PostgreSQL 16 and beyond
summary
PostgreSQL 16 adds concrete replication, storage, SQL and ops changes: logical replication from standby and binary initial table copy, parallel apply of large transactions that yields 25–40% improvement for bulk loads and lets subscribers apply chunks immediately, and subscribers can use non‑PK indexes to avoid full table scans. Storage gets up to 3x faster relation extension for COPY, experimental direct I/O, PG_STAT_IO for detailed IO counts, JSON SQL functions (json_array/object), parallel full outer joins, lz4 and zstd for pg_dump, and libpq host=random load balancing.
AI-written summary · about
chapters (10)
- 0:00Introduction to PostgreSQL evolution
- 5:28PostgreSQL 16 key features overview
- 10:55Logical replication improvements
- 18:25Storage and heap enhancements
- 23:31SQL and parallelism upgrades
- 27:05Security and privileges changes
- 30:02Miscellaneous performance boosts
- 33:23Compatibility and removals
- 35:42Future PostgreSQL developments
- 41:25Q&A session
related talks
44mThe journey towards active-active replication in PostgreSQL
43mPostgreSQL 17 and beyond
31mPast, Present, and Future: Logical Decoding and Replication in PostgreSQL
51mExploring new enhancements and improvements Fujitsu has delivered to
14mBuilding a better PostgreSQL together
28mPostgreSQL 9.4 and Beyond
51mHow Postgres is Different (read BETTER than) your database
15mDave Stokes Talks Open - Episode 14