Presented by:

4598c84307bcd518acdeee63b93e501c

Shivji Kumar Jha

from Nutanix
No video of the event yet, sorry!

PostgreSQL's popularity has surged in recent years, topping DB-Engines rankings and earning the most popular title on Stack Overflow. However, built primarily for OLTP workloads, Postgres can face challenges when handling complex analytics, queuing, and streaming. Traditionally, data is offloaded via ETL jobs to specialised OLAP systems like ClickHouse, DuckDB, and Snowflake, which excel in analytics thanks to columnar storage, vectorized computation, and advanced compression techniques.

In this talk, we’ll examine why these purpose-built databases succeed with analytics workloads and how their architectural innovations address specific demands. Further, with the rise of hybrid databases like SingleStore, TiDB, and AlloyDB, which handle both OLTP and OLAP, users face the challenge of adopting yet another database solution. This raises the question: wouldn’t it be ideal if Postgres itself could evolve to support OLAP workloads when needed?

Postgres' powerful extension framework is paving the way for such a future. We'll explore innovative extensions like Hydra, Paradedb, and pg-duckdb that are transforming Postgres into a unified OLTP-OLAP solution. Attendees will gain insights into how these extensions are driving Postgres forward, enabling it to handle analytics workloads without moving away from the familiar Postgres environment.

Date:
Duration:
45 min
Room:
Conference:
PGConf India, 2025
Language:
Track:
Database Engine Developers
Difficulty:
Medium