Guest: Bart Nieuwborg
Project/Organization: The Margo Initiative | The Linux Foundation
Show Name: The Source
Topics: Edge Computing, Open Source
 

Industrial automation is at a breaking point. Companies are moving toward software-defined operations, deploying more applications at the edge than ever before. But there’s a problem: these systems don’t talk to each other. Proprietary platforms force painful choices between innovation and operational stability. Bart Nieuwborg, Chair of the Margo Initiative, is working to change that with an open source framework designed to solve interoperability at the industrial edge.

The Multi-Vendor Reality

The Margo Initiative launched less than two years ago with six founding companies. Today, 40 vendors are contributing to solve what Nieuwborg calls “the multi-vendor reality” of industrial automation.

“Users in industrial automation have a specific challenge during their transition toward software-defined operations,” Nieuwborg explains. “They’re going to have to use more and more software at their plants, and a significant number of those software instances are going to be running at their edge—on premise in their facilities. These are things they’re going to have to manage at scale.”

The core problem is simple but stubborn. Companies can’t source all their software from a single vendor. They need best-in-breed applications for data collection, AI, edge analytics, and more. But managing these applications across different hardware platforms and fleet management tools creates operational nightmares.

Margo’s answer is a specification for interoperable application management at the industrial edge. The project recently announced Preview Release 1 (PR1), delivering the first vertical slice of this vision.

What Preview Release 1 Delivers

The preview release tackles three key deliverables: a specification, a sandbox environment, and conformance toolkits. Nieuwborg emphasizes that this is a “code-first experience” rather than just documentation.

“We didn’t want to just give people a specification that exists in a PDF file,” he says. “We built a sandbox—an open-source set of code pieces that, when brought together, create a build system. If you’re a potential adopter, you can download it, deploy it in your test environment, and actually see the interoperability happening.”

The focus for this release is application management—specifically, lifecycle management of applications at scale. This priority came directly from Margo’s advisory board, a group of industrial automation users spanning pharmaceuticals, tire manufacturing, packaged goods, and other sectors.

The preview includes install scripts, user stories from different personas (application developers, device manufacturers, fleet managers), and prototypes from community members showing how commercial products could become Margo-compliant.

How It Works in Practice

For enterprises, implementing Margo doesn’t require wholesale infrastructure changes. Nieuwborg describes a straightforward adoption path.

“They continue to source their software applications and hardware platforms from the vendors they’re already working with,” he explains. “They just ask those vendors to make their products Margo-conformant.”

This means three things: applications packaged with Margo-compliant descriptors, edge devices that express their capabilities in Margo language, and fleet management tools that can deploy and manage applications using Margo APIs.

“For the user, there is barely anything to do,” Nieuwborg notes. “The work is going to be on the suppliers. There’s going to be an effort for an app developer to package their application in a Margo-conformant way—describe what the application is capable of doing and what it requires from devices. But that one-time effort gives them market access and allows them to focus on their differentiating features.”

The Roadmap to General Availability

Preview Release 1 is exactly that—a preview. The community is gathering feedback through discourse forums and GitHub, with a structured process for turning input into specification improvements.

“We’re humble,” Nieuwborg says. “We might find mistakes. If we remove them now, the final release will be higher quality.”

The target is a general release in 2026, but it won’t cover Margo’s complete scope. The project is taking an iterative approach, releasing incremental updates rather than waiting years for a comprehensive specification.

“We didn’t want to wait another two years before we’d say our spec is done,” Nieuwborg explains. “This year we tackled application management. What’s next is likely going to be device management. We want regular cadence—not let people wait five years before they see an outcome.”

The three main pillars Margo will eventually address are application management, device management, and observability. Preview Release 1 covers application management with associated observability features.

Industry Traction and Testing

The growth from six founding companies to 40 contributing vendors in under two years signals market validation. But Nieuwborg is most excited about what users are doing.

“Our advisory board isn’t just giving advice,” he says. “They’re thinking about providing test areas in their facilities—real production environments where we can build and test in real time, not on virtual machines.”

This real-world testing is planned for 2026, allowing the community to validate the specification against actual industrial use cases before finalizing the general release.

Getting Involved

The community has built multiple feedback channels. Non-developers can provide input through discourse forums on margo.org. Developers can engage via GitHub, creating issues or submitting specification update proposals through a structured review process.

“We want to make feedback low-barrier,” Nieuwborg says. “Whether you just drop comments in our tooling or use the GitHub environment for more detailed contributions, both channels are designed to be traced and turned into usable development outcomes.”

There are no membership fees. The project follows typical open source principles—contribution-based and transparent.

For industrial automation companies facing the reality of multi-vendor environments, Margo represents a potential path out of the interoperability trap. Preview Release 1 makes that path tangible, turning abstract specifications into working code that can be tested, broken, and improved before commercial products build on it.

As Nieuwborg puts it: “Try it out, give us feedback, break it, and improve it.”