The internals of GPORCA optimizer
Shin is currently a senior principal software engineer at Pivotal Data team. Shin is a committer of GPORCA (a standalone query optimizer), GPDB (an MPP database solution), Gemfire (an in-memory data grid), and Greenplum connector (a protocol to transfer data between Gemfire and Greenplum clusters). He is also a contributor on HAWQ (a SQL solution on top of Hadoop using GPORCA). In past decade, Shin was also a seasoned former Microsoft SQL Server engineer focused on product quality of query optimizer (QO) and released column store indexing, star join optimization, and other performance features in SQL Server releases from 2005 to 2016 CTP.
No video of the event yet, sorry!
The latest OSS release of Greenplum Database (GPDB) 5.0 can switch between the planner optimizer and a standalone MPP (massively parallel processing) query optimizer called GPORCA. This query optimizer is also used in Apache HAWQ (a SQL solution on top of Hadoop). The talk will describe the architecture of GPORCA, how to develop new rules, and best practices for testing and debugging. At the end of the talk, you should be able to explore the optimizer on your own and contribute back to the community.
- 50 min
- PGConf Local: Seattle [PgConf.US]