Glossary · IIoT, data and AI
Data warehouse
Also known as: DWH
German: Data Warehouse
In information technology, a data warehouse is a central database that integrates cleaned, structured data from several operational systems in a predefined schema optimized for reporting and analysis, usually with historical data kept over long periods.
- IIoT
In one sentence
A data warehouse integrates cleaned, structured data from several systems in a fixed schema optimized for reporting and analysis.
Example
Production, quality and ERP data are loaded nightly into a data warehouse, where controllers compare manufacturing costs per product across plants.
How it applies
- Engineering: Loading a warehouse requires agreement on shared keys, such as material numbers, equipment IDs and time bases. This is where Master data problems become visible.
- Operation: A warehouse typically supports management reporting rather than shop floor decisions, since data is loaded at intervals.
- Documentation: Document the meaning of each reported measure and dimension in a business glossary, so that reports from different plants are comparable.
Data warehouse vs. data lake
A data warehouse structures data before storing it (schema on write) and suits repeatable reporting. A Data lake stores raw data and structures it when read (schema on read), which suits exploration and machine learning. Loading jobs are typically implemented as a Data pipeline, and their definitions document where each reported figure comes from.