- Open Table Formats - Data is written as Parquet files and managed using Iceberg or Delta metadata logs.
- Object Storage–Backed - Tables reside in S3, GCS, Azure Blob, or other object storage systems.
- Databricks Unity Catalog
- Snowflake Open Catalog
- Amazon S3 Tables
- You want deep integration with an existing Lakehouse platform.
- You require open, interoperable formats for analytics, AI/ML, or downstream workloads.
- You prefer to manage optimization (compaction, vacuum, retention) in systems like Databricks, Snowflake, or Polaris.