Skip to content

Search the site

An industrial "love language"? It starts with good communication

Using MQTT and MongoDB to transform industrial data.

Are you ready to get geeky? This is a story about MQTT and UNS, about SCRAM and Modbus and ISA 95 and SCADA systems. It is also about gleaming new robotic arms whirling on factory floors, and thundering rusty leviathans running critical operations in industrial facilities. 

Mostly though it’s about communication, because improving any relationship means finding a way to speak a common language; e.g. one that “engineers” with keyboards and “engineers” with engines can start to converge on. In the IT and OT worlds, this new relationship is needed.

IT and OT: Can we sit down for drinks?

Culturally the Information Technology (IT) world of “move fast and break things” and the Operational Technology (OT) world of “move gingerly and don’t break anything” have long spoken different languages. IT professionals can often “turn it off and turn it on again”; OT professionals cannot; “rebooting” a global production line is not a popular move. 

As Dominik Obermaier, CTO of German startup HiveMQ tells The Stack: “People on the operations technology side need stuff to work 24/7!”

“The timeframes you're looking at are usually much much longer than you have in IT. People [in the OT world] build systems and programme PLCs that have to work for 10-years plus; 20-years plus” he emphasises. 

That means from automotive to pharmaceuticals, OT systems are often, simply, old – and even more modern robotics hardware frequently ships with legacy operating systems or communicates using older protocols. 

“I don’t understand you”

The notion of speaking different languages is also more than just a metaphor: The IT and OT worlds’ communication protocols diverge.

Modbus is one example: Created in 1979, the robust, open automation protocol is widely used to connect electronic devices used in industrial automation. But, built as it was for air-gapped locations, it operates with no user authorisation, device authentication, nor data encryption.

For someone connecting industrial devices to the “outside world” (and increasing its threat surface) such issues become rather problematic. 

Enter HiveMQ…

“HiveMQ really sits between these two worlds, connecting Operational Technology, but also Information Technology”, says Obermaier. It has built an impressive roster of blue chip customers as a result: Audi, BMW, Liberty Global, Mercedes-Benz, Siemens, among many others.

It has done so by providing an enterprise platform for MQTT (Message Queuing Telemetry Transport); an increasingly popular, standardised data exchange protocol for Internet of Things and industrial IoT messaging that provides a scalable and reliable way to connect devices over the Internet.

(This often includes transforming Modbus data into MQTT messages…)

Enter MongoDB…

Via a new integration with MongoDB, it is also now providing a simple way to convert the millions of MQTT messages flying about a factory floor into MongoDB documents leveraging MongoDB flexible document model

Working together, the two companies can format these for streamlined data querying, and even forward MQTT messages from IoT devices to MongoDB databases in a secure and properly authenticated manner.

As MongoDB’s Dr. Humza Akhtar puts it simply: “People running factory operations are trying to modernise the way they are collecting information from machines. Those older protocols extract the data from the controllers, and then they publish it into MQTT topics. HiveMQ acts as their central nervous system, getting data from all these different lines, all these different machines, robots, various topics are being produced.”

Then once you have the data you need to store it. 

“That's where MongoDB comes in!” he says, adding: “As the data comes in, Atlas Triggers are used to run custom functions that look at documents coming in, identifies which exact MongoDB collection they have to go to (alarms, levels and so on) and store the data in the right place.”

Enter Unified Namespace 

For those looking to pull data from complex OT landscapes (to improve productivity, build digital twins, pre-empt machine failures with predictive maintenance) and use it to power new applications, this is a potential game-changer – particularly when architecting a “Unified Name Space” (UNS).

End users can, for example, use MongoDB Atlas as the data and feature store for machine learning and build predictive models for pattern identification and trend analysis; store and analyse IoT data to identify trends and patterns over time, and create powerful visual dashboards. 

See also: MongoDB takes Atlas to the Edge

The UNS approach, meanwhile, creates a “contextual data structure and common endpoint for systems to connect and exchange information.”

HiveMQ and MongoDB have been collaborating extensively on UNS design, for example for an application that can tell you “the performance of your factory over time” in as granular or top-level a way as desired. 

It’s a compelling destination, but the journey is noteworthy too. 

A central nervous system for industrial data

As Obermaier puts it: “Historically [the OT world saw] really siloed data.

“There are not a lot of incentives in the market these days to break the data up. What we've been working on is making it easy to bring non-MQTT into an MQTT landscape; a ‘central nervous system’ accessible everywhere in the organisation with a Unified Namespace.

“Earlier this year we released an open-source product called HiveMQ Edge… that connects to a lot of legacy technologies, like Modbus, OPC, UA, and other technologies in a standardised way… this is very important, because historically, you've seen a lot of vendors that are pushing their own implementations; there's not enough incentive to be interoperable.

To Humza the tools to help IT and OT connect more safely and effectively are really reaching fruition in a way that means innovation is ever easier. 

He highlights MongoDB’s “Atlas for the Edge”, which offers “bi-directional synchronisation” with the cloud: “You can have HiveMQ edge translating these [OT] protocols and MongoDB Atlas providing the flexible and scalable IIoT data storage and contextualization platform for UNS. MongoDB Atlas Edge Server enables data sync between cloud and edge devices delivering real time application experiences to the edge with high reliability.. We see a lot of opportunities for collaboration in this space…” 

Meaningful connections in a safe, organised space? 

Something even robots can get behind. 

Delivered in partnership with MongoDB.