Stand on the floor of a mid-sized factory and look around. There is a compressor that has run for eighteen years, a packaging line installed last spring, a row of pumps with analogue gauges, a robot cell from a different vendor, and a control room where someone still writes readings on a clipboard at the end of each shift. Every one of those machines produces information: temperature, pressure, vibration, speed, energy use, fault codes. Most of it disappears the moment it is generated.
Industrial IoT is the effort to stop that information disappearing. Industrial connectivity is how it actually gets moved. The two terms are often used together, sometimes interchangeably, but they describe different things. This explainer takes them one at a time, then looks at how they fit, where they struggle, and what professionals weighing a project tend to ask first.
What Is Industrial IoT?
The Industrial Internet of Things, usually shortened to IIoT, refers to sensors, machines, instruments and other equipment in industrial settings that are connected to networks so their data can be collected, analysed and acted on.
That definition sounds close to consumer IoT, and technically the building blocks overlap: a sensor, a radio or cable, a gateway, some software. The difference lies in the setting and the stakes.
A smart thermostat that drops offline for an hour is an annoyance. A pressure sensor on a chemical reactor that drops offline can trigger a shutdown. Industrial equipment also lives far longer than consumer electronics. A phone is replaced every few years; a turbine, a production line or a water pumping station can stay in service for 20 or 30 years. Industrial systems have to work alongside machines that were installed long before anyone thought about connecting them, in environments full of heat, dust, vibration, moisture and electrical noise.
So IIoT is less about new gadgets and more about a new layer on top of existing operations. It covers factories, but also energy grids, oil and gas sites, mines, water utilities, ports, rail, logistics hubs, agriculture and building systems. Anywhere physical assets do valuable work, there is a case for knowing more about how they are doing it.
What IIoT Looks Like In Practice
Most real deployments fall into a handful of patterns.
Condition monitoring and predictive maintenance. Vibration and temperature sensors on motors, pumps, fans and gearboxes show when a machine starts to drift from normal behaviour. Maintenance teams can then plan a repair instead of reacting to a breakdown.
Production visibility. Connecting machines to a central system shows how much each line produced, how long it was down and why. This is often the first step for manufacturers, because it replaces manual reporting with live figures.
Energy management. Metering at machine level shows where power, gas, steam and compressed air are actually used. With energy prices volatile and emissions reporting tightening, this has become one of the most common starting points.
Remote monitoring. Wellheads, pipelines, water pumping stations, wind turbines and telecom towers are spread across large areas. Sensors that report remotely save site visits and catch problems sooner.
Asset tracking. Pallets, containers, tools, trolleys and vehicles move around. Knowing where they are, and in what condition, cuts losses and improves planning.
Quality and traceability. In regulated sectors such as aerospace, food and pharmaceuticals, recording process conditions for every batch or part is not optional. Connected equipment makes that record automatic.
None of these is new as an idea. What has changed is cost. Sensors, radios, computing and storage have become cheap enough that connecting a single motor or a single pallet can make economic sense.
The Layers Of An IIoT System
It helps to picture an IIoT system as a stack of layers, each with its own job.
At the bottom are the devices: sensors, actuators, controllers, meters and trackers. Some are new and built for connectivity. Many are existing machines whose controllers already hold useful data but have never shared it.
Above them is connectivity, the cables, radios and networks that carry data from the device outward. This is the layer most people underestimate, and the one this article spends the most time on.
Next comes the edge, meaning computing close to the equipment. Industrial gateways and edge servers collect data from many devices, translate between protocols, filter out noise, and sometimes run analytics locally so decisions can be made in milliseconds rather than waiting for a round trip to the cloud.
Then the platform, which may sit in a data centre or the cloud. It stores data over time, manages devices, and provides tools for analysis.
At the top are applications: dashboards, maintenance systems, energy reports, quality records, planning tools, and increasingly AI models that look for patterns people would miss.
A weakness at any layer limits the whole system. But in practice, the connectivity layer is where many projects stall.
What Is Industrial Connectivity?
Industrial connectivity is the set of technologies that move data between industrial equipment, and between that equipment and the IT systems that use the data. It covers physical connections such as cables and radio links, the networks built on them, and the communication protocols that let different machines understand each other.
It solves three distinct problems.
The first is machine to machine inside a plant. Controllers, drives, sensors and robots need to exchange data quickly and reliably, often with strict timing. A robot arm and a conveyor that disagree by a few milliseconds can collide.
The second is plant floor to business systems. Data from operational technology, or OT, has to reach IT systems such as ERP, maintenance management and analytics platforms. These two worlds grew up separately, with different priorities, vendors and staff.
