Hosts Ed Fuentes and Phil Seboa welcome engineer Melvin Francis, who shares his expertise on real-time industrial communications, Edge AI’s impact on IIoT, and overcoming OT-IT integration roadblocks. These topics matter for readers seeking better performance and reliability from their automation systems.
Real-Time Data and Time-Sensitive Networking in Industrial Automation
The growing use of smarter, connected devices on the factory floor means that getting accurate, on-time information is no longer optional—it’s essential. Melvin Francis points out that, "If you make a mistake in IoT...your Alexa is going to switch on the bulb a few seconds later. But what happens in industrial automation is people are going to die. There's going to be huge catastrophes that are going to happen."
Real-time data is not just about speed; it’s about delivering information reliably at specific, predictable intervals. This requirement is met by technologies like Time-Sensitive Networking (TSN), which allows critical control data (such as robot commands or safety interlocks) to coexist with less critical information (like video or maintenance logs) on the same Ethernet backbone. TSN ensures every data packet gets its guaranteed place and time on the network, eliminating unpredictable delays and mishaps. Melvin demonstrates this vividly, explaining, "When you bring in the component of TSN...all the devices [are] time synchronized in a level of nanoseconds, plus or minus 10 nanoseconds...everybody gets a bandwidth split."
For those responsible for industrial safety and uptime, investing in TSN-backed communications directly improves reliability and the ability to integrate newer technologies without sacrificing performance or security.
Edge AI: Moving Intelligence Closer to the Plant Floor
Cloud computing began as a panacea for data analysis, but for industrial environments, security concerns and latency issues limit the feasibility of sending all data to the cloud. Edge AI brings analysis and decision-making capabilities directly onto the factory floor, close to the machines and sensors generating the data.
As Melvin Francis explains, "People want to have their AI models running at the edge. Decision needs to be available in their factory floor and they want to be as secure as possible." Edge AI enables real-time monitoring, predictive maintenance, and anomaly detection, all without having to send sensitive operational data off-site. Melvin shares a case where Edge AI flagged issues with robotic palletizing in a packaging plant: By monitoring motor torque locally, the system detected empty boxes and process bottlenecks—even when these problems occurred sporadically.
Edge AI also maximizes existing equipment through “brownfield” deployments, layering advanced analytics onto legacy systems. This means companies can access the benefits of AI without a costly rip-and-replace of their automation investments.
Bridging the IT-OT Divide: Secure and Manageable Integration
The traditional gap between IT (information technology) and OT (operational technology) poses major headaches, especially as integration becomes critical for advanced analytics and remote access. Security, data latency, and conflicting priorities between business systems and plant operations often complicate integration projects.
Melvin Francis describes, "The ultimate thing is when you want the real time data or the robot control data that is really, really critical to be sent out...your real time data should not be interrupted by your IT data." By segmenting networks with TSN and adopting communication standards like OPC UA (which offers flexible publish/subscribe and client/server options), plants can keep critical operations safe and deterministic while still providing IT systems with the insights they need.
The right protocols and architecture ensure clean, secure, and actionable information for all stakeholders—without putting safety or core processes at risk. “OPC UA can use MQTT...OPC UA is the middleman who's going to say this is how your information modeling needs to be done. This is how you need to pack your data,” Francis notes. This approach leads to fewer headaches, better uptime, and a clearer path for future technology rollouts.
Key Quote From The Episode
"When the data has to be delivered, it should be delivered on time. It can be slow, but if I say that it has to reach at 12 o'clock, it should be 12 o'clock on time with plus or minus 10 microseconds of jitter." - Melvin Francis
Key Takeaways
[00:06:40] Safety and reliability depend on true real-time data. Industrial operations require precise, dependable timing that general-purpose networks cannot provide.
[00:10:51] Edge AI brings advanced analytics and anomaly detection directly to the plant floor, minimizing security risks and latency.
[00:12:35-00:20:48] Integrating IT and OT successfully requires advanced protocols like OPC UA and TSN to separate, prioritize, and protect critical communications.
Wrap Up
Real-time networking, Edge AI, and secure IT-OT integration are no longer luxury features— they’re requirements for any organization aiming to increase safety, productivity, and innovation in industrial automation. Readers should assess existing network capabilities, investigate TSN and Edge AI solutions, and work toward standards-based integration using protocols like OPC UA. Practical steps include piloting Edge AI at a single machine, segmenting network traffic by criticality, and engaging in continuing education to keep skills current.
About the Guest
Melvin Francis is a research-driven engineer specializing in industrial Ethernet, deterministic networks, and embedded automation. He currently leads technical marketing at BE Services in Germany, focusing on TSN, OPC UA, and Edge AI deployments for Industry 4.0. Francis shares frequent educational content for the automation community on LinkedIn and is dedicated to making advanced automation accessible and practical.
