Join the Insider! Subscribe today to receive our weekly insights
Willem (00:00)
Welcome, you’re listening to the IT/OT Insider Podcast. I’m your host Willem and in this special series on industrial data ops, I’m joined by my co-author David. And today, today we brought you Dominik Obermaier from HiveMQ.
Dominik (00:14)
Thank you so much for having me.
David (00:16)
This is actually the final episode of our DataOps series, and we go out with a bang. I think HiveMQ is probably one of the, or maybe the best known name when we’re talking about MQTT. So I’m obviously really happy to have Dominik here with us.
13 years ago, we co-founded HiveMQ and today he’s their CEO. yeah, Dominik, also from my side, thank you for joining me and Willem. And let’s start with maybe your personal backstory.
Dominik (00:48)
Absolutely. Thank you. So my name is Dominik Obermaier. I am the CEO and also one of the founders of HiveMQ. It’s really interesting that these days I spend a lot of time with OT people, but also pretty much on the IT-OT rift that many of our customers are observing. Because like when going back to my start of the career before I started computer science and then founded the company,
I was working at the MES vendor for the German automotive industry. So I started pretty much like the very first two weeks of my job was on the shop floor of a very large German car manufacturer. So learn a lot what’s happening if you see downtime. this, think, was an experience that I still pay dividends all across my career.
But going forward, then I started the company Hive MQ in 2012. So quite a while ago, I bootstrapped the company together with some very great customers, especially in Germany. We started to work with customers like BMW very early on. And since 2014, when the MQTT specification was pretty much developed, and I was one of the people who specified MQTT in 2014,
Since then, we’re all in on the MQTT technology, which is really interesting because now moving forward to, I would say 2020, this is really where the concept of the unified namespace started to get a lot of traction, especially on the OT side and especially manufacturing, but also in energy, be frank, not only manufacturing, also in energy sector. And since then, also here at HiveMQ, we spend a lot of time connecting devices and assets.
on the shop floor and really from Edge to the cloud. So what we do here at HiveMQ is we have a flagship product that’s also called HiveMQ and that is used by the best and also largest companies in the world to connect all of their physical assets from Edge to cloud and we provide this full platform based on the open standard of MQTT to connect pretty much all of the sites a company has, all of the assets.
provides real-time visibility. And with the new product, what we call HiveMQ Pulse, that was released in February this year, we now also offer a full, unified namespace solution for small companies, but also for the world’s largest companies. Example would be Lilly in the pharma space, who is the largest pharma company by market capitalization. But if you look at our customer wall, you will see
that in logistics, in energy, in manufacturing, and in automotive, that all of the top 10 leading brands in all of their sectors are our customers.
David (03:50)
That’s amazing and that trick is already a couple of follow-up questions, but actually I want to start with, like, so you are now 12, 13 years in the journey from bootstrapping to becoming a world leader in this space. How is the mission from HiveMQ today different than it was 12 years ago or is it still the same?
Dominik (04:17)
Yeah, it’s a great question. Our vision didn’t really change. Here at HiveMQ, we believe that every company is an IoT company. And what we mean with this is when we look back, let’s say to the 80s, computers were a big deal. And then we had companies like Microsoft who had the vision that a computer should be on a desk of every person.
which back then was like a, was a dream. And I mean, look at us now today as society, we don’t even have a computer on the desk. We have a computer in our pockets 24 seven these days. So the world changed quite a bit. And what also happened in this business change. So especially a bit before the eighties actually, but in the eighties and nineties in particular, computers were used everywhere. So every company got a computer company. If they thought of themselves as a computer company or not, computers are the backbone of what they did.
And in the nineties and obviously in the two thousands, the same happened with the internet. So every company today is an internet company. I recently was with my wife on a restaurant here locally in Lenswood, it’s near Munich. And I mean, we just booked it online and we paid with cashless. We paid with our debit card. And everything that was happening is this restaurant sees itself as a restaurant, but it’s actually a computer company, but also an internet company.
