How Orange and HiveMQ are Streamlining Automated Edge Rollouts

Rolling out industrial edge infrastructure at scale is difficult. Each site has unique equipment, network constraints, and security requirements. Manually setting up gateways, brokers, and data pipelines is not only time-consuming but also invites inconsistencies that can lead to errors or downtime.

That’s why forward-thinking operators are turning to automation. By containerizing edge components and using configuration-as-code tools, they can deploy consistent, secure environments across dozens or even hundreds of facilities in a fraction of the time.

From Challenge to Blueprint

Orange wanted to simplify the way industrial customers bring new sites online, while keeping data quality and security uncompromised. Large industrial and telecom operators face the same issue everywhere: connecting fleets of equipment across multiple locations can balloon operational costs if every edge device has to be configured by hand. Manual rollouts are slow, inconsistent, and prone to misconfiguration, while automation provides the speed and repeatability needed to keep systems reliable.

Containerizing edge components and managing them with infrastructure-as-code tools changes the equation. It allows updates and patches to be applied as easily as in modern DevOps pipelines, and pairing that with a clear Unified Namespace means data is organized and ready for analytics, AI, or digital twins from day one.

To validate this approach, Orange and HiveMQ built an end-to-end pipeline modeled on a “real-world equipment” template from a global food manufacturer. The scenario focused on a liquid-mixer setup, incorporating realistic sensor data structures and tag definitions to simulate production conditions. 

The rollout was based on OPC UA connected machines and a GitOps operating model, with all edge configurations version controlled and delivered via a CI/CD approach. Local groupings followed ISA-95 and ISA-88 standards, enabling structured, scalable deployments that reflect how manufacturing systems are typically organized. Metadata from OPC UA was carefully mapped into the Unified Namespace (UNS), with thoughtful handling based on relevance: some metadata was embedded directly into the JSON payload, while others were carried in MQTT 5’s User Properties. This configuration ensured that the data was not only production-ready but also enriched, versioned, and fully contextualized from edge to enterprise.

Inside the Automated Edge Architecture

At the heart of the solution is a fully containerized edge environment that standardizes how data is collected, transported, and contextualized across sites. Its key building blocks include:

• Simulated manufacturing data generated by a custom datapoint tool, feeding an OPC UA server.

• HiveMQ Edge gateways running in Docker containers at the factory level.

• Factory and HQ brokers handling reliable, MQTT-based data transport.

• Automated configuration with Ansible, ensuring every broker and gateway inherits a global template while allowing site-specific overrides.

• Contextualized data enriched with metadata such as units of measure and valid ranges.

• A Unified Namespace (UNS) organizes all tags into a clean hierarchy so teams can instantly understand where each data point comes from.

This design supports a true “deploy once, repeat everywhere” model, letting new locations start from a proven baseline while integrating cleanly with the broader manufacturing tool ecosystem.

How the Automated Edge Rollout is Shaping Future Deployments

Early results show that automation delivers measurable gains. Rollouts that used to take weeks can now be completed in hours, thanks to templated deployment and CI/CD pipelines—a reduction in rollout costs estimated at 60–70%. Full version control through Git ensures every configuration is traceable and reproducible, while reducing manual handling significantly lowers the risk of errors.

The automation also strengthens operational resilience. By avoiding bad configurations and unplanned downtime—each of which can cost tens of thousands per incident—operators maintain consistent, high-quality data flows. Integration costs from the UNS to other systems have been reduced by up to 70%, and the data is now 100% available, consistent, and fully contextualized across sites.

The collaboration positions Orange to offer a repeatable, edge-ready solution to other enterprise customers, while HiveMQ continues to provide the MQTT backbone and Unified Namespace foundation for reliable, contextualized data flows. As the two teams build on this blueprint, it’s set to become a model for how telecom operators and manufacturers can jointly unlock the value of Industrial IoT.