperatingsystem;activeinbothstorageprocessorsNoimpactHost-BasedWriteSplittersSupportedonAIX*,Solaris*,andWindowsLightweightcodethatinstallsoneachserverthataccessesareplicatedLUNSplitsor‘Duplicates’writestoproductionLUNsandsendsacopytotheRPA*AIXandSolarishost-basedwritesplittersarenotsupportedbyRecoverPoint/SEHOST(WINDOWS)UserSpaceKernelFileSystemPartitionManagerKshsplit.sysDiskClassDriver(e.g.,Disk.sys)MultipathDriverHBAMiniport/StorportDriverMonitorthesplitterprocessListenforLogRequest,collectandtransfertoRPAoverIPCommunicationwiththeRPAoverFCWatchdogServiceSplitterUserSpaceHIC/HLRVariousapplicationprocesses......HBAHBANICFabric-BasedWriteSplittersRecoverPointleveragesintelligentfabricservicesSupportsBrocadeFabricApplicationPlatformSupportsCiscoSANTapSupportsEMCandnon-EMCStoragearraysFullsupportforVMware,includingVMFSNotsupportedbyRecoverPoint/SEApplicationsTargetAPIsCopyofwrite(splitwrite)ApplianceNotinprimarydatapathOriginalwriteBrocadeFabric-SplitterPlatformsDepartmentandDirectorversionsConnectrixAP-7600BfordepartmentPB-48K-AP4-18switchingmoduleforConnectrixB-seriesED-48000BandED-DCX-BdirectorsFeaturesHigh-performancestorageprocessor164Gb/sports2GigabitEthernetportsNoRecoverPointorBrocadelicenserequiredNotsupportedbyRecoverPoint/SEPB-48K-AP4-18ConnectrixAP-7600BCiscoFabric-SplitterPlatformsMDS-9000usesthe18/4blade9222icanusethe18/4bladeandhasnativefabricsplittingsupportBothMDSand9222irequireaCiscoStorageServicesEnabler(SSE)license,noRecoverPointlicenserequiredNotsupportedbyRecoverPoint/SEConnectrixMDS-9222iNativesupportorwithMDS18/4Multi-ServicesbladeConnectrixMDS-9000SerieswithMDS18/4Multi-ServicesbladeMDS18/4Multi-ServicesBladeCLARiiON-BasedWriteSplitterRecoverPointsplitterincludedatnochargewithFLARE26andlaterSupportsCLARiiONCX3andCX4arraysSupportsCelerraNSarrays(CelerraUnified)SupportsopensystemsCanbemixedwithotherRecoverPointsplittersEasiesttouseandconfigureCanbesharedbyuptofourRecoverPointclustersSolarisVMwareSolarisVMwareCLARiiONsplitterrunsoneachCLARiiON;nohostorfabricagentrequired—cansharebetweenclustersSolarisVMwareLinuxWindowsPRODUCTIONAPRODUCTIONDISASTERRECOVERYSolarisVMwareLinuxWindowsRecoverPointSANSAN/WANRecoverPointSANRecoverPointSANFC/WANRecoverPointSANSplitterSupportbyOSOS/SplitterTypeHost-basedBrocadeCiscoCLARiiONAIX5.2,5.3YesYesYesHP-UXNoYesYesYesHyper-VHyper-VandHyper-VR2YesYesYesLinuxNoYesYesYesSolaris10,11YesYesYesVMwareVMwithWindowsRDM/PonlyYesYesYesVMwareVMwithotherNoYesYesYesWindows32and64bit,Server2003,2008,2008R2YesYesYesOtherOSNoSeeESMSeeESMSeeESMSplitterguidelinesDifferentsplittertypescanbeintermixedMultipleinstancesofasplittertypeissupportedCLARiiON-basedwritesplittercanbesharedTheCisco-SANTapIntelligent-FabricsplittercanbesharedwithanRPQMultipleversionsofsplittersaresupportedSplitterHost-basedBrocadeCisco-SANTapCLARiiONHost-basedYesYesYesYesBrocadeYesYesNoYesCisco-SANTapYesNoYesYesCLARiiONYesYesYesYesLocalProtectionProcess—ContinuousDataProtection(