Data Virtualisation
also called Query Federation, Logical Data Warehouse
Querying data where it lives across multiple systems through a single logical layer, without physically consolidating it.
The appeal is obvious: one query interface across the warehouse, the lake, an operational database and a SaaS API, with no copying, no pipeline and no duplicated storage. For discovery, for joining a small reference table to a large one, and for avoiding a copy purely to satisfy one query, it is genuinely useful.
The physics are what limit it. A federated join must move data across the network between systems with different capabilities, and the optimiser has poor statistics on the remote side, so plans are frequently bad in ways that are hard to diagnose. Performance is bounded by the slowest source, and queries against an operational database put analytical load exactly where it does least good.
Where it fits well: low-volume joins, reference data, exploratory work, and providing a governed access point that enforces permissions consistently across heterogeneous stores. Where it fails: high-volume analytics, anything needing predictable latency, and any workload that would put sustained pressure on a transactional system.
The mature position is that virtualisation complements consolidation rather than replacing it — it is the right tool for the queries not worth a pipeline, and the wrong tool for the ones that are.