StarRocks+极速统一的湖仓分析.pptx
S匧鸟ta絡r?R涸o廩c?k?s區
StarRcoksStarRocksStarRocks
01StarRcoks
?//?→????
SingleSourceofTruth:
StarRocksLakeHouseDatawarehouseDataLakeLakeHouseDataqualityPerformanceRealtime???GovernanceOpen?Singlesourceoftruth的ScalabilityCostefficiencyStarRocks?catalog??OpenTableFormat
02StarRcoks
LakeHouseStarRocksFESpark/FlinkComputeengineCatalogTableformatFileformatStorageHMSIcebergParquetHDFSObjectstorageORCHudiStarRocksFileGlue/DLFStarRocksTableStarRocksCN统一开放的Lakehouse架构,分层解耦设计Lakehouse?Catalog?HDFS
CBOOptimizerVectorengineComputeEngineMVExternalTableStatisticsTrinoCompatibleCatalogExternalCatalogUnifiedCatalogTableFormatHive/HudiMetadataCacheIcebergMetadataCacheIcebergColocate/BucketJoinFileFormatNativeReaderIOCoalesceLateMaterializationCost-BasedLateMaterializationDataCacheFooter/PageCache
DataCacheCacheBlocklevelDataCacheCache??+Hash??Cache务LakeBEExecutDataCacheBEExecutDataCacheBEExecutDataCacheBalancedDistributionGranularCachingScanRange0CN0CN1ConsistentHashScanRange1CN3CN2ScanRangeN
DataCache?DataCache12%Cache吐?ExcutionEngineDataCacheHDFS/S3ScanBusyScan
ParquetReaderColumnCol1Col2Col3IORowGroup务IOFooter/PageCache/Page/FooterIO
ParquetReader-Col2务Col1Col2Col1SKIPScanandFilterAND
ParquetReader-Col2Col1?MaskMaskCol1?去Enumeratetheorder
Hive/Hudi?cacheHMS/Glue?Table/Partition/ListFile?refreshexternaltabletbl_name??Cacherefresh_interval,cache_ttl,cacheListFileCacheOptimizerCoordinatorRefreshThreadRefreshTableRefreshManualRefreshAutoMetadataCacheTableC\aPcahretitionRefereshAsync\ExpireRefereshAsync\ExpireHMS\GlueStorage
Iceberg?IcebergV2TablePositionDeleteFilesColocate/BucketJoinbucketpartitiontransformJoinshuffleMemory/Diskmanifest,?MemoryMemoryDiskdatalayermetadatalayercurrentmetadatapointerIcebergCatalogcurrentdmbe1t,atdabatlea1pointermetadatafiles0metadatafiles0s1malnisitfestmafniliefestmafniliefestmalnisitfestmafniliefestdatafilesdatafilesdatafiles
UnifiedCatalogHMS/GlueHive/Hudi/Iceberg/DeltaLake?Hive/Hudi/Iceberg/DeltaLakeCatalogCatalogHMS/GlueUnifiedCatalo
网址:StarRocks+极速统一的湖仓分析.pptx http://mxgxt.com/news/view/1143843
相关内容
让数据分析极速统一!StarRocks和阿里云一起干了件大事如何利用 StarRocks 加速 Iceberg 数据湖的查询效率
StarRocks和帆软软件达成战略合作,助力企业开启“极速统一”数字分析
StarRocks:从概念到应用的下一代分析型数据库
跨越速运张杰:基于 StarRocks 提升运单分析时效|爱分析活动
StarRocks Summit 2023 技术交流峰会圆满落幕
StarRocks Summit Asia 2024落幕,Lakehouse引领数据技术新趋势
StarRocks 2024 数据技术峰会圆满收官,Lakehouse引领数据技术新趋势
StarRocks数据集成
StarRocks 2.1 新版本特性介绍