Apache Iceberg.
The open standard for the data lakehouse
Why Iceberg matters.
ACID transactions on data lakes
Full insert, update, delete, and merge support on object storage. No more append-only limitations or manual partition management. Iceberg makes your data lake behave like a proper database.
Schema evolution without rewrites
Add, rename, or reorder columns without rewriting existing data. Iceberg tracks the full schema history, so old readers and new writers coexist without breaking.
Time travel & rollback
Query any historical snapshot by timestamp or snapshot ID. Roll back a bad write in seconds. Iceberg maintains a complete audit trail of every table change: built-in, no extra tooling.
Starburst is the enterprise layer on Iceberg
We don't just support Iceberg. We're one of its largest contributors. Our Managed Icehouse product automates the operational tasks that make Iceberg production-ready at scale: compaction, snapshot expiry, orphan file cleanup, and partition evolution.
The result: you get the openness of Iceberg with the reliability your business demands.
90%+
of Starburst customers run on Apache Iceberg
- Icehouse LakeOps: automated compaction & maintenance
- Native Iceberg reads & writes via Trino
- Full ACID support on any object store
- Time travel, rollback, snapshot management
The Iceberg ecosystem
Iceberg is vendor-neutral. Your data stays yours, readable by any compatible engine.
Start with Starburst + Iceberg.
Open data lakehouse storage, automated, governed, and AI-ready.