IIoT on blanpa

Securing OT Networks — The Purdue Model, OPC-UA Encryption & Segmentation

Sat, 18 Jul 2026 00:00:00 +0000

Every other post in this series has been about connecting things — PLCs to dashboards, protocols to brokers, edge to cloud. This one is about the uncomfortable flip side: every connection you add is also an attack surface. Industrial systems were designed in an era of physical isolation, and we’ve spent the last decade enthusiastically networking them without always asking what that means for security. This is the post I wish more IIoT developers read before their first deployment, not after their first incident.

Modbus in Node-RED — The 47-Year-Old Protocol Everyone Still Uses

Sat, 11 Jul 2026 00:00:00 +0000

I’ve written about OPC-UA, EtherNet/IP, and a whole zoo of modern protocols. But if I had to bet on which protocol you’ll actually encounter most often on a real shop floor, my money is on the oldest one in the room: Modbus. Designed by Modicon in 1979, it’s on power meters, VFDs, temperature controllers, flow meters, HVAC gear, and ten thousand other devices. It’s gloriously simple, slightly maddening, and absolutely everywhere. Here’s how to talk to it from Node-RED.

Talking to Allen-Bradley PLCs from Node-RED — EtherNet/IP & CIP Explained

Sat, 27 Jun 2026 00:00:00 +0000

I’ve written about getting a Siemens S7 onto a dashboard via OPC-UA. But walk into a North American plant and you’re far more likely to find an Allen-Bradley PLC — a ControlLogix in the big cells, a CompactLogix on the smaller machines, maybe a dusty SLC 500 still running a line nobody dares touch. And Allen-Bradley doesn’t speak OPC-UA out of the box. It speaks EtherNet/IP and CIP. This is the guide to bridging that gap with Node-RED.

Apache Kafka on the Shop Floor — Event Streaming for Manufacturing

Sat, 20 Jun 2026 00:00:00 +0000

Kafka has a reputation problem in manufacturing. Half the people I talk to think it’s the answer to everything; the other half think it’s a heavyweight cloud thing that has no business near a PLC. Both are wrong. Kafka is a fantastic tool for a specific set of shop-floor problems — and a terrible choice for the rest. This post is about telling the two apart, and then actually wiring Kafka into a Node-RED stack once you’ve decided it fits.

Lessons Learned: Publishing 7 npm Packages for Industrial IoT

Sat, 23 May 2026 00:00:00 +0000

Update (July 2026): the family has since grown to eight packages — node-red-contrib-kafka-suite joined after this post was written. Everything below still applies.

Since November 2025 I’ve published seven npm packages for Node-RED, all targeting the IIoT space. Some took off, some needed multiple rewrites, and all taught me things I couldn’t have learned from tutorials alone. This post is the honest rundown: what worked, what didn’t, and what I’d do differently.

From Process Engineer to IIoT Developer — My Career Switch

Sat, 16 May 2026 00:00:00 +0000

Two years ago, I was a process engineer in a manufacturing plant. I wore steel-toed boots, carried a clipboard, and spent my days optimizing cycle times and reducing scrap rates. Today, I build Node-RED modules for industrial IoT, maintain open-source packages, and write code that connects factory machines to dashboards and analytics systems.

This is the story of how I switched — what pushed me to leave, what I had to learn, what transferred surprisingly well, and what I’d tell someone considering the same move.

CI/CD for Node-RED Flows — Automated Testing and Deployment

Sat, 09 May 2026 00:00:00 +0000

Node-RED is fantastic for building IIoT integrations quickly. Drag, drop, wire, deploy — you have a working MQTT-to-database pipeline in minutes. But then comes the question every production deployment forces you to answer: how do I version this? How do I test it? How do I deploy it without clicking buttons?

The answer is treating Node-RED flows as code — putting them in Git, writing automated tests, building Docker images, and deploying through CI/CD pipelines. This post shows the complete setup, from repo structure to GitHub Actions.

CAN Bus Reverse Engineering with Node-RED and Raspberry Pi

Sat, 02 May 2026 00:00:00 +0000

If you’ve ever stared at a VFD or servo drive and thought “I know this thing has useful data — I just can’t get to it,” CAN bus is probably the answer. Most industrial drives, controllers, and even some sensors speak CAN internally. With a Raspberry Pi, a $15 CAN hat, and some patience, you can tap into that data stream and make it visible in Node-RED.

This post walks through the entire process: hardware setup, Linux socketcan configuration, traffic sniffing, frame decoding, Python scripting, and finally Node-RED integration for real-time dashboards.

Siemens S7 + OPC-UA + Node-RED — From PLC to Dashboard in 30 Minutes

Sat, 25 Apr 2026 00:00:00 +0000

This is the tutorial I wish I had when I first tried connecting a Siemens PLC to Node-RED. It took me two days of frustration — fighting with certificates, firewall rules, and data type mismatches — before I got a single value to appear on a dashboard. With this guide, you’ll do it in 30 minutes.

Prerequisites
#

Before starting, make sure you have:

Docker vs K3s on the Shop Floor — Edge Deployment Strategies

Sat, 18 Apr 2026 00:00:00 +0000

Containers on the shop floor used to be a hard sell. “Why not just install the software directly?” plant engineers would ask. After rolling back a failed manual update at 2 AM for the third time, the answer becomes obvious: reproducibility, isolation, and the ability to roll back in seconds instead of hours.

The real question isn’t whether to use containers in industrial environments — it’s whether to use Docker Compose or go all in with K3s (lightweight Kubernetes). I’ve deployed both. Here’s when each makes sense.

