Industrial IoT (IIoT)
The Industrial IoT Stack Explained: A Plain-English Buyer's Guide for Tennessee Plant Managers
Published by IOT Tennessee Research & Editorial Team

- Think in Layers, Not Products
- Layer 1: Getting Data Off the Machine
- Layer 2: Moving It Reliably
- Layer 3: Deciding Where the Thinking Happens
- Layer 4: Turning Data into Work
- Layer 5: Protecting the Whole Thing
- The Tennessee Reality: Mostly Brownfield
- Build, Buy, or Bring in an Integrator
- A Sensible First 120 Days
- Challenges You Should Expect
- Where This Is Heading
Ask three people at the same Tennessee plant what "Industrial IoT" means and you will usually get three answers. The maintenance lead thinks of vibration sensors. The controls engineer thinks of PLC data and OPC UA. The plant manager thinks of a dashboard someone promised would pay for itself. All three are describing a piece of the same system, and most stalled projects trace back to nobody having drawn the whole picture before buying the first device. This guide lays the system out in plain terms so you can ask better questions of vendors, integrators, and your own team.
Think in Layers, Not Products
An industrial IoT deployment is easier to plan when you treat it as five layers that each have their own owner, cost, and failure mode. Skipping one layer is the most common reason a pilot works in a demo and fails on the floor.
Layer 1: Getting Data Off the Machine
There are two sources. The first is data your equipment already produces: most machines built in the last two decades run a PLC that tracks cycle counts, temperatures, pressures, and fault codes that nobody is collecting. The second is data you add with retrofit sensors, such as clamp-on current sensors, vibration accelerometers, and temperature probes. Reading what the PLC already knows is almost always cheaper and faster than adding hardware, so start there and add sensors only where the machine is blind.
Layer 2: Moving It Reliably
Older equipment often speaks Modbus or a proprietary serial protocol, while newer systems support OPC UA, a vendor-neutral standard designed for industrial data exchange. A gateway translates between them and passes data upstream, commonly using MQTT, a lightweight messaging protocol that works well over the modest, sometimes flaky links found on a plant floor. Insisting on open standards at this layer is the single best protection against being locked into one vendor for the next fifteen years.
Layer 3: Deciding Where the Thinking Happens
Some decisions must happen at the edge, on a device or gateway near the machine, because they need a response in milliseconds or must keep working when the plant network drops. Others, such as trend analysis across several lines, belong on a plant server or in the cloud. A sound design keeps safety-relevant and time-critical logic local and sends summarized data upstream, which also holds down data transfer costs.
Layer 4: Turning Data into Work
A dashboard is not the goal. The value shows up when a condition triggers a work order in your maintenance system, a quality hold in your MES, or a message to the right supervisor. Ask every vendor to show you the path from an alert to a completed task, not a screenshot of a chart.
Layer 5: Protecting the Whole Thing
Connecting machines to a network creates a new attack surface. Operational technology networks should be separated from office IT networks, devices should have unique credentials, and firmware should be updatable. Treat security as a design requirement in layer two, not a patch after go-live.
| Layer | Question to Ask | Common Pitfall |
|---|---|---|
| 1. Data source | What does the PLC already track? | Buying sensors for data the machine already holds |
| 2. Connectivity | Does it use open standards like OPC UA and MQTT? | Proprietary gateways that lock in your data |
| 3. Edge or cloud | What must keep working if the network goes down? | Sending everything to the cloud, then paying for it |
| 4. Application | How does an alert become a work order? | Dashboards no one opens |
| 5. Security | Is OT separated from office IT? | Flat networks with shared passwords |
The Tennessee Reality: Mostly Brownfield
Very few plants in the state get to design around IoT from day one. Most are running a mix of equipment from different decades, from newer automated cells in the automotive supply chain to machinery that predates the internet. That mix is the norm, and it is workable. Our guide to retrofitting legacy textile and furniture factories shows how plants with genuinely old equipment still get usable data, and our look at the auto and chemical manufacturing corridor shows what a more heavily automated environment demands.
Build, Buy, or Bring in an Integrator
Three paths exist, and the right one depends on your in-house skills more than your budget. Building with your own controls and IT staff gives you the most control but assumes people with time to spare, which few plants have. Buying a packaged platform gets you running faster but ties you to its assumptions. Hiring a systems integrator sits in the middle and works well when the integrator commits to open standards and documents what they build so your team can maintain it. Tennessee has an active base of automation and controls firms, and you can browse them in our directory of Tennessee IoT and automation companies when you are ready to compare options.
A Sensible First 120 Days
- Inventory what you already have. List the PLCs, drives, and meters on your critical lines and note which already expose data.
- Pick one line and one business question. "Why does this line stop on second shift?" is a better starting point than "let's monitor everything."
- Design layers two and five before buying hardware. Decide on protocols and network separation first.
- Connect, then observe for a few weeks before acting. Let the data show you what normal looks like.
- Wire one alert to one work order. Prove the full path from signal to action on a single use case.
- Document it. Write down tags, network addresses, and owners so the system survives staff turnover.
Challenges You Should Expect
Integration effort is routinely underestimated, and it is usually the network and data mapping work, not the sensors, that consumes the budget. Skills are the second constraint, since IIoT sits between controls, IT, and maintenance, and few plants have someone fluent in all three. The third is scope creep: a successful first line creates pressure to expand faster than your support capacity. Setting a clear owner and a written standard before expanding keeps the second line from repeating the first line's mistakes. For more practical tactics, see our piece on reducing downtime with IoT sensor strategies, and browse the full Industrial IoT section.
Where This Is Heading
Expect open standards to keep gaining ground, more analysis to move to the edge as hardware gets cheaper, and buyers to demand security documentation the way they now demand warranty terms. For a Tennessee plant manager, the practical advice does not change: understand all five layers, start with one problem, and keep your data in formats you control.
Frequently asked questions
This article was reviewed by the IOT Tennessee Editorial Team for accuracy, clarity, and relevance. Information may be sourced from publicly available treatment resources, government agencies, and healthcare references where applicable.
Last reviewed: September 2026Related articles
Industrial IoT (IIoT)
Industrial IoT (IIoT)
The Utilization Surprise: What Machine Monitoring Reveals in...
Industrial IoT (IIoT)
Instrument First, Automate Second: A Practical Industrial Au...
Industrial IoT (IIoT)
