StarRocks 构建极速统一的湖仓新范式

1 .StarRocks 

2 .01 02 03 StarRocks & 04 StarRocks 

3 .01 

4 .StarRocks 

5 .StarRocks 

6 .StarRocks 

7 .StarRocks 

8 .

9 .02 

10 .Data Warehouse BI DataSource ETL Data Data Marts Warehouse Report Data Lake BI DataSource On Demand Data Lake Report 

11 .常见的大数据技术架构 研 模 研 务 研 Ø TP ETL ETL TP TP TP Ø MPP ETL TP MPP TP Ø ETL ETL ETL TP TP MPP 

12 . 研 TP 研 研 TP MPP 研 kappa 

13 .研 研 的 研 研 l l l l l l l l l 

14 .

15 .

16 .03 StarRocks & 

17 .StarRocks 

18 .Killer Feature - 极简的存算分离架构 

19 .Killer Feature - Catalog StarRocks catalog Ø StarRocks 务 Ø Apache Hive Apache Iceberg Apache Hudi Delta Lake Paimon JDBC 务 

20 .Killer Feature - 

21 . External table ETL Query CDC BI,Reporting,Applications,Ad Hoc Ingestion 

22 .• • • / • A/B Test • • BI • / • 2~3 

23 .Killer Feature - StarRocks Compute engine Spark/Flink … StarRocks CN Unified catalog CBO Vector engine Query cache Catalog HMS Glue/DLF … StarRocks FE Realtime update Partition/Bucketing/Sort Table format Iceberg Hudi … StarRocks Table Colocate Statistics Index File format Parquet ORC … StarRocks File Bitmap/HLL Type Storage HDFS S3 n Lakehouse n StarRocks 

24 .StarRocks DataSource Lakehouse Data App HDFS/S3 研 Data Application Data Lake Kafka 研 Table Sink 研 Data Lake Data Flink/ Hive/Iceberg/Hudi Governance Spark 

25 . StarRocks湖仓融合常见范式 StarRocks 

26 .04 StarRocks湖仓落地实践 

27 .StarRocks - 1 2 Hive Kylin 3 Trino 

28 .StarRocks - 升级StarRocks湖仓 1 7 2 MV 40 3 Trino 99% 

29 .StarRocks 1 EB 2 Presto 3 Presto