If both networks use Ethernet, why do so many problems start when people treat them as the same thing? The label looks familiar, but the job is not.
John Rinaldi of Real Time Automation makes that point in plain terms. In his view, IT and OT run on different assumptions, and those assumptions shape how networks are built, timed, and addressed on the factory floor.
Table of Contents
The first split is purpose, utility versus production
In most IT environments, the network is a utility. It needs to be available, stable, and shared across the business, much like power, water, or HVAC. That mindset sets the tone for everything that follows. The network is a service, and the IT team supports many departments without becoming part of any one department’s daily process.
OT starts from a different place. On a plant floor, the network is tied to the machine and the work the machine does. Rinaldi uses a tea bag line as an example. You can’t peel the controls and the control network away from the system that makes the tea bags and still expect production to keep moving. The switches, PLCs, drives, and devices are part of one operating system in the plainest sense of the phrase.
OT networks are part of production, not a background office service.
That broad split lines up with Cisco’s OT vs. IT overview, which separates data-focused IT systems from systems that control physical processes.
A quick side-by-side view makes the contrast easier to see.
| Area | IT networks | OT networks |
|---|---|---|
| Purpose | Provide shared network service across the business | Keep machines and processes running |
| Timing | Delays are often tolerable | Cyclic traffic must arrive inside tight windows |
| Addressing | Each device needs a unique address | Duplicate private addresses may repeat across similar machine cells |
Once you see that purpose gap, the rest of the differences stop looking strange. They start to look like design choices that fit the job.
How IT and OT networks operate under pressure
The second difference is how the network behaves when work is in motion.
OT traffic lives on tighter timing
Office traffic can usually absorb delay. A printer can pause. A file transfer can take a few more seconds. A database can lock for a moment while it finishes some background task. People may notice, but the process often survives the wait.
On the plant floor, delay can be a problem in itself. OT networks often carry cyclic messages that must arrive within a set time and within allowed jitter. That doesn’t always mean hard real-time behavior, but it is close enough that timing matters a lot. If a control message shows up late, the controller may miss its window. Then the process can drift, fault, or stop.
Because of that, OT engineers don’t look at the network as a best-effort pipe. They care about predictable delivery. Bandwidth matters, of course, but predictability matters more when control traffic is on the line.
VLANs and rings mean different things in OT
The same split shows up in network design. Rinaldi points out that IT and OT often use VLANs for different reasons. In many IT settings, VLANs help organize and monitor switch infrastructure. On the plant floor, VLANs are often used to segment equipment or isolate parts of a machine area for a clear operating reason.
Rings show the culture gap even more clearly. Traditional IT design works hard to avoid loops because loops can create storms and instability. In OT, ring topologies can make sense when uptime is the priority and the process needs communication to survive a fault. A ring can help the network stay available if a cable is cut or a device fails.
Palo Alto Networks makes a similar point in its difference between IT and OT, noting that OT is built around control of physical operations. What looks odd to an office network team may be a sensible choice on a production line.
Why duplicate addresses can make sense in OT
The third difference is addressing, and this is where OT often breaks one of IT’s core habits.
In IT, every device on the network is expected to have its own unique address. That rule is basic, and for good reason. Shared business networks fall apart when address conflicts appear. Office systems assume one address points to one device, full stop.
OT can work differently, especially inside repeated machine cells. Picture one PLC and its connected devices on a switch. Then picture another PLC elsewhere in the plant doing the same job with the same device layout. In that case, OT engineers may assign the same private address pattern to both machine sections. Device X in one cell can have the same address as Device X in the next cell. A drive in one unit can match the address of the corresponding drive in another.
That sounds wrong from an IT seat, yet it solves a real plant-floor problem. If engineers fix a bug, tune a program, or add a feature to one PLC, they can copy that program to the matching machine with far less rework. The address structure stays familiar because the machine structure stays familiar.
This is one reason Rinaldi argues against dropping all factory devices onto a shared IT network. Duplicate addressing inside controlled OT spaces can be useful. On a broad office-style network, it becomes a conflict.
Why this view pushes back on IT/OT convergence
Rinaldi’s strongest point is about architecture. He is not arguing that data should stay trapped inside machines. Real Time Automation’s work is built around moving data around the factory floor and getting machine data into other applications. The issue is how that exchange happens.
In his view, full IT/OT convergence goes too far when it assumes one network model can cleanly fit both office systems and plant-floor control. A network built for printers, databases, laptops, and general business traffic does not automatically fit controllers, drives, and cyclic I/O. The goals are different. The timing rules are different. Even the addressing logic can be different.
That is why he calls the idea of placing everything on the same network “silly and dangerous.” His point is blunt, but the reasoning is consistent. If the network is part of production, you cannot treat it as if it were only a shared business service.
The line between IT and OT still matters
When people say IT and OT are both Ethernet, they describe the cable and miss the job. The harder differences are purpose, timing, and addressing, and those differences shape the whole network.
