What Is SCADA and How Does It Work?

Why supervisory control matters

Operators cannot manage what they cannot see. Pipelines, water networks, power distribution and production lines spread equipment over distances no one can walk every shift. Without a supervisory layer, someone has to visit each site to read a gauge, and a fault is found when it has already caused damage.

SCADA closes that gap. It turns thousands of field readings into one live picture, raises an alarm when a value leaves its safe range, and lets a person act from a control room instead of a truck. The engineering question is rarely whether to have SCADA. It is how the field, the network and the software should divide the work.

SCADA

Supervisory control and data acquisition: a system architecture that gathers data from distributed field devices, presents it to operators and sends supervisory commands back. Fast, deterministic control stays in the field controllers.

SCADA architecture: the four core components

SCADA is not one software tool. It is a layered system that bridges physical equipment and the applications people use.

Remote terminal units (RTUs)

An RTU is a ruggedised, microprocessor-based unit at a remote site. It reads analog and digital signals from sensors, such as pressure, temperature, flow or a valve position, converts them into data the system can use, and reports them upstream. RTUs are built for sites with poor power and intermittent links, so they typically buffer data and timestamp events locally.

Programmable logic controllers (PLCs)

A PLC provides real-time automated control for fast processes. It runs a deterministic scan cycle: read inputs, execute control logic, write outputs, repeat. Interlocks and emergency stops belong here, because they must work whether or not the central system or the network is available. In practice the line between RTU and PLC has blurred; many sites use PLCs with communication modules in both roles.

Human-machine interfaces (HMIs)

The HMI is the visual layer. It turns process data into mimic diagrams, trends, alarm lists and controls, so an operator can see the state of a line or a network and intervene. An HMI can be a panel next to one machine or a set of screens in a control room that covers many sites.

Communication infrastructure

The network carries readings up and commands down. It can be fibre, Ethernet, licensed radio, cellular or satellite, often mixed within one system. Whatever the medium, the protocols must tolerate delay and loss, and the links must be secured, because they are a direct path to physical equipment.

How a SCADA system works, step by step

Two patterns move the data. In polling, the master asks each device for its values on a schedule. In report-by-exception, a device sends data only when a value changes beyond a threshold. Polling is predictable; report-by-exception uses far less bandwidth, which matters on cellular and satellite links.

The protocols are a mix of old and new. Modbus remains common at device level. DNP3 and IEC 60870-5-104 are widely used in utilities because they carry timestamps and handle unreliable links. OPC UA connects SCADA to plant software, and MQTT is often used to move SCADA data into IoT and cloud platforms.

HMI vs SCADA: what is the difference?

The two terms are often used together, and they are not the same thing. An HMI is one component: a screen through which a person sees and operates a process. SCADA is the whole system around it, including data acquisition from many devices, communication, alarms, history and supervisory control across sites.

A PLC can run without an HMI or SCADA: it keeps executing its program. What is lost is visibility and the ability for a person to intervene remotely.

Where SCADA is used

• Energy: generation and distribution assets spread over large areas report to a central control room, where operators reroute power and respond to faults.
• Water and wastewater: pump stations, reservoirs and treatment steps such as chemical dosing run continuously and are supervised remotely.
• Manufacturing: line and machine data reach one control room, so operators spot bottlenecks, quality drift and emergencies early.
• Oil and gas: wellheads, tanks and pipelines report pressure, flow and level from remote locations, often over cellular or satellite links.
• Transport: traffic signals, signs and roadside sensors feed regional traffic management centres.

Types of SCADA systems: from monolithic to IoT

Architectures evolved in generations. Most real installations mix them, which is why modernisation is mostly an integration problem.

What SCADA changes for operations

• From reactive to planned maintenance: logged deviations in pressure, temperature or run time show an asset degrading before it fails, so repairs can be scheduled.
• Fewer manual readings: data arrives automatically, so technicians no longer drive to sites only to copy numbers, and transcription errors disappear from the record.
• Faster response to unsafe conditions: alarms fire the moment a value leaves its limits, and automated shutdown logic in the field isolates the problem.

None of this comes automatically. Alarm limits that are too tight flood operators with noise; history that no one reviews prevents nothing. The value depends on choosing the right signals, sensible limits and a clear owner for each alarm.

Building IoT monitoring software for SCADA: the VikingSCADA project

VikingSCADA provides monitoring for industrial equipment at remote oil, water and gas sites. Before the project, readings depended on manual tracking: technicians drove long distances to wellheads and processing stations to copy values from local control modules.

• Travel time and cost: field staff spent hours driving between remote sites.
• Data quality: manual notes introduced transcription errors into the historical record.
• Late detection: data was reviewed only periodically, so a malfunction could go unnoticed for days.

Indeema built a web-based IoT SCADA monitoring platform that captures and stores operational data from remote client sites. The ingestion layer was designed for SCADA traffic over low-bandwidth cellular and satellite connections. Network segmentation and strict access control keep the cloud interface separated from the field equipment.

• Automated data collection replaces routine site visits made only to take readings.
• Automated logging records process values without manual transcription.
• Real-time alarms notify field personnel as soon as an equipment value leaves its safe range.

Security: the cost of connecting SCADA

Classic SCADA relied on isolation. Every web client, cloud link and remote-access path removes some of it. Many field protocols were designed without authentication, so a network path to a controller can be a path to the physical process.

Before you modernise a SCADA system

A short review that turns a modernisation idea into requirements:

Common mistakes

1. Moving control logic into the cloud

   A link outage or latency spike turns into a process upset or a safety issue.

   Instead: Keep control and interlocks in PLCs; use the cloud for supervision, history and analytics.

2. Designing for the best-connected site

   Remote sites on cellular or satellite links drop data or flood the link.

   Instead: Size polling, payloads and buffering for the worst link in the fleet.

3. Exposing field devices for remote access

   Unauthenticated field protocols become reachable from outside the plant.

   Instead: Use segmented zones, outbound-only gateways and per-user access through a managed entry point.

4. Alarming on everything

   Operators learn to ignore alarms, including the one that matters.

   Instead: Rationalise alarms: clear limits, priorities and an owner for each.