Business

Databricks Unity Catalog Keeps Data in Customer Storage

Databricks has highlighted how its Unity Catalog managed tables allow organizations to retain full ownership of their cloud storage while automating data layout and governance.

Databricks AI2 days agoBusiness
Image: Databricks AI

Databricks is emphasizing its open storage model for Unity Catalog managed tables, which allows enterprises to keep their data within their own cloud accounts. Unlike competing data platforms that lock information into proprietary formats or provider-controlled environments, Databricks writes managed table data directly to customer-owned Amazon S3, Azure Data Lake Storage (ADLS), or Google Cloud Storage (GCS) buckets. This approach ensures that organizations maintain complete visibility, auditability, and physical control over their underlying files.

Even though the data resides in customer-owned storage, Databricks automates the layout, tuning, and cleanup of these managed tables. The system supports open formats like Delta and Iceberg, ensuring that users are not locked into a single vendor. External engines such as Apache Spark, Flink, Trino, Kafka Connect, and Snowflake can read and write to these managed tables using the Iceberg REST Catalog and Unity Catalog open APIs. This architecture allows external tools to interact with governed data without requiring expensive data duplication.

Administrators can define where managed table data lands at the metastore, catalog, or schema level, with more specific settings overriding broader defaults. This flexibility is crucial for organizations that must isolate physical storage to comply with regional data residency laws or allocate cloud costs to specific business units. When organizational needs shift, administrators can use commands like ALTER CATALOG or ALTER SCHEMA with the SET MANAGED LOCATION clause to redirect new tables to a different path without disrupting existing data.

Additionally, the platform simplifies the migration of legacy data. When converting an external table to a managed table using the ALTER TABLE command, Databricks automatically copies the data and transaction logs into the designated managed storage location. This ensures that newly managed assets immediately align with the organization's current storage policies and governance frameworks.

This is our own summary of reporting by Databricks AI

More in Business