KubeCon + CloudNativeCon Europe 2024 | Vikas Maurya & Anders Trosten

Clips
News
More from KubeCon + CloudNativeCon Europe 2024

Vikas Maurya

Global Product Line Manager

ABB

Anders Trosten

Software Architect

ABB

ABB brings cloud-powered automation and AI integration to the edge

Almost everything is automated today, with computer systems performing a lot of heavy lifting required to keep industries across the globe running.Technologies such as artificial intelligence are especially attractive to companies globally, showcasing the ability of data automation and large language models. Even with its perks and popularity, delivering automation across industries can be a constant challenge for companies, such as ABB Group, recently in the news for acquiring AI-powered robot navigation company Sevensense, who continue to innovate automation every day

play_circle_outline ABB as an industrial automation company serving process automation industries

play_circle_outline ABB Ability Edgenius as a platform for enabling innovation at the edge

play_circle_outline Role of open source software in enabling innovation in industrial OT

play_circle_outline Advantages of AI at the edge for predictive maintenance and optimization

Info

Vikas Maurya & Anders Trosten | ABB

Vikas Maurya

Global Product Line Manager ABB

Anders Trosten

Software Architect ABB

>> Good afternoon cloud community and welcome back to spectacular Paris. We are so thrilled to be broadcasting live here from KubeCon Cloudnativecon. Rob Strechay, my co-host. My gosh, what a magical day.

>> This is awesome. I mean, it's been so great. The energy has been wonderful.

>> It's buzzing here.

>> It's totally buzzing. And I think that, you know, AI and everything and automation and how you're going to simplify operations and I think this is a great way.

>> What an interesting lead-in you just said there for me.

>> Just segue into our great guests here.

>> We have two fabulous guests from ABB. We have Anders and Vikas. Vikas, thank you so much for being here. How you guys doing?

>> We are good. Awesome.

>> Glad to be here. Enjoying here.

>> Yeah, I can feel your energy. It's buzzing off the floor. I can imagine. Is this a big show for you?

>> Yes. I would say it is a quite a big event. We are here more or less the first time, so that's also very exciting. First time.

>> In Kubecon Paris, yes.

>> Oh my gosh. Wow, that's exciting.

>> Yeah.

>> Well welcome. We're happy you're here. And you came all the way from India.

>> Sorry.

>> You came all the way from India.

>> Yes.

>> Yeah, wow. It's quite a journey. So talking of the conversation process automation, give us a rundown on ABB, just in case the audience doesn't know what's going on here.

>> Okay. So we are ABB. So we are an industrial automation company and we particularly serve into process automation area. So, which is basically industries like oil and gas, like power generation, like chem and rev, pulp and paper mining. These are the typical industries we serve from ABB and we provide automation. So we have a technology which we call DCS, distributor control system. And that is what is kind of core of all the automation solutions, digital solutions we provide to all these companies.

>> And how does that compare to the world of IT automation and what are the similarities or differences?

>> Yeah, so IT and OT, so we are dealing with operation technology, right? And IT is something which is a little bit creeping into it, but the challenges we face in OT is really different. I would say it's not similar to IT. We have automation systems which serves to the critical infrastructure of our customers and their reliability, safety and and efficiency have been the key of it, right? So been always been ensuring that is there 24/7, reliable operation which continue to be there. We cannot stop the plant. it need to run 24/7 and reliably. That is very, very important. That's what we have offered so far. IT, we do not see the similar kind of pattern. We also deal into kind of an environment which is more, I would say resource constraint. It is also in a harsh environment. That means we deploy our solutions and systems into conditions which is environmentally not very, I would say conducive. So we have high temperatures sometimes, and we build PCs, which is like industrial grade PCs. So you don't get similar kind of resources you typically see in IT for deploying our software stack or things like this. And also we have regulatory compliances, I would say, which is we have to follow because we are serving to critical infrastructure. And this is, I mean, as I said, like energy industries, we pick like water or we pick oil and gas. They have regulatory compliances, which we have to ensure our systems follows. In IT, you'll not see it more or less in the same way. So that's something we have to really operate and ensure this is not.