CDP)1.DataissplitandsenttotheRecoverPointapplianceinoneofthreeways4. Theappliancewritesdatatothejournalvolume,alongwithtimestampandapplication-specificbookmarks5. Write-order-consistentdataisdistributedtothereplicavolumesProductionvolumesReplicavolumesJournalvolume2a.Hostsplitter3. WritesareacknowledgedbackfromtheRecoverPointappliance2b. Intelligent-fabricsplitter2c.CLARiiONsplitter/A/C/BrArCrBDataFlow-CDPRemoteProtectionProcess—ContinuousRemoteReplication(CRR)Localsite2a.Hostsplitter2b. Intelligent-fabricsplitter7. DataiswrittentothejournalvolumeRemotesiteJournalvolume5. Dataissequenced,checked,compressed,andreplicatedtotheremoteappliancesoverIPorSAN1. DataissplitandsenttotheRecoverPointapplianceinoneofthreeways8.Consistentdataisdistributedtotheremotevolumes2c.CLARiiONsplitter3. WritesareacknowledgedbackfromtheRecoverPointappliance4.AppliancefunctionsFibreChannel-IPconversionReplicationDatareductionandcompressionMonitoringandmanagement6. Dataisreceived,uncompressed,sequenced,andverifiedDataFlow-CRRRecoverPoint/SEandRecoverPoint*RefertotheEMCSupportMatrixfordetailsFeaturesRecoverPoint/SERecoverPointOperatingsystemsupportedWindowsServer2000,2003,2008HeterogeneouswithCLARiiONsplitterHeterogeneous*StoragearrayssupportedCLARiiONAX4-5,CX4,CX3,andCXseries;blockLUNsfromFibreChannelCelerraNS20,NS40,NS80,NS-120,NS-240,andNS-960Heterogeneous*NumberofarraysTwoinCRR/CLR,oneinCDPUnlimitedConcurrentlocalandremote(CLR)dataprotectionSupportedSupportedLicensingPerreplicatedcapacityPerreplicatedcapacityNumberofappliancesTwotoeightpersiteTwotoeightpersiteBandwidthreductionBuiltinBuiltinJournalcompressionNotsupportedBuiltinVirtualizationsupportHyper-V,VMwarevCentermonitoring,VMwareSRMHyper-V,VMwarevCentermonitoring,VMwareSRMMulti-pathingEMCPowerPathonlyHeterogeneousCapacityAslicensedfor4TB,8TB,16TB,24TB,32TB,and33TB-300TBin1TBincrementsAslicensed(upto600TB)SynchronousreplicationSupportedSupportedIntelligentfabricNotsupportedSupportedRecoverPoint的特点AgendaRecoverPoint的价值任意时间点数据的恢复可支持异构存储降低总体拥有成本RecoverPoint的使用场景Agenda生产站点灾难恢复（远程）站点灾难恢复UNIXWindowsUNIXWindows初始同步更新更新CRR目标阵列源阵列SANSAN目标CG3目标CG2目标CG1WANRecoverPoint应用装置RecoverPoint应用装置源CG3源CG2源CG1灾难恢复生产故障将故障切换到CRRLUN更新CRRLUN灾难生产站点UNIXWindows灾难恢复（远程）站点UNIXWindowsCRR目标阵列源阵列WANRecoverPoint应用装置RecoverPoint应用装置更新SANSAN源CG3源CG2源CG1灾难恢复（远程）站点灾难恢复恢复重新同步LUN回切到生产恢复UNIXWindowsUNIXWindowsCRR目标阵列源阵列RecoverPoint应用装置RecoverPoint应用装置CRR源阵列目标阵列更新生产站点SANSANWAN操作恢复从生产卷到本地目标的本地CDP复制对每个写操作拍摄快照在日志中跟踪所有更改零RPO根据策略整合快照将恢复点整合到每日、每周或每月恢复点通过保留更多快照来扩展RPO降低每个PRO消耗的存储量生产站点UNIXWindows日志本地拷贝用于数据挖掘远程拷贝用于备份可写快照用于测试备份、测试、决策支持通过复制避免影响生产应用程序生产站点远程站点生产应用程序生成决策支持工具同步复制软件升级测试备份同步或异步复制可写快照阵列快照数据迁移“近即时”时间点(PIT)回滚从生产数据到备用阵列高效克隆映像提供更及时的信息访问在本地站点上利用数据进行备份、测试、决策支持易于为多个服务器撒播数据源站点生产CRM阵列1阵列2阵列3阵列4CDPPIT1拷贝CDPPIT2拷贝CDPPIT3拷贝CDPPIT4拷贝总结最低的总体拥有成本采用服务器、存储整合来减少CAPEX（资金开销）采用WAN带宽节约来减少OPEX（运营开销）针对灾难恢复和本地数据保护的任意时间点恢复应用程序一致性复制和恢复最高的可用性和性能针对EMC和非EMC存储、多供应商主机和SAN的异构支持RecoverPoint提供...RecoverPoint数据保护、复制和灾难恢复—针对物理和虚拟环境