Industrial IoT (IIoT)
IoT and Industrial Automation in Indiana: Where the PLC Ends and the Data Layer Begins
Published by IOT Indiana Research & Editorial Team

- Control Versus Monitoring: Two Different Jobs
- A Map That Began in Indiana: The Purdue Reference Model
- Where IT and OT Meet
- Integration Patterns That Protect the Machine
- Safety Systems Stay Separate
- Robots, Cobots, and Autonomous Vehicles
- Retrofitting Automated Equipment With IoT: An Illustrative Case
- Cybersecurity for Automated Plants
- Change Control: The Discipline That Prevents Surprises
- Funding and Local Help
- Common Mistakes
- Where IoT and Industrial Automation in Indiana Is Heading
Industrial automation has a long history in Indiana plants. Programmable controllers run presses, conveyors, ovens, and packaging lines, and they have done so reliably for decades. The arrival of IoT raises an awkward question for the people responsible for that equipment: what changes? The honest answer is that the controller should keep doing what it does best, controlling the machine, while a separate data layer collects, moves, and analyzes information around it. Confusion about where one ends and the other begins is the source of most of the risk in this space, including downtime and cybersecurity trouble. This guide explains the split, uses a framework that originated at Purdue University to describe it, and offers practical patterns for bringing IoT to automated equipment without endangering it.
Control Versus Monitoring: Two Different Jobs
Traditional automation exists to make a machine do something reliably, in real time, every cycle. IoT exists to observe and learn. They have different requirements, and mixing them casually creates problems.
| Aspect | Control Layer (PLC, drives, safety systems) | IoT and Data Layer |
|---|---|---|
| Main purpose | Run the machine correctly and safely | Collect, transmit, analyze, and present data |
| Timing | Deterministic, in milliseconds | Best effort; seconds or minutes are usually fine |
| Consequence of failure | Production stops or safety is at risk | Data is delayed or lost; production continues |
| Change frequency | Rare, carefully controlled | More frequent as analytics evolve |
| Who owns it | Controls engineers, maintenance | Operations, IT, continuous improvement, or shared |
| Guiding principle | Availability and safety first | Insight and flexibility |
The rule of thumb that keeps plants out of trouble: the data layer may read from the control layer, but it should not be able to interfere with it. If an IoT gateway crashes, the machine should not notice.
A Map That Began in Indiana: The Purdue Reference Model
Engineers describing how factory networks should be organized often use a model that originated at Purdue University, commonly called the Purdue Model or Purdue Enterprise Reference Architecture. It divides an industrial environment into levels, and it remains a useful mental map for deciding where IoT components belong and where firewalls should sit.
| Level | What Lives There | Examples |
|---|---|---|
| 0 | Physical process | Sensors, motors, valves, actuators |
| 1 | Basic control | PLCs, drives, controllers |
| 2 | Area supervision | HMIs, supervisory control and data acquisition systems |
| 3 | Site operations | Manufacturing execution systems, historians |
| 3.5 | Demilitarized zone (added in many modern designs) | Buffer systems between plant and business networks |
| 4 and 5 | Business and enterprise | ERP, corporate IT, cloud services |
Modern IoT strains this tidy picture, because data needs to flow from level 0 or 1 up to cloud services at level 4 or 5 and beyond, skipping intermediate steps. The model still helps: it reminds you that every path between levels is a place to apply controls, and that traffic should move through deliberate, monitored connections rather than ad hoc links. Many plants place IoT gateways and data brokers in a buffer zone, so that plant equipment never connects directly to the internet and outside systems never connect directly to controllers.
Where IT and OT Meet
The term IT and OT convergence describes information technology, which handles business systems, joining operational technology, which runs the plant. In practice it means two cultures with different priorities must collaborate. IT worries about confidentiality and frequent patching. OT worries about availability and avoids changing anything that works. Neither is wrong. Problems arise when an IT team applies an office security tool to a controller network without understanding the consequences, or when an OT team connects a device to the network without IT knowing. Successful plants assign clear ownership, hold regular joint reviews, and agree on change procedures that respect both concerns. For small manufacturers with one person wearing both hats, the same principle applies: document who decides what.
Integration Patterns That Protect the Machine
- Read-only access to existing controllers. A gateway reads values from a PLC through a supported interface such as OPC UA or an industrial protocol and forwards them. It has no permission to write. This is the safest starting pattern, and it suits monitoring and analytics.
- Parallel sensing. Instead of touching the controller at all, add independent sensors, such as current clamps or vibration sensors, with their own gateway. The automation system is untouched, which is attractive for older or fragile equipment and for machines under warranty or customer control.
- Historian or data platform in between. Machine data flows to a local historian, and analytics read from that, not from the controllers. This decouples analysis from control and keeps queries from loading the PLC.
- Controlled write-back. Only when there is a clear, tested benefit should the data layer send anything back to a controller, such as a recommended setpoint. Use explicit approval steps, limits, and logging, and never let a cloud service directly command a running machine.