And even if you don’t think of it like that. And what we believe here at HiveMQ is that every company will be an IoT company. This means you’re blending the physical world and the digital world together. the physical world very often in the company connects our assets and the digital world very often our applications. And this now also leads us exactly to what we see when it comes to IT and OT, because this is exactly the conversions that also is happening in the industry.
Willem (06:10)
Maybe a question to get things started, because I’m sure that some listeners already know a lot about MQTT. For others, it’s maybe a word they heard, I think coming from the horse’s mouth, Dominik, how would you describe MQTT?
Dominik (06:26)
Yeah, MQTT is a technology. first of all, mean, technologies like exist and they’re neither good nor bad. But the interesting thing is now the question, why are so many people talking about MQTT and why are most of the digitization projects you find today in factories based on the MQTT standard? And when we go back to history, the computing history showed us that all of the technologies that we use today,
are based on open protocols and open standards. So many people have heard of TCP IP. This is pretty much the backbone of the internet. Then we have communication protocols like HTTP. This is used if you access, for example, with a browser or Wikipedia or Google or something like this. So it’s really built for the internet of humans. But there’s also other communication protocols. For example, the email protocol, there’s SMTP to send emails and all of these kind of things. And
I mean, usually people don’t care about technologies, right? We just want to use something that works. But the reason why we are interoperable and the reason why we actually can enjoy like an internet and why we can enjoy technologies today is because people standardize on some technologies. So it’s actually really important that you don’t go into proprietary technologies and as but into open standards that everybody’s using. And so in 1999, there was basically a communication protocol invented called MQTT.
for all pipeline monitoring. And back then nobody cared about open standards. Also the internet was pretty young. mean, HTTP was just a few years old. this way, we really liked the children days of the internet also. And then this MQTT protocol was used for all pipeline monitoring and it has a few key advantages. First of all, it was super lightweight on the bandwidth. Bandwidth was expensive. If you talk to Dr. Andy Stanford-Clark.
who was one of the inventors of the protocol, he once said in an interview that every byte, which is the most minimal amount of data you can send over wire, costs $10,000 per year. So every byte they could save was worth $10,000. So they were coming into the project to think, what’s the minimum amount of technology we can add to move the data around? And this also in a way that works for mobile networks, but also back then for cellular networks.
And so they built this, it was great, it was awesome. And then it was shelved, IBM used it for some other projects. And around 2010ish, the specification was opened up by IBM. And it was pretty much a promise, hey, if somebody’s building on top of this technology, we will not sue you. Which is not good enough for being an open standard. It’s good enough to get the best view, but it’s not. And so it took a few years.
David (09:17)
No.
Dominik (09:22)
until 2014 actually before our specification happened. And I was very lucky to meet the other inventor back then called R. Nipper. Back then he worked at a company called Eurotech. Before he was at Archem these days, he has a company called Serious Link, which is also in the MQTT space. And yeah, I met him and he told me about the technology. I thought, this is the answer to my question. What’s the technology for the internet of things?
what would be the winning technology because the technologies for the internet of humans is clearly not good enough. yeah, and so, I mean, at the end of the day, as with my fellow co-founders founded the company, built the first product, won the first customers and started like this with the company. And now again, fast forwarding to today, MQTT these days is going full circle. It was really invented more for operations technology, almost like for scaler systems. This was pretty much a scaler you get.
And all these characteristics, it’s open. It’s based on what’s called a published subscribe pattern. This means you’re decoupling the producers of data with the consumers of data. It’s very lightweight. And another characteristic, it’s super trivial to implement a consumer and a producer of data. So it’s very easy to implement it on any applications, even on PLCs and so on.
compared to other standards like OPC UA, I mean, just to bring it into perspective, OPC UA, depending on how you count, and it’s like 6,000 plus pages that you should implement when you’re implementing the full specification and all of the stack. MQTT in the first version had 40 pages or 46 pages. So this also shows MQTT is simple. And this is also why a lot of adoption was being done. First of all, it’s simple and it implements a decoupled pattern.
