An architect is describing the watsonx.data architecture to a new team. Which statement correctly characterises the relationship between the storage layer and the query engines?
Select an answer to reveal the explanation.
Short Explanation
Separation of storage and compute is the whole architectural idea. The data sits once on object storage, the catalog says what it means, and engines are compute you can add, size, pause or delete without touching a byte of it. Get that straight and most of the fundamentals domain falls into place.
Full Explanation
The correct answer is the statement that engines are compute that reads a shared storage layer through catalog metadata, so storage and compute scale independently. watsonx.data separates storage from compute. Data persists on object storage in open table formats, catalogs describe it, and engines are compute resources that read through that metadata. Because the engines hold no authoritative copy, they can be added, resized, paused or removed independently of the data they serve.
The statement that each engine owns its own storage, and the platform synchronises data between engine storage areas, is incorrect because it describes a shared-nothing warehouse, which is the architecture the lakehouse replaces. There is no engine-owned storage and therefore no synchronisation between engines - that absence is the point.
The statement that storage is a cache the engines populate on first query and discard when they stop, is incorrect because object storage is the durable system of record, not a cache. Engines may cache for performance, but the data does not live in the engine and is not discarded when it stops.
The statement that engines and storage scale together, since an engine's capacity is derived from its bucket size, is incorrect because it couples the two dimensions that the design deliberately decouples. Bucket size affects how much data you can hold; it has no bearing on how much compute an engine has.
For the exam, be ready to state the layers in order: object storage, open table format, catalog metadata, engines, governance, and the management interfaces above them.