Most plants should stay with the first three patterns for a long time. Write-back is where the real risk lies, and the gains from monitoring alone are substantial.
Safety Systems Stay Separate
Machine safety functions such as emergency stops, light curtains, and safety interlocks are designed and certified to a different standard than ordinary controls. They should not depend on IoT infrastructure, and connecting to them for data should be done carefully, if at all, by qualified people. It is fine to record that an emergency stop was pressed. It is not fine to let a network problem, a software update, or an analytics tool affect whether it works. If a proposed IoT integration touches safety circuits, involve your safety engineer before proceeding.
Robots, Cobots, and Autonomous Vehicles
Automation is expanding beyond fixed machines. Collaborative robots work near people on tasks such as machine tending, and autonomous mobile robots move materials without fixed tracks. Both generate useful data and rely on connectivity, and Indiana's Manufacturing Readiness Grants have listed cobots and advanced robotics among eligible technologies when funded. The IoT questions are familiar: what data does the robot expose, how does it connect to other systems, who is responsible for updating its software, and how is it protected? A cobot that reports cycle counts, faults, and utilization can quickly show whether the investment is being used well.
Retrofitting Automated Equipment With IoT: An Illustrative Case
Imagine a packaging line with a decade-old PLC running a filler, capper, and labeler. Downtime is frequent and no one knows why, because the operator's log records only that the line was down. The plant chooses the least intrusive path. A gateway is connected to the PLC network with read-only access, pulling fault codes and machine states. A few current sensors go on motors that the PLC does not monitor. The data goes to a local server in a segmented network zone, and a simple dashboard shows stop causes by shift. After two months the team learns that a third of downtime stems from a single label sensor that trips repeatedly, and another portion from waiting on cartons. Neither was visible before. The fixes are inexpensive, and the PLC program was never modified. This is the typical shape of a good first IoT project on automated equipment: observe first, change control logic later and only if warranted.
Cybersecurity for Automated Plants
- Segment the network. Keep controller networks separate from office and guest networks, and place IoT gateways in a controlled zone between them.
- Control remote access. Vendors often need remote support. Use a logged, approved method, and disable access when it is not needed rather than leaving a permanent open door.
- Inventory everything. You cannot protect devices you do not know about. Keep a list of controllers, gateways, and network equipment, with firmware versions.
- Change default credentials and use unique accounts where devices allow.
- Back up controller programs and store them securely, so a failure or incident does not mean rebuilding logic from memory.
- Plan updates carefully. Patching a controller can stop a line. Schedule it, test it where possible, and have a rollback plan.
- Train people. Many incidents begin with a USB stick or a laptop moved between networks.
Larger customers increasingly ask suppliers about these practices, so documenting them helps win and keep business as well as reducing risk.
Change Control: The Discipline That Prevents Surprises
Before connecting anything to an automated line, agree on a simple change process. Write down what will be connected, how, and to what. Identify what could go wrong and how it would be detected. Schedule the work during planned downtime where possible, and define how to reverse it. Afterward, record what was done and update drawings and documentation. It sounds like bureaucracy, but it is far cheaper than debugging a mystery outage that started the day someone plugged in a new device.
Funding and Local Help
Indiana's Manufacturing Readiness Grants have supported advanced communications, industrial IoT, cobots, sensors, and advanced robotics when funded, though availability has depended on state budget cycles, so check the Conexus Indiana page for current status. Purdue's Manufacturing Extension Partnership provides consulting and training for manufacturers, including digital transformation topics, and can help plants plan integrations before they buy. Controls integrators and machine builders can also be valuable partners, especially for older equipment where original documentation is scarce.
Common Mistakes
Letting the data layer touch control. Start read-only and stay there until there is a compelling, tested reason.
Ignoring who owns what. Unclear ownership between IT, OT, and vendors causes both outages and security gaps.
Connecting first, documenting later. Undocumented connections become tomorrow's mystery.
Forgetting backups. Controller programs should be saved and tested.
Treating safety like data. Keep safety systems independent and involve qualified engineers.
Where IoT and Industrial Automation in Indiana Is Heading
Controllers are increasingly shipping with built-in support for modern protocols and secure communication, which simplifies safe data access. Edge devices are handling more analysis close to the machine, and standards such as OPC UA are making connections more uniform. At the same time, customers and insurers are paying closer attention to how operational networks are protected. The plants that do best will keep the control layer stable and dependable, build a well-managed data layer around it, and treat the boundary between the two as something to be designed, documented, and guarded. Indiana's long experience with automation is an advantage here, because the discipline that keeps a line running is exactly the discipline that keeps an IoT project from breaking it.
Frequently asked questions
Related articles
Industrial IoT (IIoT)
Industrial IoT (IIoT)
Predictive Maintenance Becomes More Important Across Indiana Factories
Industrial IoT (IIoT)
Cummins and the Connected Factory: What Industrial IoT Means for Indiana Manufacturing
Industrial IoT (IIoT)
