Apache Iceberg

Apache Iceberg is an open table format that turns Parquet files in object storage into database-grade tables — ACID commits, schema evolution, hidden partitioning, and time travel — governed by the Apache Software Foundation rather than any single vendor.

Its distinguishing achievement is neutrality: Snowflake, Databricks, BigQuery, AWS, Trino, and Flink all read and write Iceberg, and its REST catalog specification is what engines standardise against. The v3 spec (GA across major platforms in 2026) added deletion vectors, a variant type for semi-structured data, and geospatial types. For an open, multi-engine lakehouse, Iceberg is the default choice; the livelier question is which catalog governs it.

Go deeper: Iceberg vs Delta Lake: How to Actually Choose in 2026