No video of the event yet, sorry!

Sharding is a challenge for PostgreSQL developers. It is general purpose tool to achieve query speedup and database scalability. Wherein, PostgreSQL architecture is not ideal for managing relational data in shared-nothing computing architecture. Nowadays, hardly anyone can say with confidence what a practically valuable implementation of sharding in PostgreSQL will looks like. The Development of sharding is moving forward in small steps across hot discussions. In this presentation i want to describe out current look on sharding in postgres. Also I want to tell about our new features, that you can see in our product - PostgresPro Enterprise and deeply study in the hackers mailing list. This year we developed support of shippable tables. It can be used as a base to the global tables implementation. Second feature is global snapshot. It is implemented to solve a problem of snapshot isolation in repeatable read and read committed levels of distributed transaction. Probably it can be used for solving the problem of making the consistent backup of distributed database. Third new feature is fast COPY FROM into a sharded table. And finally i want to describe a statistics dump/restore feature for speedup of analyze command on large sharded tables.

Date:

2020 November 20 16:10 CST

Duration:

40 min

Room:

Virtual - English Sub-Conference A

Conference:

CHINA 2020 And PGConf.Asia 2020

Language:

Track:

Performance

Difficulty:

Medium