Glossary · Data, timing and communication
Stale data
In systems engineering, stale data is data that is structurally valid but too old to support the intended decision safely. Whether data is stale depends on its age and on the decision it is used for.
- Data and interfaces
- Functional safety
In one sentence
Stale data looks valid but is too old for the decision at hand; its maximum age must be defined per use, checked by timestamp and handled.
Example
An HMI keeps showing a tank level of 40% received ten minutes ago, because the gateway froze and forwarded the last value without marking it.
How it applies
- Integrated systems: Define a maximum age for each data item and use: a temperature trend may tolerate minutes, a position used for collision avoidance only milliseconds. Check the age with the timestamp against a consistent time base (time synchronization).
- Functional safety: Stale data passes structural checks, which is what makes it dangerous. Treat an exceeded age as a fault and apply a defined reaction, such as a safe default.
- Technical documentation: Documentation can be stale in the same way: valid in form, but describing an earlier state of the machine. See documentation latency.
Stale data vs. bad data
Bad data fails validation or is flagged by its quality status. Stale data is correct in form and may even be flagged “good,” but time has made it unreliable.