# OrioleDB — Cloud-Native Storage Engine for PostgreSQL

> OrioleDB is an open-source PostgreSQL extension that replaces the default storage engine with a modern cloud-native design, eliminating table bloat, reducing write amplification, and enabling S3-native storage.

## Install

Save in your project root:

# OrioleDB — Cloud-Native Storage Engine for PostgreSQL

## Quick Use
```bash
# Run OrioleDB-enabled PostgreSQL via Docker
docker run -d -p 5432:5432 -e POSTGRES_PASSWORD=secret orioledb/orioledb:latest
# Create a table using the OrioleDB storage engine
psql -h localhost -U postgres -c "
  CREATE EXTENSION orioledb;
  CREATE TABLE events (id int8 PRIMARY KEY, data jsonb) USING orioledb;
"
```

## Introduction
OrioleDB is a table access method extension for PostgreSQL that replaces the heap storage engine with a modern design based on undo logging and copy-on-write checkpointing. It solves long-standing Postgres pain points like table bloat, VACUUM overhead, and write amplification while remaining fully SQL-compatible.

## What OrioleDB Does
- Replaces the default Postgres heap storage with a no-VACUUM table access method
- Eliminates table bloat by using undo logs instead of multi-version heap tuples
- Reduces write amplification through copy-on-write checkpointing (no full-page writes in WAL)
- Provides row-level WAL for lower I/O overhead on write-heavy workloads
- Supports S3-compatible object storage as a native storage backend for cloud deployments

## Architecture Overview
OrioleDB implements a Postgres table access method that stores data in a dual-pointer B-tree structure instead of the traditional heap. Dead tuples are tracked via an undo log and reclaimed automatically without VACUUM. Checkpointing uses a copy-on-write mechanism that avoids the full-page write penalty of the standard Postgres WAL. The storage layer abstracts block access so that pages can be stored on local disk or in S3-compatible object storage.

## Self-Hosting & Configuration
- Run the official Docker image, which packages PostgreSQL with OrioleDB pre-built
- Build from source against a patched PostgreSQL (patches are included in the repository)
- Enable per-table by specifying USING orioledb in CREATE TABLE statements
- Mix OrioleDB and standard heap tables in the same database
- Configure S3 storage backend via postgresql.conf parameters for cloud-native deployments

## Key Features
- No VACUUM required; dead tuples are cleaned up automatically via undo logging
- Reduced write amplification saves SSD lifespan and lowers cloud storage costs
- S3-native storage enables disaggregated compute and storage architecture
- Drop-in compatibility with standard PostgreSQL SQL, extensions, and tooling
- Per-table opt-in allows gradual migration of high-churn tables

## Comparison with Similar Tools
- **Standard PostgreSQL heap** — The default storage engine that OrioleDB replaces; requires VACUUM and suffers from bloat on write-heavy workloads
- **InnoDB (MySQL)** — MySQL engine with undo logging (similar concept), but tied to the MySQL ecosystem
- **Neon** — Serverless Postgres with disaggregated storage, but modifies the entire Postgres stack rather than offering a per-table access method
- **AlloyDB (Google)** — Managed Postgres-compatible service with a custom storage layer, but proprietary and not self-hostable
- **CockroachDB** — Distributed Postgres-compatible database; different scope focused on horizontal scaling rather than single-node storage efficiency

## FAQ
**Q: Can I use OrioleDB tables alongside standard Postgres tables?**
A: Yes. OrioleDB is a table access method, so you can choose it per table. Standard heap tables continue to work normally in the same database.

**Q: Does OrioleDB eliminate the need for VACUUM entirely?**
A: For OrioleDB tables, yes. The undo log mechanism reclaims dead tuple space without a separate VACUUM process.

**Q: Is OrioleDB production-ready?**
A: OrioleDB is under active development. It is used in production by some organizations, but the maintainers recommend testing thoroughly for your specific workload before deploying broadly.

**Q: What Postgres versions does OrioleDB support?**
A: OrioleDB tracks recent PostgreSQL major versions. Check the repository for the current supported version matrix.

## Sources
- https://github.com/orioledb/orioledb
- https://orioledb.com

---
Source: https://tokrepo.com/en/workflows/24d16e17-3c72-11f1-9bc6-00163e2b0d79
Author: AI Open Source