SCADA Historian Storage calculator
What the same data costs once it is archived rather than transported.
SCADA / OT · Interactive Operations Model
Bit rate · Daily and monthly data
Change the polling assumptions and see what the link is really carrying — point count, poll interval, payload size, protocol overhead and the number of sites sharing it. At small payloads the framing can be most of the traffic.
Values requested each cycle. Reading a contiguous block costs far less than the same count scattered across registers.
Time between cycles. With report-by-exception, use the effective update rate rather than the scan rate.
Framing, addressing, TCP/IP and acknowledgements on top of the application data. Small frequent reads carry proportionally far more of it.
Remote sites sharing the same circuit or data plan.
Start from an example: One remote site, 500 points every five seconds →
Move a parameter and the link answers: what each poll cycle puts on the wire, how much of it is framing rather than measurement, and what that comes to in a day.
The link carries the figures you entered, in the part of the address that is never sent to a server.
Change one assumption at a time. The model updates from the same calculation used for the reported result.
Keep the number of sites and points fixed. Shorten the polling interval and watch how quickly the required link capacity changes.
We can measure what the poll actually costs on your circuit, model batching and retries, and design the acquisition so the data plan stops being the constraint.
Values per second = points per site ÷ poll interval × sites—Application bytes per second = values per second × bytes per point—Bandwidth = bytes per second × (1 + overhead) × 8 bits—Monthly data = bytes per second × 2,592,000 seconds—This is a screening estimate. Real traffic depends on the protocol, on how many points fit in one request, on retries and keep-alives, and on whether the link carries anything else. Cellular plans also bill rounded sessions rather than exact bytes. Confirm against a measurement on the live link before committing to a plan or a circuit.
A link is not sized by the number of points but by how often you ask them. Polling is a rate, and every value carries protocol framing with it — charged per value, not per cycle, so at small payloads the framing can be most of the traffic. The model separates the measurement from the overhead, and the sites from each other, so you can see which part of the link you are actually paying for.
What the same data costs once it is archived rather than transported.
Twelve questions on polling, report-by-exception, protocols and the OT/IT boundary.
Cloud SCADA and distributed data acquisition around real infrastructure.