A SCADA system supervises an industrial process by acquiring data, presenting state, managing alarms and trends, and supporting authorized operator actions. A useful design must show each component, trust boundary, data path, failure mode, and operating responsibility rather than reducing SCADA to a screen or server.
The short answer
SCADA combines field interfaces and controllers, industrial networks, acquisition and application servers, operator clients, engineering functions, alarm and history services, and controlled links to other systems. It supervises operations; it does not replace the basic or safety control layers. The architecture must remain understandable during normal operation, degraded modes, maintenance, recovery, and cybersecurity incidents.
- Draw physical, logical, and trust boundaries separately and identify every data producer and consumer.
- Define data quality, timestamp source, units, update rate, alarm semantics, retention, and ownership.
- Use least privilege, controlled remote access, segmentation, monitored conduits, backups, and tested recovery.
- Design degraded operation and loss-of-view behavior before selecting redundancy or virtualization options.
Decision guide
| Layer | Primary responsibility | Acceptance evidence |
|---|---|---|
| Field and control | Measure and control the process through instruments, PLCs, RTUs, and safety systems | I/O, loop, logic, interlock, and fail-state tests |
| Supervisory services | Acquire data, manage alarms, visualization, users, trends, and history | Tag, display, alarm, audit, redundancy, and recovery tests |
| Operations clients | Present authorized situational awareness and commands | Role, navigation, performance, usability, and abnormal-condition tests |
| External integration | Exchange approved information with historians, MES, reporting, or enterprise services | Interface contract, quality, security, buffering, reconciliation, and ownership |
A practical workflow
- 01
List operational use cases, users, commands, alarms, reports, history, availability, and recovery objectives.
- 02
Create physical, logical, data-flow, and trust-boundary diagrams with protocols and accountable owners.
- 03
Define the tag model, data quality, timestamps, alarm lifecycle, users, audit events, and retention rules.
- 04
Specify segmentation, identity, remote access, hardening, patching, backup, restore, and incident response.
- 05
Test normal load, communications loss, server and client failure, failover, restore, security controls, and operator acceptance.
Evidence and acceptance
Acceptance should connect each requirement to configuration identifiers, test procedures, timestamped results, defects, retests, approved deviations, recovery evidence, and the released backup set. The evidence must distinguish a passed technical test from operational authorization and cybersecurity acceptance.
Common mistakes
- Treating SCADA as a collection of graphics while leaving alarms, users, time, history, and recovery undefined.
- Flattening control, safety, supervisory, and enterprise traffic into one undocumented network.
- Assuming redundancy is proven without testing common dependencies, failover, resynchronization, and return to service.
- Publishing data externally without an interface owner, quality model, buffering behavior, or reconciliation process.
- Keeping backups without proving that the complete system can be restored to an approved state.
Engineering and security boundary
Primary sources
Technical claims and product boundaries were checked against these primary sources. Fact-checked
Frequently asked questions
Is an HMI the same as SCADA?
An HMI is an operator interface. SCADA usually adds system-wide acquisition, alarms, users, history, communications, and supervisory services, although product naming and project scope vary.
Does SCADA replace PLC or safety logic?
No. PLC, DCS, and safety systems execute the approved control and protection functions. SCADA provides supervisory visibility and authorized interaction according to the architecture.
What should a SCADA FAT prove?
It should trace requirements through tags, displays, alarms, users, communications, redundancy, failure handling, audit, backup, restore, performance, and defect closure within the agreed test environment.
Does AutoNexit collect live SCADA data?
No. AutoNexit records project documentation and evidence; it does not monitor plant signals or replace operational platforms.