And these days, you find it into all of the factories globally for device-to-device communication, but also for, let’s say, shop floor to cloud communication and back.
David (11:31)
That’s a super cool backstory. It’s also, you know, it also shows that you have to be at the right place at the right time and something can happen. then the standard or at least the protocol exists already for so long, but it’s just that one thing.
Dominik (11:52)
Yeah. And it was also interesting. mean, the question is now, a question I get often is like, why is there so much pressure of companies to digitize? This is a question I get often these days. It’s like, mean, if you look at manufacturing companies, I mean, for a very long time, you could get around by not digitizing. And there is, of course, the spearhead of companies that are highly digitized. This is great. And everybody talks about them. their vast majority,
David (12:16)
Yeah, yeah.
Dominik (12:19)
of manufacturing companies are still, let’s say, not yet in the digital age. So for many companies, a huge success to actually have tablets, like digital tablets instead of physical tablets. so digitization is not evenly distributed across industry. Let’s put it like this. one of the things that really changed is since, I mean, actually since the 2020s.
Since 2018, we now have the understanding, okay, AI is more than a password and it’s more than machine learning. I mean, if look back, I recently did a keynote where I also talked about this. So if you look at the industrial revolution, you have the first industrial revolution, the second and the third, and then you have the steam engine, then you have electricity, then you have computers. You have all these kinds of key milestones that, and everybody back then was clear. It’s not like,
Like, we have to see mentioned, nobody believes in this. majority of people saw, okay, this is a huge game changer. Not everybody understood how it will get changed again, similar to electricity. mean, electricity is not like, electricity, nobody will need this. People are actually really bullish on this and think it’s a great technology. We’re just unsure how to use it. And with computers the same. mean, most everybody today believes computers is great. It’s not a lot of people think, let’s not do computers for manufacturing. It’s stupid, right?
David (13:47)
Hahaha.
Dominik (13:48)
I understand for
AI, we now have this technology, everybody understands we need that, this will change the game, but everybody’s figuring out how we will change the game. And in 20 years, we’ll look back and say, it’s so obvious, of course you do it like this, similar to, of course you use this dimension, of course you use electricity, but we’re now figuring it out with AI. In order to do AI, you need data. And you don’t only need some data, you need actual data that’s usable, that’s contextualized.
David (14:03)
Yeah.
Dominik (14:15)
is also governed in a proper way. is especially important in regulated industries. And also, you need data in real time. It doesn’t help you if you have old data, because decision making needs to happen fast. And then you only have the foundation to actually think about AI. And this is also why a lot of the digitization pressures, at least in our customer base, is something to be observed.
David (14:41)
I just wanted to point, let’s say one short remark you made here, because I think it’s really, true, is that in general, indeed, the manufacturing world still has lots of steps to take to become digital. And what you actually see is that obviously you have some lighthouse companies everywhere, and obviously every technology vendor and every consulting company, they promote their lighthouses. And it’s good that you say that because…
When I’m in discussion with many discussions I also have on a daily and a weekly basis, sometimes people say like, yeah, yeah, I know David, but compared to the rest, yes, we are indeed really slow. I go like, no, compared to the rest, you’re actually doing what the rest is doing.
Dominik (15:28)
Yeah, exactly. Exactly.
David (15:31)
But I would say to make it one step more concrete or more tangible. So we have this thing which we call the capability map. It’s a bit the common threads throughout the series where we talk from connectivity and contextualization and data quality, broker, analytics, sharing and visualization. What I think makes sense, Dominik, is to map what you are offering.
a bit to our capability map for two reasons. In the sense that first of all, you’re more than only a broker. I think that’s interesting also for our audience to understand that. And then secondly, you already briefly mentioned Pulse, which is a new, I would say a new announcement. And maybe you could also comment on what Pulse is then also adding to those capabilities.
