Glossary · Change, versions and configuration
Data model version
In systems engineering, a data model version is the version of the semantic data structure, including its fields, meanings, units, ranges and enumerations. It changes when the meaning of data changes, not only when its format does.
- Data and interfaces
- Systems engineering
In one sentence
A data model version identifies fields, meanings, units, ranges and enumerations, so peers can tell whether they interpret data the same way.
Example
A packaging line adds a new enumeration value “jam cleared” to its machine status model and increments the data model version so that the MES and the HMI can detect the change.
How it applies
- Integrated systems: Increment the data model version for changes in meaning, units, ranges or enumerations, even if the byte layout is unchanged. Otherwise peers risk semantic drift.
- Machine safety: Safety-relevant values (speeds, positions, zone states) need a defined unit and range in every data model version. A changed engineering unit is a potential hazard.
- Technical documentation: Interface specifications and data dictionaries should reference the data model version they describe.
- AI and retrieval: Retrieval and AI assistants that read machine data need the data model version to interpret historical records correctly.
Data model version vs. protocol version
The data model says what the data means. The protocol version says how it is transported. One protocol can carry several data model versions, and one data model can travel over several protocols.