Data Warehouse
A separate database holding a read-only copy of data from all company OLTP systems, reshaped for analysis — so analysts can query freely without endangering production.
What is it?
Since the late 1980s/early 90s, analytics moved off OLTP systems. Four reasons: data is scattered across many systems (data silos), OLTP schemas are awkward for analysis, analytical queries are expensive enough to degrade production, and OLTP systems sit on locked-down networks. The warehouse answers all four: one read-only analytical copy.
What I learned
The warehouse’s real job isn’t just analytical speed — it’s being the one place where hundreds of separate operational databases can be joined. HTAP engines don’t kill that.
Learned from
Applied in
No practical applications connected yet.