Dominik (16:03)
Yeah. Yeah.
Yeah.
Yeah. Yeah. This is great. So thank you. And I think just for the listeners and to the viewers of this podcast, I mean, I really didn’t talk a lot about what HiveMQ does. So what you’re offering is really a full platform end-to-end based on open standards. And this is very important because HiveMQ customers don’t require other HiveMQ software to work. You can also use open source. You can use other vendors. We play very nice.
Of course, we believe that all of the components we have are the best in the market, of course, but everyone will tell you this. but we believe customers should have the freedom of choice, especially when they build a backbone for the next 20 to 30 years. the decisions, so it’s very hard to predict the future and to make the right decisions. And so what I urge all the customers is make some decisions based on principles. As an example, I want to use open standards, not based on vendors. I mean, you don’t know what vendors will exist in 30 years. Like hopefully all of them.
But the reality is that you need to make smart choices. Your business is relying on a digital backbone. So I urge you to open standards. And MQTT is the obvious choice here. The good thing is there are other vendors in HiveMQ. But of course, I will now talk about HiveMQ So what you’re offering is the flagship product is what is called broker. This is really the server that moves the data around from the producers to the consumers. And we do this at a…
We do a few things really well, would say. First of all, reliability. Most of our customers are super mission critical. So we talk about customers who run literally national energy grids. So this means if there would be a problem, then there is a huge problem for a lot of consumers in North America. We talk about almost all of the car manufacturers in Europe.
David (18:14)
Yeah.
Dominik (18:23)
We talk about the same as in Asia, for many in Asia. We talk about also in the US, many car manufacturers. We talk about some tier one car manufacturers who connect all of the factories with our software. We talk about a connected fleet of the biggest trucks brands in the world, almost all of them that are connected with our software. So this is mission critical, really. The business relies on what we do.
And so we have a huge responsibility here. And we have some case studies. An example, Mercedes-Benz is a customer with us for more than 10 years now. We have a public case study that people can also see online where they say, okay, they didn’t have a single downtime for more than 10 years while still going through three iterations of our software. So I think reliability and also high availability is something that we got right. This is pretty much why people come to HiveMQ. But then also you have a lot of
other functionality on top beyond moving the data. So first of all, scale. As an example, mean, of course, factory scale. Factory scale is compared to some of the use cases we have pretty small, to be frank. So we have a customer in Germany. It’s a connected car platform. They have 30 million devices, like actual devices online on the same installation. We talk about a topic namespace for the people who are familiar with MQTT.
this would be a topic namespace of almost 400 million tags. So this is a single installation. So this is one of the larger ones we have, but just to bring us into perspective. So scale is something every time people talk about scale, don’t forget about scale. Like we don’t talk about scale. Like we run the largest IoT installation in the world. We’re moving gigabytes of data around every single second. And this is for factories, but also for the platforms. So this is, I think this is what we got really right.
David (19:56)
Wow.
Dominik (20:21)
And then also integrations to the other systems. So this is the data broker. And then what we added a few years ago is an open source product. So it’s free to use, it’s free of charge. Also there is a commercial supported version, but every user can just use it and download it and use it. Something called HiveMQ Edge, which is a software gateway that allows you to connect to all the proprietary protocols out there, including OPC UA, Modbus, and BACnet and what you have here.
And this is also free and open source for you. So you can get all of the proprietary data from your PLCs into a standardized, let’s say, data backbone, but also move the data back. So it goes bidirectional, which is also really important because a key problem most of our customers have is data silos. And there’s unfortunately some vendors out there who really enjoy the fact that data is in their silo.
