Lakehouse

A lakehouse is a newer design that tries to combine the best of both worlds: the cheap, flexible storage of a data lake plus the structure, reliability, and fast SQL of a data warehouse. It puts a metadata-and-transaction layer (using open table formats like Delta Lake, Apache Iceberg, or Apache Hudi) on top of plain files so that messy lake storage starts behaving like clean warehouse tables.

The appeal is having one place for everything — raw files and analytics-ready tables — instead of maintaining a lake and a separate warehouse with data copied between them. In practice the lakehouse is still maturing, and teams choose it to cut duplication and let both data scientists (who like raw files) and analysts (who like tidy tables) work from the same store.