PostgreSQL BI Performance & News from XL project
summary
Benchmarks show vanilla Postgres improves modestly across 9.1–9.5 (~20%), then gains 3× speed on 10GB with 9.6 parallelism (22 TPC‑H queries drop to ~2–3 minutes). At 100GB it still runs for hours. A distributed fork, Postgres‑XL, runs the same 3TB workload across eight nodes and cuts load/query time dramatically — about 5–6× faster on big queries and a 3TB load drops from ~79 hours on one server to ~12 hours on the cluster. Postgres‑XL tracks core releases (alpha2 on PG10) and supports partitioning, logical replication, and common tools.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 6:25PostgreSQL 96
- 8:35PostgreSQL 100GB
- 10:15How many terabytes
- 13:45Postgres Excel
- 15:42PostgreSQL Excel
- 18:23PostgreSQL 10 Marriage
- 19:53PostgreSQL Admin
- 21:07Excel
- 22:20ESA
- 23:22Core Infrastructure Initiative
- 25:13Performance Data on Excel
- 30:50Conclusion
related talks
59mUsing Postgres XL
50mPerformance Archaeology
29mPerformance Archaeology - 20 years of improvements
50mSupersized Postgres: Postgres-XL for OLTP and Big Data Analytics
34mPostgres 10, Performance, and You
54mScaling out by distributing and replicating data in Postgres XC
56mPerformance Improvements in PostgreSQL9 2
43mPostgreSQL in the cloud: Theory and Practice