And especially some PLC vendors don’t like it when the data gets into other ecosystems. And we help our customers also pretty much connect all of their different PLCs together. And this is important if you have, for example, a factory in Germany, you might have whatever Siemens here, and then you have a factory in the US, might have whatever Allen Bradley or Rockwell. And then you still can get the data together in the standardized backbone by plugging in
these different PLCs and still have some data formats. So, HiveMQ Edge helps us contextualize the data. And then lastly, what we now added, we also have other things on the data stream, Data Hub, example, does data quality, heals the data on the fly, proper permission management, it’s really the enterprise-grade security and the enterprise-grade data management that you would expect from a solution like HiveMQ that we offer. And now what we launched is a new
called HiveMQ Pulse, which is a distributed data intelligence platform. So imagine you can have all of your factories globally, all of your assets, all of your devices in a unified information model, and you can manage your unified namespace there. And you can govern and manage your unified namespace. So it’s a true UNS product that allows you to
to manage the data and the metadata, allows it to distribute compute, allows it to generate insights directly where data is being produced and not need to wait until data hits a data lake or some historian. And we do this also completely open with open standards. So this is pretty much not only the messaging backbone, this is the actual digital backbone. There are also the largest companies in all of the sectors are now
building all of their data needs on top to get real-time insights into what’s anywhere on the globe in all of their factories.
David (23:27)
There is this one typical follow-up question when you talk about brokers is where does historical data live? If you don’t mind, I think it’s also a question many people in our audience have.
Dominik (23:40)
Yeah.
Yeah. And this is a great question. So people are familiar with the unified namespace concept, which is really not a product. It’s really more like a, what’s an architecture pattern that emerged in the 2000s was a bit forgotten. And then 2020, 20 years later was rediscovered.
So the problem with unified namespaces right now is that if you build it, you only get real-time data. But in order to ask some questions, like as an example, in my last shift, what was the maximum temperature value I observed here? This might be a question that you might ask. The current state of the system is not enough to answer these questions. Usually, what you’ve done in the past is you had some kind of historian, then you had some
hopefully some dashboard space in the historian and then you hope you can actually have this data otherwise you would go to a query. You might not have access to it and so on depending on who you are. Not every OT engineer gets access to historians all the time. And if you don’t have a historian but own a data lake, good luck as an OT engineer getting access to this. This doesn’t happen in many companies. So with Pulse what we have is first of all, we integrate with the historic data sources. This means we integrate with data lakes, we integrate with historians.
But also, we realize the fact that many companies either don’t have this or have teams that need more convincing that they should be used as a data source. Most of the time, the problems customers have in digitization are not technological problems, but organizational problems. And access to systems is a big organizational problem. So what Pulse also offers is a history. You can configure this like
how you want it. It has some defaults that are really saying for most companies, but you can really go overboard and almost abuse it as a historian, which you shouldn’t be doing. We saw some people already trying to do that. Or you can also integrate it with other data sources. And so you can have the history and Pulse also preserves that history and also generates insights continuously also based on history. And moving average would be an example.
of some data, minimum, maximum. Again, this is the trivial example, but we allow, as an example of insights would be, you can ask the system questions like, okay, here’s a batch I produced. It had 70 steps of all of the machines happening here. And today my yield is lower than yesterday. What changed? So you can trade back and then you can figure out, oh,
Actually, there was a reconfiguration of some system here on step 45. And so you can pretty much trace back all of the issues that occurred. So you can get these real-time insights. And this is not AI or nothing. This is really just the information model that gives you these kind of answers. And of course, you can use AI and ML on top. But you don’t need this for that. So if you have a proper information model, you do not need AI and ML. And sometimes,
people forget that you don’t need fancy technologies to solve problems that require them.
Willem (26:59)
A sort of follow up question on that one. So with your new product, you work on data contextualization. You already mentioned the importance of data management and governance. Now also again, the importance of having a very good data model because that will solve a lot of questions without having to find out some crazy solutions. How do you bring that message to the shop floor where
that’s usually not their world. And how do you connect with the central IT organizations in those big companies?