The third is remote and mobile assets. Equipment spread across fields, pipelines, roads and oceans needs wide-area links, often without mains power or wired infrastructure.
Each problem has its own family of technologies.
Wired Industrial Connectivity
Wired links still carry most of the critical traffic inside factories and process plants. Cables are predictable, hard to interfere with, and can deliver power as well as data.
Fieldbus
The oldest layer is fieldbus: serial communication standards developed from the late 1970s onwards to replace individual wires from each sensor with shared digital networks. Modbus, introduced by Modicon in 1979, is still everywhere because it is simple and open. PROFIBUS, DeviceNet and others followed. In process industries, HART allows digital information to ride on top of the traditional 4-20 mA analogue signal used by millions of field instruments.
Fieldbus systems are reliable and well understood. Their limits are speed, and the fact that each family speaks its own language.
Industrial Ethernet
From the 2000s, industrial versions of Ethernet began replacing fieldbus in many applications. PROFINET, EtherNet/IP and EtherCAT are among the best known. They use standard Ethernet cabling and hardware, but add mechanisms for the real-time performance that motion control and safety systems need.
The next step is Time-Sensitive Networking, or TSN, a set of IEEE standards that lets time-critical control traffic and ordinary data share the same Ethernet network with guaranteed timing. The appeal is a single converged network instead of separate ones for control and information.
Ethernet In Hazardous Areas
Process plants such as refineries and chemical sites have been slower to adopt Ethernet at field level, because instruments often sit in explosive atmospheres and far from control rooms. Ethernet-APL, the Advanced Physical Layer, was developed to close that gap. It runs at 10 Mbit/s over a single twisted pair, reaches up to 1,000 metres, powers field devices over the same two wires, and is designed for intrinsically safe use in hazardous locations. It is backed by organisations including the FieldComm Group, ODVA, the OPC Foundation and PROFIBUS & PROFINET International.
For process automation, that is significant. It means a flow meter deep inside a plant can, in principle, join the same IP-based network as everything else.
Wireless Industrial Connectivity
Wireless is growing fastest, for obvious reasons: it avoids the cost and disruption of running cables, and it reaches assets that move or sit far away.
According to IoT Analytics' October 2025 update, there were about 21.1 billion connected IoT devices worldwide in 2025, up 14 per cent on the year, with a forecast of 39 billion by 2030. Wi-Fi, Bluetooth and cellular together accounted for about 80 per cent of connections. Those figures cover all IoT, consumer and industrial, but they show where the wireless market's weight lies.
Short Range: Wi-Fi, Bluetooth And Industrial Wireless
Wi-Fi is widely used in factories and warehouses, especially for mobile devices, scanners and vehicles. Bluetooth Low Energy is common for battery-powered sensors and indoor location. For process plants, WirelessHART and ISA100.11a were designed specifically for industrial instruments, with mesh networking and the reliability that control engineers expect.
Long Range, Low Power: LPWAN
Low-power wide-area networks trade speed for range and battery life. A sensor sends small amounts of data a few times an hour and can run for years. LoRaWAN operates in unlicensed spectrum and can be deployed privately by a company or through a network operator. NB-IoT and LTE-M are cellular standards run by mobile operators. These networks suit metering, environmental monitoring, agriculture, asset tracking and remote infrastructure: cases where data volumes are small but distances and device counts are large.
Private LTE And 5G
For larger sites that need high capacity, mobility and more control than Wi-Fi offers, private cellular networks have moved from pilot to mainstream. A private network uses LTE or 5G technology but is dedicated to one organisation's site, such as a factory, port, mine or airport.
The GSA's June 2026 report on private mobile networks counted 2,003 organisations worldwide with private network deployments above €100,000 in value, across 88 countries, with customer references growing at a compound annual rate of 37 per cent since 2019. Manufacturing remains the leading sector by number of deployments, and 5G accounts for more than half of deployments announced since 2022.
Private 5G is often pitched with promises of ultra-low latency and massive device density. In practice, many early deployments use it for more ordinary needs: reliable coverage across large sites, connecting automated guided vehicles, video for quality inspection, and bringing mobile workers online.
Satellite
For assets beyond terrestrial networks, such as ships, remote pipelines, mines and farms, satellite IoT has become far more accessible. New low-Earth-orbit constellations and direct-to-device services are lowering costs, and a growing number of modules can switch between cellular and satellite as coverage changes.
The Language Layer: Protocols That Make Data Useful
Connecting a machine is only half the job. The data also has to be understood. A temperature reading is useless if the receiving system does not know which machine sent it, in what unit, and what it means.
This is where industrial communication protocols come in, and two have become especially important.