REST vs OPC-UA vs GraphQL for Manufacturing Data

Sat, 11 Apr 2026 00:00:00 +0000

“Just give me the temperature and status of that machine” — simple enough, right? But the how matters enormously. Do you poll a REST endpoint? Subscribe via OPC-UA? Query a GraphQL schema? Each paradigm makes different trade-offs between simplicity, performance, and expressiveness, and in manufacturing those trade-offs have real consequences.

I’ve used all three in production factories. Here’s what I’ve learned.

The Same Query, Three Ways
#

Let’s start with a concrete example. We want to get the current temperature and status of a CNC machine with ID cnc-001.

Node-RED vs Kepware vs Ignition — IIoT Platforms Compared

Sat, 04 Apr 2026 00:00:00 +0000

Choosing an IIoT platform is one of those decisions that haunts you for years. Pick the wrong one, and you’re locked into expensive licenses, limited integrations, or a tool that doesn’t scale. Pick the right one — or the right combination — and you’ve built a foundation that grows with your factory.

I’ve worked with all three platforms extensively: Node-RED as my daily driver for data transformation, Kepware as the connectivity workhorse, and Ignition as the SCADA powerhouse. Here’s my honest comparison.

Building a NATS-based Edge-to-Cloud Pipeline for Industrial IoT

Sat, 21 Mar 2026 00:00:00 +0000

MQTT got us far. Every IIoT project I’ve worked on started with Mosquitto, some topics, and QoS 1. It works — until you need message replay, exactly-once delivery, or synchronizing configuration across 30 factory sites. Then you start bolting on Kafka for persistence, Redis for key-value state, and a custom sync mechanism for edge-to-cloud. Three systems to operate, three failure modes to debug at 2 AM.

NATS replaces all three. It’s a single binary that handles pub/sub messaging, persistent streams (JetStream), key-value stores, and edge-to-cloud synchronization (leaf nodes) — all with built-in clustering and zero external dependencies.

Running ML Models at the Edge with ONNX and Node-RED

Sat, 14 Mar 2026 00:00:00 +0000

Every time I hear “just send the data to the cloud for inference,” I think about the CNC spindle that threw a bearing 200 ms after the anomaly appeared. The round-trip to AWS takes 150 ms on a good day — and that’s assuming the factory Wi-Fi doesn’t drop the packet. By the time the cloud responds, the spindle is already damaged.

Edge inference means running ML models directly on the device next to the machine. No cloud latency. No bandwidth costs. No dependency on internet connectivity. This post covers how to train a model, export it to ONNX, and deploy it on an edge device through Node-RED.

Predictive Maintenance with Node-RED — From Vibration Sensor to Health Score

Sat, 07 Mar 2026 00:00:00 +0000

A bearing doesn’t fail without warning. It screams — in frequencies you can’t hear. The trick is listening at the right frequency, recognizing the pattern, and acting before the $80 bearing destroys the $40,000 spindle.

This post walks through building a complete predictive maintenance pipeline in Node-RED: from raw accelerometer data to a health score your maintenance team can act on.

Why Predictive Maintenance?
#

Three maintenance strategies exist. Only one saves money without accepting unplanned downtime:

i3x — The Open Manufacturing API That Breaks Data Silos

Thu, 05 Mar 2026 00:00:00 +0000

Every factory I’ve worked in has the same problem: data is everywhere, but accessible nowhere. The historian speaks SQL. The MES has a SOAP API. The SCADA system uses OPC-UA. The new IoT platform wants REST. Connecting all of them means writing custom adapters for each combination — and maintaining them forever.

i3x changes this. It’s an open REST API specification that gives you one interface to all manufacturing data, regardless of what platform stores it. I built a Node-RED integration for it, and in this post I’ll explain what i3x is, how it works, and why it matters.

MQTT vs Sparkplug B vs NATS vs OPC-UA — Which Protocol for Industrial IoT?

Thu, 05 Mar 2026 00:00:00 +0000

Choosing the right messaging protocol for an Industrial IoT project is one of the most impactful architectural decisions you’ll make. Pick the wrong one and you’ll spend months working around its limitations. Pick the right one and data flows effortlessly from sensor to cloud.

In this post I compare MQTT, Sparkplug B, NATS, and OPC-UA — four protocols I work with daily in industrial environments. No theoretical fluff — just practical differences, real trade-offs, and concrete examples.

IIoT on blanpa

Securing OT Networks — The Purdue Model, OPC-UA Encryption & Segmentation

Modbus in Node-RED — The 47-Year-Old Protocol Everyone Still Uses

Talking to Allen-Bradley PLCs from Node-RED — EtherNet/IP & CIP Explained

Apache Kafka on the Shop Floor — Event Streaming for Manufacturing

Lessons Learned: Publishing 7 npm Packages for Industrial IoT

From Process Engineer to IIoT Developer — My Career Switch

CI/CD for Node-RED Flows — Automated Testing and Deployment

CAN Bus Reverse Engineering with Node-RED and Raspberry Pi

Siemens S7 + OPC-UA + Node-RED — From PLC to Dashboard in 30 Minutes

Prerequisites #

Docker vs K3s on the Shop Floor — Edge Deployment Strategies

REST vs OPC-UA vs GraphQL for Manufacturing Data

The Same Query, Three Ways #

Node-RED vs Kepware vs Ignition — IIoT Platforms Compared

Building a NATS-based Edge-to-Cloud Pipeline for Industrial IoT

Running ML Models at the Edge with ONNX and Node-RED

Predictive Maintenance with Node-RED — From Vibration Sensor to Health Score

Why Predictive Maintenance? #

i3x — The Open Manufacturing API That Breaks Data Silos

MQTT vs Sparkplug B vs NATS vs OPC-UA — Which Protocol for Industrial IoT?

Prerequisites
#

The Same Query, Three Ways
#

Why Predictive Maintenance?
#