Dominik (27:40)
love this question because it actually is the, I think it’s the crux for all digitization at the end of the day. mean, the, what many companies are in the situation is, OT is, has very different needs than IT and IT is very often depends on the companies, very often a cost center, it very sits up until the CEO sometimes into very different departments, up until the C-suite. And so there is an organizational rift in here.
And by the way, I know this is not the question, but just as a quick side note, the best pattern that I’ve seen with almost all of our customers is have this kind of bridge team. Very often they have the name of a digitalization team. They very often have titles like IIOT architect or similar, but there must be somebody responsible to bridge that gap because it won’t happen. And this is an uphill battle for many, companies, especially if it’s just like, it doesn’t have budget and all of this.
So you cannot fix an organizational problem with technology. This is, think, the very first thing that people need to understand. And once you’re committed to solve this organizationally, then you can also come back to technology. So in our case, it’s… I mean, we come from the IT side, So we know how to sell into IT. We know how to use… This is where we’ve grown up. And…
When it comes to OT, first of all, you sell very differently and the needs are very different. And so, I mean, it’s really interesting. We work with both very well and it depends a bit where this is driven. Most companies, this is our observation, in most companies, the digitization is not driven by OT. There might be some OT people involved and there might be a lot of visionaries here. And usually these are the ones who…
who have the issues and they want to get it fixed. But usually also the budget is on it here with OT. It’s very rare that you have a lot of budget to spend, which makes total sense. OT very often, at least with our customers, this is CapEx, while if you look at IT, it’s very often OPEX. And this makes an actual difference. mean, this sounds like for many technologies, like, yeah, why should I care about this? Yeah.
Companies do care. There’s a big, difference how you look at investments and who pays for what. So I think, I mean, I don’t have a good answer for your question, but we very often get brought in by this bridge team, not by, it’s very rare that an OT engineer reaches out to us and drives the digitization project. It’s not happening. It’s very rare that an IT person drives the digitization project. There must be a clear mandate and there must be a clear ownership. And if this is happening,
then we also can work with the company. If this is not happening, we don’t sell to these companies. And frankly, they’re not ready. I don’t believe anybody sells successfully to these people. If yes, please reach out to me. want them to sell out.
Willem (30:51)
No, think David and I will agree. Unless it’s part of the desire of the company to do this. I think even if you’re a very enthusiastic engineer in a plant, your focus is probably better spent on selling it internally and getting closer to IT, maybe trying to create that bridge team. I think we have a couple of patterns for that.
David (30:56)
Yeah.
Yeah.
And also, Willem,
the fact that what Dominik said, fixing the organization is always more important than coming with new tech.
Dominik (31:29)
Yeah. And if I just may add one thing, a success pattern that I’ve observed with the companies who implemented the fastest. And I just give you an example. I mean, why does it even matter? Why does it matter to digitize? We don’t do it for the sake of fancy dashboards. This is not the reason. So one of our customers, they have rolled out the majority of the more than 200 factories globally. And they say,
Willem (31:44)
What? But they look pretty!
Dominik (31:59)
before any kind of new, let’s say, use case, they rolled out. As an example, a very basic predictive maintenance use case for some compressors, as an example. It took them six months before. And now with the additional backbone of building with HiveMQ they could reduce this to four to six weeks per use case rolled out globally into every single factory. And…
And so the speed of market, the speed of iteration is incredible. And so what’s happening is the competitive advantage this company has compared to the peer is off the charts. And of course, this looks like not very significant right now, but just imagine you’re outperforming all of your peers in execution by an order of magnitude. 10 times. This is crazy.
Willem (32:42)
I think even
in the teams and on the shop floor, I mean, if you say, I would like a dashboard that shows me ABC or I want to integrate something and you come as central team even, and you say it’s going to take you a year and it’s a huge project, people will think twice about asking things from you. Whereas if you come and you say, sure, let’s sit together and in a month I’ll show you something, by the way…
you can roll it out to the weather or plant you think is useful. That changes how people will work with data. Absolutely.
Dominik (33:16)
Yeah.
