IIoT Data Budget
Estimate bandwidth and storage from tag count, sampling interval and retention, and see what report-by-exception saves.
| Measure | If every sample is sent | With report-by-exception |
|---|
How is it calculated?
- Value rate = number of tags ÷ sampling interval (values/s).
- Message size = (tags per message × payload per value) + message overhead.
- Message rate = value rate ÷ tags per message; bandwidth = message rate × message size × 8.
- Daily volume = bandwidth × 86,400 s; storage = daily volume × retention ÷ compression ratio.
The “report-by-exception” column assumes that only the stated percentage of values is not sent (those suppressed by deadband or on-change filtering). The real saving depends on the process, the filter threshold and how the signal behaves; there is no fixed figure, so the default is only a starting point to experiment with. A larger number of tags per message spreads the overhead over more values, but the message size must stay within the limits of the broker and the network.
Frequently asked questions
How close are these figures to real network traffic?
It is an estimate. TCP and TLS headers, retransmissions, MQTT QoS acknowledgements and the payload format all change real traffic. Treat it as a planning figure and compare it with measurements during the pilot.
What difference does report-by-exception make?
Sending only meaningful changes can cut traffic and storage clearly for slowly changing signals (temperature, level); the saving drops for fast-changing signals. Set the filter threshold (deadband) with process knowledge.
Are the values I enter stored?
No, the calculation runs entirely in your browser.
Related: Store-and-Forward · Deadband · Edge Computing · Edge or Cloud? Industrial Data Architecture Decisions
Want this running automatically in a real project?
Let’s design an architecture with HighByte Intelligence Hub and GoodData so the data flows from machine to dashboard automatically.