SCADA in one sentence

SCADA — supervisory control and data acquisition — is the system that lets operators monitor and control assets spread across a wide area: pumping stations, electrical substations, pipelines, water networks, wind farms. The defining feature is *geographic distribution* connected over long-distance communication.

The five layers

A classic SCADA architecture has five layers:

  1. Field devices: sensors, actuators, and instruments that measure and act on the physical process.
  2. RTUs and PLCs: remote terminal units and PLCs acquire data locally and execute control logic at the site. They are the bridge between the physical field and the network.
  3. Communication infrastructure: serial, radio, cellular, satellite, or IP links connecting remote sites to the control centre. This is often the most failure-prone part of the system.
  4. SCADA servers (MTU): the master terminal unit software collects data from all remote sites, maintains the live database, executes supervisory logic, and serves the operator workstations.
  5. Operator workstations and historian: HMIs present mimics and trends to operators; the historian stores time-series data for analysis and reporting.

How data flows

Each RTU polls its local instruments and holds the latest values. The SCADA server polls the RTUs (or receives unsolicited reports) over the communication network. Data arrives at the server, is time-stamped and quality-checked, and is then: displayed on operator screens, written to the historian, and evaluated against alarms.

Control is usually supervisory: the operator or the server sends setpoints and commands down to the RTU, and the RTU handles the fast local loop. This split keeps the system resilient — if the communication link drops, the RTU keeps controlling its local process.

SCADA vs DCS vs IoT

  • DCS supervises a continuous process within one plant over a high-speed plant network.
  • SCADA supervises distributed assets over long-distance, often lower-bandwidth links.
  • IoT/IIoT platforms collect high-volume data for analytics, with less emphasis on deterministic control.

The lines blur — modern SCADA runs on the same servers and protocols as IIoT platforms — but the design requirements differ: SCADA must prioritize availability, deterministic polling, and operator workflows; IIoT prioritizes scale and analytics.

Key point: SCADA's defining constraint is the communication network between remote sites and the control centre. Architecture decisions — polling rates, protocol choice, redundancy — follow from that constraint.