PostgreSQL at Amazon
summary
Amazon moves from Oracle to PostgreSQL at massive scale: migrated 7,500 Oracle databases and 75 PB of data into Postgres variants, drove Redshift (initially 1.6 PB MPP) and launched RDS and Aurora with separated compute/storage. RDS ships community releases day‑one and now supports 92 extensions; instance sizes reach 196 vCPUs/4 TB RAM and AWS previews U7 with 1,920 vCPUs/32 TB. Upstream contributions target Postgres 17/18 features like B‑tree IN optimizations, built‑in collations, parallelized major upgrades and stats export/import, plus Aurora DSQL and automatic sharding.
AI-written summary · about
chapters (15)
- 0:00Amazon’s PostgreSQL vision
- 1:27AWS launches managed PostgreSQL
- 3:55Aurora’s compute-storage split
- 5:02Oracle to PostgreSQL migration
- 6:49PostgreSQL in daily ops
- 7:55Scaling PostgreSQL to exabytes
- 8:57AWS extension plans
- 10:23Aurora’s automatic sharding
- 11:45Amazon’s PostgreSQL team
- 13:09PostgreSQL 17 optimizations
- 14:37Always-available PostgreSQL focus
- 15:40Reducing PostgreSQL downtime
- 16:43Next-gen hardware for PostgreSQL
- 17:30PostgreSQL for GenAI challenges
- 18:32Multi-region PostgreSQL solutions
related talks
18mPostgreSQL at Amazon
39mAWS Aurora
10mPostgres on AWS: Run 5X faster on metal, migrate anywhere
40mPushing past PostgreSQL’s vertical scaling limits - Running Amazon RDS
53mPostgres in Amazon RDS
14mBuilding a better PostgreSQL together
47mPostgres & Beyond
41mRunning Highly Available Postgres in AWS RDS
