大數(shù)據(jù)平臺建設(shè)2014年EMC—某著名企業(yè)（南陽站）高端客戶研討會(huì)Source:EMCDigitalUniversewithResearchandAnalysisbyIDC,TheDigitalUniverseofOpportunities:RichDataandtheIncreasingValueoftheInternetofThings,April2014.20204.4

ZETTABYTES44

ZETTABYTES10x

MORE數(shù)字化宇宙20142013ZETTABYTE

=

1,000,000,000,000,000,000,000bytesFirstExabyteQuarterFirstExabyteYearQ3Q2FirstExabyteMonth200920102011201220132007200820052006FirstExabyteShippedEMC’sExabyte旅程Source:EMCDigitalUniversewithResearchandAnalysisbyIDC,TheDigitalUniverseofOpportunities:RichDataandtheIncreasingValueoftheInternetofThings,April2014.2013202028

MILLION

ITPROS

WORLDWIDE36

MILLION

ITPROS

WORLDWIDE8

MILLIONMORE

OFYOU數(shù)字化宇宙2014Source:EMCDigitalUniversewithResearchandAnalysisbyIDC,TheDigitalUniverseofOpportunities:RichDataandtheIncreasingValueoftheInternetofThings,April2014.2020230

GBPERITPRO1230

GBPERITPRO5x

MORE數(shù)字化宇宙20142013大數(shù)據(jù)無處不在Video&

GraphicsGenetic

SequencingSatellite

PhotosComputer

ModelingSeismic

ExplorationMusic/AudioHealthRecordMedical

Imaging大數(shù)據(jù)地圖定義大數(shù)據(jù)湖MachinelatencyInteractivereportsBatchprocessingHumaninteractionsMillisecondsSecondsSeconds,MinutesMinutes,Hours分布式流計(jì)算，MPPDBReal-timeanalytic/MPP/SQLonHadoopNear-real-timeanalytic不同的響應(yīng)時(shí)間的要求決定了不同的功能實(shí)現(xiàn)MapReduceofflineanalytic/Pivotal的大數(shù)據(jù)平臺服務(wù)能力企業(yè)級Hadoop商用平臺1橫向擴(kuò)展的MPP數(shù)據(jù)倉庫2基于分布式內(nèi)存的實(shí)時(shí)數(shù)據(jù)平臺3大數(shù)據(jù)平臺咨詢以及架構(gòu)規(guī)劃4大數(shù)據(jù)平臺實(shí)施運(yùn)維以及優(yōu)化5PivotalDataLake融合架構(gòu)CentralizedManagementSystemmonitoringSystemmanagementUnifiedDataManagementTierDatamgmt.servicesMDMRDMAuditandpolicymgmt.ProcessingTierWorkflowManagementDistillationTierHDFSstorageUnstructuredandstructureddataIn-memoryMPPdatabaseUnifiedSourcesFlexibleActionsReal-time

ingestionMicrobatch

ingestionBatchingestionReal-timeinsightsInteractiveinsightsBatchinsightsWorld’sLeadingExpertsPivotalLabs–PivotalDataLabsOnDemandServicesPivotalDataDispatchBATCHBATCHNEARTIMENEARTIMEHAWQGreenplumDBPivotalHDREALTIMEREALTIMEGemFireXDGemFirePivotal的大數(shù)據(jù)平臺企業(yè)數(shù)據(jù)架構(gòu)分析數(shù)據(jù)集市MPPDatabase實(shí)時(shí)數(shù)據(jù)庫In-MemoryDB實(shí)時(shí)分析系統(tǒng)In-MemoryObject企業(yè)數(shù)據(jù)倉庫RDBMS數(shù)據(jù)緩沖沉淀系統(tǒng)報(bào)表與展現(xiàn)DataVisualization數(shù)據(jù)采集系統(tǒng)Stream/CEPAnalytic

DataMartsOperational

IntelligenceRun-Time

ApplicationsEnterpriseData

WarehouseDataStaging

PlatformTraditionalBI/ReportingDataVisualizationDataIngestionSystemPivotal全套對應(yīng)方案分布式計(jì)算之MPPShared-Nothing架構(gòu)演進(jìn)共享磁盤例如：InformixXPSOracleRACDB2pureScaleDBSAN/共享磁盤DBDBDB網(wǎng)絡(luò)SAN/FC完全共享例如：SMP服務(wù)器DB磁盤無共享例如：GreenplumDBDBDBDB網(wǎng)絡(luò)磁盤磁盤磁盤磁盤MasterMPP注：藍(lán)灰色表示共享資源Mainframe,MiniComputerTerminalsLAN/InternetClient/Server

PCMobileCloudBigDataSocial

MobileDevices1STPLATFORM2NDPLATFORM3RDPLATFORMMILLIONSO