使用 QuestDB 设计高性能时间序列数据库.pdf

《使用 QuestDB 设计高性能时间序列数据库.pdf》由会员分享，可在线阅读，更多相关《使用 QuestDB 设计高性能时间序列数据库.pdf（32页珍藏版）》请在三个皮匠报告上搜索。

1、High Performance Time-Series Database Design With QuestDBWhat you will learn-Evolution of OLAP and Time Series DBMS-What data model databases are converging on-A little bit about QuestDB and how we look to the future of DBMSIntroVlad Ilyushchenko,Co-Founder and CTO at QuestDBMy responsibilities are:

2、-We build what people want-We work as a team,because building databases is hardIntroduction to Time Series DatabasesOrganisations use Time Series DBMSs(TSDB)to help acquire and extract information from nascent data.Examples of nascent data workflowsOLAP DBMSorganisations use On-Line Analytical Proce

3、ssing systems to extract information from existing data sets.OLAP and TSBS Design Journey-Monolith-Shared disk DBMS Engine-Lakehouse DBMS EngineMonolith DBMSMonolith DBMS-DBMS manages shards-DBMS manages data availability-Compute elasticity is limited-Database provides ingress/egress endpoints-Data

4、is vendor locked2010s TSBS-TSBS replaced OLTP-All TSBS are monolith-OLAP have moved on from Monolith to Shared Disk EnginesOpenTSDBShared Disk EngineShared Disk Engine-Object Store manages data(S3 etc)-Object Store manages data availability-Data store is limitless;removed the need for sharding-Elast

5、ic compute-Database provides ingress/egress endpoints-Data is vendor locked2010s OLAPBy 2010s OLAP databases have already moved on from Monolith Data Warehouse to the Shared-Disk Engine model.Lakehouse Engine2020s-OLAP Lakehouse systems-Better schema control and versioning.-Allow solve ingress issue

6、s by depositing new data to the shared file system(object store)-CRUD support-Ingress into row-first log-structured files with indexes-Merge newly added data2020s Overview-Lakehouse Engine architecture offers the most flexibility,TSDBs are not there yet.-TSDBs double-down on the ingress performance,