And I think what needs to happen, and this is again, that one of the success, successes, that, that was sorry, what’s the ingredients for success is vocabulary and you need to establish a common vocabulary. And what this means is not only vocabulary, what it means is you need to get them on the same page. What does it mean to have a unified namespace? What is the definition? Why are we doing this? And so it also very often starts with training programs. And so what we do here at HiveMQ is
almost all our customers, when we onboard them, we also create, we have a lot of material at HiveMQ when it comes to technology, when it comes to, let’s say, the business side integrations, also for system integrators. So we do a lot of content and education for the market. And we have our, what we call university that we also use in order for purpose-built onboarding programs for customers. So they can establish vocabulary, they can certify their people, and they can actually
This is good for the people working because they have something on their CV. It looks great. You get a certificate. But also on the other side, you can also make sure like who’s actually a tenant is. Is there some departments who are really into it and who really did the work? And is there some departments who missed out and have not yet enjoyed some education on digitization? And this turns out to be really, really huge for most customers because once you establish this, you establish a common language. And this also means you can
solve problems together, especially when you are in different parts of the organization.
David (34:48)
And if it comes about, or if you talk about education, I think with the IT/OT insider, we can only agree that there is need for much, much, much, much, much more information available on the internet. Really? Yeah, in a common language, absolutely. Yeah, absolutely, absolutely. Because that’s one of the things that all vendors, you mentioned in the beginning as well, all vendors obviously say like buy our solution because we are the best.
Willem (35:01)
and a common language because there’s ideas going everywhere.
David (35:17)
But to come back to the bit more to the UNS topic, so you’ve created quite some content. If I’m company X, whatever I’m producing, and I want to start, I would say, let’s step away from technology. Have you seen some tips and tricks where you go, if you do this, this, and this, this is a good way of starting to build your data model?
Dominik (35:44)
Yeah. mean,
my recommendations, I mean, first of all, you need some, you really need somebody in organization who wants to drive things, like especially for companies who don’t have it clearly on their agenda. And then very often, there’s really two ways. The one is there is a top down decision board level, CO level or C-suite level, is as you know, digitizing and this is how we fund it. This is great. This is amazing because this then is going to happen.
Not every company has a luxury or that wish on a leadership level yet. This will come, but not everyone enjoys this luxury. The second best thing is bottoms up adoption. And the way how to do this is not demonstrate toys, but demonstrate value. And very often the people who drive bottoms up innovation, they are technologists, love technology. don’t very often they care about the business, but it’s not like they don’t get everything. They’ll say, yay, that’d be whatever.
create 5 % more efficiencies, it’s awesome. Very often people really try to build something. There’s so much great people out there who are great builders and they cannot really contribute as much as they want to the company in the setting they are in. And so very often what’s happening is they take a product like HiveMQ Edge, which is open source, free to use, everybody can download it. You don’t need to talk to sales for doing this. So you download it and then you connect your PLCs. This is really funny. We have many of our, let’s say, champions in our companies.
They have actual PLCs at home. Because these people live and breathe their craft. And so this is amazing. And so they connect it and then they’re, oh, let me just add the data and then let me put it, we also have what we call high-vehicle cloud. It’s free of charge. You can connect up to 100 devices for free. So you don’t even need to host anything. This means you have a free stack. You just route the data from the PLCs into the cloud.
David (37:13)
Yeah.
Yes, yes.
Dominik (37:38)
And then from there, you just route the data anywhere. This could be an application. This could be another open source tool. could be dashboard. And so this is how people get started. And once they understand the technology, they usually have an idea. tell colleagues and from there you get this adoption. But at some point of time, you need to demonstrate, okay, how can I bring this in and what’s the value? And then the key questions very often, what’s the very first use case?
I recommend not, let’s please not go too fancy. Why don’t you just quote unquote get real-time insights on this one alarm that otherwise you would have a downtime or something would break and then you get the real-time insights as a push notification, as a villain in a dashboard and so on. And once you actually got the funding and you can demonstrate the actual value, this is then where I recommend talk to us and then we can also help you actually structure a real project, drive the value.
