- It uses a schema-on-write approach.
- It stores data in its raw format.
- It is optimized for write-heavy operations.
- It is more agile than a data lake.