OPC UA, released by the OPC Foundation in 2008, is a platform-independent framework for exchanging industrial data securely. It can run anywhere from an embedded microcontroller to cloud infrastructure, and builds in encryption, authentication and auditing. Its real strength is information modelling. Rather than sending a bare number, OPC UA can describe a machine's structure, its components and the meaning of each value. Industry groups have built more than 150 companion specifications that define standard models for specific equipment types, from robots to machine tools. OPC UA is also standardised internationally as IEC 62541.
MQTT takes a different approach. It is a lightweight publish-subscribe messaging protocol, originally created in 1999 for monitoring oil pipelines over unreliable satellite links. Devices publish messages to a central broker, and any authorised application can subscribe. It is simple, efficient and widely supported by cloud platforms. The Sparkplug specification adds a standard structure to MQTT messages for industrial use.
Many architectures now use both: OPC UA to model and access data at machine and plant level, and MQTT to move it efficiently to wider systems. The idea of a "unified namespace", in which all operational data is published to a shared, structured hub, has become popular among manufacturers trying to break down silos.
Where IT Meets OT
Industrial connectivity forces two cultures together.
Operational technology teams care about uptime, safety and determinism. Their systems were often isolated by design, and "if it works, don't touch it" is a sensible rule when a mistake can stop a line or endanger people.
IT teams care about data, integration, scalability and security. They are used to regular updates, cloud services and standard networks.
IIoT sits in the middle. It needs OT's data and IT's tools. Projects that treat it purely as an IT initiative often run into resistance on the floor. Projects run purely by OT can struggle to scale beyond one site. The organisations that make progress usually create joint teams and agree early on who owns what.
The Hard Parts
For all the attention IIoT receives, many projects stall after the pilot stage. The reasons are consistent.
Brownfield reality. Most plants are full of older equipment with proprietary interfaces or none at all. Connecting them requires protocol converters, retrofit sensors or controller upgrades, and each machine can be a small project of its own.
Security. Connecting equipment that was once isolated exposes it to new threats. Ransomware attacks on manufacturers and utilities have made this a board-level issue. The IEC 62443 series of standards provides a framework for securing industrial automation systems, but applying it takes time, expertise and discipline.
Data without purpose. It is easy to collect data and hard to use it. Projects that start with "connect everything" often produce dashboards nobody checks. Those that start with a specific problem, such as unplanned downtime on one critical line, tend to show value faster.
Skills. People who understand both control systems and modern IT are scarce. Many companies rely on systems integrators to bridge the gap.
Long lifecycles. A network chosen today may need to work for 15 years or more. Standards evolve, vendors change strategy, and radio technologies are retired. Choosing open, widely supported standards reduces, but does not remove, the risk of being stranded.
Choosing Connectivity: The Questions That Matter
There is no single best industrial connectivity technology. The right choice depends on the job, and engineers tend to work through the same questions.
How much data, and how often? A vibration sensor sampling at high frequency needs very different bandwidth from a tank level reading sent once an hour.
How fast must it arrive? Motion control needs deterministic timing measured in microseconds or milliseconds. Energy reporting can tolerate minutes.
How far, and through what? Inside a steel building, across a large yard, along a pipeline, or across an ocean.
Is there power? Mains-powered devices can use almost any link. Battery-powered sensors push the choice toward low-power options.
Does it move? Fixed equipment can use cables; mobile assets need wireless, and often more than one type.
What is the environment? Hazardous areas, extreme temperatures, washdowns and heavy electrical noise all narrow the options.
Who runs the network? A company can own a private network, rely on an operator, or combine the two. Each option shifts cost, control and responsibility.
How long must it last? The expected life of the equipment should shape the choice of standard.
In most real sites, the answer is a mix: industrial Ethernet on the lines, Wi-Fi or private cellular across the facility, LPWAN or cellular for remote and mobile assets, with OPC UA and MQTT tying the data together.
Why It Matters Now
Industrial IoT has been discussed for well over a decade, and some of the early promises did not materialise on schedule. What is different now is the maturity of the underlying pieces. Industrial Ethernet is standard. Low-power cellular networks are widely available. Private 5G has thousands of real customers. Ethernet-APL is bringing digital communication into hazardous process areas. OPC UA and MQTT have given the industry common ways to describe and move data.
At the same time, the pressure to use that data has grown. Energy costs, emissions reporting, supply chain disruption, labour shortages and the arrival of practical AI tools all reward companies that know what their equipment is doing.
Industrial IoT is the goal of turning physical operations into usable information. Industrial connectivity is the unglamorous infrastructure that makes it possible. For anyone building, buying or investing in this space, understanding the second is usually the fastest way to judge whether the first will work.
