I have written quite a lot on the concept of a Unified Namespace or UNS. In a nutshell, this is the central data system that extracts real time data from all other systems in the plant (SCADA, MES, ERP, etc.). The UNS only keep track of the present, i.e. the most recent value. The past is discarded.
If the UNS doesn’t track historical data, where do we save it?
The UNS communicates his current values to other systems for historical storage. Depending on the use case for the data, this storage could either be:
- On premises (in a data historian) for low-latency access for edge applications
- In the cloud (in a data lake) for easy integration with ML models
The final result is a beautiful architecture where all current data is always found in one place (the UNS), and all historical data can be retrieved from either the cloud or a local historian.