Hello,

Noob alert - Am new to Lakehouse Arch and trying to understand it better. A Lakehouse such as Iceberg/Hudi - Is that (usually) going to run a a continuous process/service the way a DB engine would run?

My understanding is that:

• Storage and compute are being decoupled.
• Iceberg/Hudi would run on a cloud/clustered nodes (Inside Spark eg), provide table services as an abstraction over underlying Parquet files/other open formats. This acts as a centralized "blackbox" to provide CRUD and Transaction services to the data.
• Any other data consumers looking to use the data above (Lets say a Java application or a Spark Application), would use Iceberg/Hudi API to interact with the "blackbox" above to read/write the data?

Thank you!