PostgreSQL 16 and beyond: Amit Kapila - PGCon 2023

"Discover the latest features and improvements in PostgreSQL 16 and beyond, including I/O enhancements, JSON advancements, and performance boosts, as Amit Kapila shares his insights from PGCon 2023."

Key takeaways

PostgreSQL 16 plans to improve I/O capabilities, including a read/add mechanism to make it production-ready.
JSON arrays can have different value expressions and JSON objects.
A new isJSON predicate helps distinguish between JSON objects and arrays.
The isJSON function returns true for objects and false for arrays.
Parallel full outer joins and right joins have been introduced, improving parallelism.
A new self-auto-tuning module is being worked on to improve processing.
Case-insensitive text collation is being improved, allowing users to define custom rules for text sorting.
Large relation files are planned to be introduced to improve storage.
Performance improvements include up to 3x improvement for bulk loads in logical replication setups.
Direct I/O capabilities are being worked on, allowing for direct access to disk without going through the buffer pool.
Index-only scans have been introduced, improving query speeds.
TOAST (type-ahead overflow storage) is being further improved.
Logical replication from standby and DDL replication have been improved.
Automatic vacuuming and cleaning up temporary files have been enhanced.
Storage area improvements include reduced memory usage and improved locking mechanisms.
Case-insensitive text collation is being improved, allowing users to define custom rules for text sorting.
Postgres now supports minimum Windows 10 for Windows installations.
Various partitioning related improvements have been made.
CPU acceleration using SIMD instructions is being supported for both x86 and ARM architectures.

PostgreSQL 16 and beyond: Amit Kapila - PGCon 2023

More talks