That is why IT and OT should not be treated as one uniform system. They can exchange data, but they do not follow the same rules on the factory floor.
![Voltage Sag vs Interruption: Causes, Impact, and Fixes A plant can lose a production line from a blink of power, even when the lights come back almost at once. If you've seen a VFD trip, a contactor drop out, or a PLC reset after a split-second dip, you've seen power quality turn into a production problem. The issue is often not a full outage. It's a short voltage event that sensitive equipment can't ride through. Start with the basics, and the failure starts to make sense. What voltage sag and interruption mean A voltage sag is a short drop in RMS voltage below normal, usually to 10% to 90% of rated voltage, for 0.5 cycles up to 1 minute. In a 415 V system, a brief drop to 280 V or 250 V is a sag, not a blackout. Duration matters. If voltage stays low for more than a minute, that is usually undervoltage, not sag. A sag arrives fast, recovers fast, and can still stop a machine. This quick comparison makes the difference easier to see: EventWhat happensTypical durationVoltage sagVoltage drops but does not go to zero0.5 cycles to 1 minuteVoltage interruptionVoltage is zero or near zeroLess than 1 minuteUndervoltageVoltage stays below normal for longerMore than 1 minute An interruption is more severe because supply is lost completely, or almost completely, for less than a minute. If it clears in a few seconds after auto-reclosing, it is a momentary interruption. If it stays off beyond a minute, it becomes a sustained interruption. Why these events happen The most common cause is a fault on the power system. That could be a single line-to-ground fault, line-to-line fault, double line-to-ground fault, or a three-phase fault. When fault current rises, voltage drops across the network until protection clears the problem. If the fault is on your feeder, you may see a sag first and then an interruption when the breaker opens. If the fault is on another feeder from the same substation, your breaker may never trip, but your plant can still see a bus voltage dip. That is why equipment can trip even when "our feeder never opened." Large motor starting is another frequent cause. An induction motor can draw five to seven times full-load current during start. In a weak system, or where the motor is large compared with the transformer, that inrush can create a temporary sag. Transformer energization, capacitor switching, welding loads, arc furnaces, and sudden heavy loading can do the same. Why a tiny dip can stop a large machine > The main motor may ride through a sag, but the control power often won't. Older plants had more electromechanical loads, and many of them tolerated short dips. Modern plants rely on PLCs, VFDs, servo drives, electronic power supplies, sensors, relays, and SCADA. Those devices make automation possible, but many are more sensitive to voltage dips than the motor they control. Massive steel control panels and heavy machinery dominate the floor as overhead lights cast a chaotic, flickering glow. Sharp shadows and sparks suggest a sudden surge in the facility power grid. [https://user-images.rightblogger.com/ai/f382171e-d1b1-4320-b7eb-289d9b53ee27/industrial-factory-power-instability-93e17dc7.jpg] A short sag may not stop a spinning motor because inertia keeps it moving. Still, the contactor coil can drop out, the VFD can detect undervoltage, and the PLC power supply can reset. Once the control chain breaks, the process stops. In process plants, that can mean lost batches, reset time, scrap, labor loss, and delayed delivery. Magnitude and duration both matter. Some equipment can tolerate 80% voltage for five cycles, but not 40% for the same time. That is why ride-through curves matter, and why event recording matters too. Good monitoring tools, such as monitoring power quality with PME 2024 R2 [https://www.interestingautomation.com/schneider-pme-2024-r2/], help capture minimum voltage, duration, and affected phases. Practical ways to reduce voltage sag problems The most cost-effective fix starts with the weak point. If a 200 kW machine trips because a 230 V PLC supply resets, you usually do not need to protect the whole machine. You need to protect the control power. * Specify ride-through performance when buying critical PLCs, drives, relays, and controls. * Add a small UPS, DC backup, or capacitor ride-through module for control power. * Use a voltage sag compensator or dynamic voltage restorer for sensitive process loads. * Apply online UPS systems where transfer time cannot be tolerated. * Consider motor-generator or flywheel systems where short interruptions happen often. * Use static transfer switches only when the two sources are truly independent. Source quality matters too. Utilities reduce events with better protection coordination, faster fault clearing, line maintenance, tree trimming, and feeder automation. On the plant side, grid automation and fault visibility also help, which is why tools for using Easergy T300 for fault detection [https://www.interestingautomation.com/brief-explain-easergy-t300-features-benefits-and-complete-guide/] are relevant in systems that need faster disturbance response. Final thoughts A blink in voltage can do more damage to production than a short outage, because the failure often happens inside the control system before anyone sees a breaker trip. That is the core lesson behind voltage sag and interruption studies. The best fix is rarely the biggest one. Find what actually trips, measure how deep and how long the event lasts, and protect the most sensitive part first. A brief dip should not turn into hours of downtime.](https://www.interestingautomation.com/wp-content/uploads/2026/05/Voltage-Sag-vs-Interruption-Causes-Impact-and-Fixes-150x150.jpg)
