Glossary · Automation software engineering and architecture
Execution jitter
Also known as: Task jitter, Cycle jitter
German: Ausführungsjitter
In real-time systems, execution jitter is the variation in the timing of a task's execution from one period to the next, such as deviations in its start time, completion time or cycle length relative to the intended schedule.
- Software engineering
In one sentence
Execution jitter is the variation in a task's start, completion or cycle time from one period to the next.
Example
Measurements show that the nominal 1 ms motion task starts between 0.98 and 1.04 ms after the previous cycle; this 60 µs jitter is within the drive's tolerance.
How it applies
- Engineering: Sources of execution jitter include interrupts, higher-priority tasks, cache effects, varying code paths and operating system activity. Real-time operating systems and careful task design keep it bounded.
- Commissioning: Controllers and runtimes often provide task timing statistics. Measure jitter under realistic load, including communication and diagnostics, not only on an idle system.
- Documentation: Document the acceptable jitter for time-critical tasks and the measured values for the delivered configuration. Service documentation should explain which diagnostic values show task timing and what to do when limits are exceeded.
Execution jitter vs. communication jitter
Execution jitter concerns when a task runs on a processor. Communication jitter concerns when messages arrive over a network. Both add up in control loops and synchronized motion, so the general term Jitter is often qualified to make clear which one is meant.