Event-driven reporting is efficient but not foolproof - buffers overflow and devices restart. The integrity poll is the periodic full read that guarantees the master's picture is eventually correct.
Integrity Poll in one line: An integrity poll is a request from a DNP3 master that reads all current static values from an outstation at once, not just the recent events. It re-synchronizes the master's copy of the data with the outstation's actual state. Integrity polls run on a slow periodic timer and after events like a restart or a reported buffer overflow, providing a reliable full snapshot that backstops the efficient but potentially lossy event-based unsolicited reporting.
Modern SCADA links lean on event-based reporting - the outstation sends only changes, not every value on every poll. This is efficient, but it has a gap. Event buffers are finite. If a link is down long enough, the buffer fills, and the oldest changes are discarded. When the link recovers, the master has missed those changes and its data is quietly wrong.
The integrity poll closes this gap. By periodically reading all current values regardless of events, the master gets an authoritative snapshot that overwrites any drift. Even if events were lost, the next integrity poll restores an accurate picture. This is the mechanism that turns event reporting from best-effort into eventually-consistent-and-reliable.
In DNP3 an integrity poll is a read of Class 0 (all static, current-value data) usually combined with Classes 1, 2, and 3 so any pending events come along too. Masters run it on a configurable interval - often every several minutes to tens of minutes, far slower than event reporting because it is a heavier request that returns everything.
Integrity polls also fire on specific triggers. After an outstation restart, the master issues one to rebuild its full view from scratch. When an outstation sets a buffer-overflow flag, the master responds with an integrity poll to recover the values that may have been lost. The equivalent operation in IEC 60870-5-104 is called a general interrogation, and it serves exactly the same purpose.
There is a tuning trade-off. Frequent integrity polls give a fresher guaranteed snapshot but consume more bandwidth, which is precisely the cost event reporting was meant to avoid. The interval is chosen to be long enough not to burden the link, but short enough that any undetected data drift is corrected within an acceptable window.
Field RTUs on cellular, radio, and satellite links go offline routinely, so the buffer-overflow scenario the integrity poll guards against is common, not theoretical. Operators configure unsolicited event reporting for fast, efficient change notification and a slow integrity poll underneath it as the guarantee that the central view will always reconverge on reality after any outage.
The two together give both speed and correctness. An alarm or pressure change is reported immediately by an unsolicited response; the periodic integrity poll ensures that after a long comms gap or an RTU reboot, nothing has been silently lost. Without integrity polls, a busy or intermittent site could accumulate small, invisible errors in its historical and live data over time.
Setting the interval is a practical decision balanced against link cost. On a metered cellular connection carrying many sites, engineers stretch the integrity-poll interval to conserve data while keeping it frequent enough that any drift is caught well within reporting and operational tolerances.
Normal event-based operation transmits only changes. An integrity poll reads all current values at once to give a complete, authoritative snapshot. It runs much less often than event reporting because it is a heavier request, and its job is to guarantee correctness, not efficiency.
On a slow periodic timer, and on triggers such as an outstation restart or a reported event-buffer overflow. Any of these situations means the master may have missed data, so a full read re-synchronizes its view with the device's actual state.
Yes, conceptually. The integrity poll is the DNP3 term; IEC 60870-5-101/104 calls the equivalent full-read operation a general interrogation. Both recover a complete current snapshot to backstop event-based reporting.
This page references the protocol specifications published by the organizations below. Editions, product capabilities, and documentation change over time - confirm current requirements and specifications directly with the source.
Last reviewed: July 27, 2026. Merobix is not affiliated with, endorsed by, or sponsored by these organizations; their names are used only to identify the standards and products discussed.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.