We can help you and also communicate to your managers, to your leadership team, how to actually get the adoption in. It takes time, be frank, it’s not something you do in weeks. It takes time, but we are also willing to help on this. We’ve done this multiple times, very often, every single day actually. And so we know how to navigate also these kinds of things.
Willem (38:59)
Okay, now I’m getting signs that I need to ask my last question. So one final question I definitely want to ask you. David and I also get questions about, UNS left, UNS right. And given that HiveMQ is quite involved in this topic, what are some misconceptions you would like to clarify around UNS ? Because sometimes I hear people and they think it’s going to solve all my problems.
Dominik (39:30)
Yeah. It’s, yeah. And also especially if you look at the popular LinkedIn things, forums and so on. mean, a lot of, a lot of, a lot of people are just using words and just pretty much do marketing. Sometimes I think like, really, this is kind of marketing department. This used to be a technician and now pretty much there is very misinformation. it’s, yeah, sometimes it’s bit disappointing that people spread misinformation. So.
Willem (39:51)
That’s LinkedIn nowadays. It’s just like a new Facebook.
David (39:52)
Yeah.
Dominik (40:00)
I think there’s a few people who really clarify the concepts and do great content. One of them is Kudzai Manditereza who also now works at HiveMQ, but he didn’t used to work at HiveMQ when he created a lot of his content, but he now joined HiveMQ a while ago. The other one is Walker Reynolds, who also has some very good content, and he’s one of the key drivers of this. And they usually have very good content.
So this would be, I think, a starting point for everybody who wants to look at this. I think the very next step after, I mean, now I can’t do misconceptions actually on the question. The misconceptions are very often that, I first of all, UNS is not a historian. cannot, Data Lake is not a unified namespace. It cannot be. I see some people claiming Data Lake is a unified namespace. I’m like…
I mean, it’s, it’s, it’s, need to bend the definition very, very far to even remotely. Does it make sense?
David (41:03)
Yeah, it’s
almost by definition not a structured environment.
Dominik (41:09)
Yeah, and it’s
very, very tough. then there’s other people. I saw this message a while ago where people are like, oh, OPC UA is a unified names business, whatever, 2018 or 16 or so, which also clearly has some misconceptions. So I think it’s very dangerous. And the reason is there’s no standardization. And you cannot weaken the message of MQTT. It’s clear what MQTT is, it’s not. You cannot weaken the message in OPC UA why it’s in standard.
And when it comes to architecture patterns, like if a namespace, there is no standardization. And I think this is good, but this also means people need to think. People shouldn’t believe everything they read on the internet on this one, as usual. And I mean, what we take pride in really is some curated content, also from the best experts in the industry, also at hivemq.com where we educate people and in a…
And all of this stuff works without using any HiveMQ software. This is really important. We’re an open standards company. We believe we have the best products to solve these problems. We are also not for everybody. I also must admit this. HiveMQ is really for mission-critical use cases. And it’s a commercial software. But we still want everybody really to drive the adoption, if there are customers or not. so HiveMQ.com is one of the resources that we take really pride of.
we believe you have the best content around this UNS concept.
David (42:39)
And I can only agree with that. I’ve used your content and I watched your content already multiple times. So thank you for that and keep on doing that. I think that’s also the mission we try to, I would say, to subscribe to at the IT/OT Insider. And I think with that, Dominik, this is again a wrap of this episode. So again, of the…
the final episode of our DataOps series where we explored how to make industrial data work for us. Thank you so much for joining us.
Dominik (43:16)
Yeah. Thank you all for the opportunity. It was really fun.
David (43:19)
and to our listeners for tuning in again. If you enjoyed the conversation, don’t forget to subscribe at itotinsider.com or go to hivemq.com for information about MQTT, UNS and more. And until we meet again, bye bye.