>> Yeah, no, I mean I think it's funny because I started my career in IT a long time ago and I actually, one of my first places I was, was actually I worked on an oil refinery. So to your point about, it's funny, that never went down. You could not have it go down unless there was big plant maintenance and then you're losing money and all of that. And like you said, critical infrastructure for this country was the country of Trinidad's oil refinery. That was how their entire economy operated. So, and are those the main kind of, you were saying critical infrastructure and things of that nature, so water power, things like that?

>> Yes. That's the kind of critical infrastructure we're talking about. Of course we have other industry verticals as well, but I say not all are same, but there's the criticality in all the verticals for sure.

>> But like transportation and stuff like that as well.

>> Robotics, a lot of the work that you're doing is at the edge, right? And there are some, I mean you talked about these harsh environments. How do you approach that, Anders? How do you help your customers solve that problem?

>> Yeah, well I mean one of the things that are different also with like IT and OT. In IT you can kind of undo, rewind, redo, right? In OT you have changed like the real world. Like you change the physical environment and.

>> Right.

>> There's no undo in that world, right? So that is kind of.

>> There's no undo yeah, no command Z on that. Yeah. Control Z, yeah.

>> So the way to approach this is really that we try to bit separate like this more kind of edge computing environment from the core OT so that we can handle more IT like things and still interact with the OT to make a separation here. And that's kind of the, let's say the idea on how to bring this in here.

>> Yeah, yeah. How is it challenging to, I guess, okay, let me ask a better question. When customers come to you, 'cause you're dealing with legacy brands a lot of the time, right? Older and more deep entrenched industry.

>> Yep.

>> Do they know what problem they're solving for? Or do they just want things to be more efficient?

>> Yeah, so maybe I can.

>> Yeah, go for it.

>> So as I said, the foundational requirements which every customer have or which we serve is reliability, safety, and efficiency. But we are seeing the industry, the space is also a little bit dynamic. And also becoming demanding because we are looking about sustainability. They want solutions which can also optimize their plant, can run more energy efficient. And I think this is where we are getting some new kind of requirements. We are expected to provide kind of solutions which can innovate fast, right? And that's where we are trying to kind of get it through, trying to get something which can help for our future systems to enable that kind of possibilities.

>> Yeah.

>> Right, so that's where I would say the new agile innovation kind of thing, which is more a demand coming through by ensuring, is still ensuring that the previous requirements, which we have served till date is untempered. We should not be touching that or compromising on that, but we still want this to happen. Those are kind of little bit opposing requirements.

>> I was going to say competing priorities there a bit.

>> Colliding. Yeah, yeah, yeah.

>> Yes really, right? That's what's going on.

>> So that's the tricky thing we are trying to deal with.

>> Yeah, you're almost like moderators in that space, I bet. Yeah.

>> We kind of are like the interface to the data. We are the interface to the OT and by being that, we can enable this innovation if we can get the right environment there. So that's what the edge brings to us, that we can bring in say, a modern cloud native platform where you can have access to data and access to new technology and marry this together to to to run at the edge. That's kind of our little bit mission here to bring this into the OT world.

>> Is that where the ABB ability, EdEdgenius comes in and did I pronounce that right?

>> Exactly.

>> Exactly, so that's where we are converting the OT and IT more or less in the middle. And the goal we have is to kind of keep the core process control, which provides this foundational requirements on left side, leave them within a lifecycle, release cycle, update cycles, which is much slower as it was there in the past. But then we extract the data from this DCS data, process data from this space, safely and securely to a kind of borderline where Edgenius is coming into, getting the data safely out of this and bring at the edge. And then we start the whole kind of monitoring and optimization kind of space where we bring application solutions, we can optimize the plan. So Edgenius is somewhat sitting in the middle and trying to be a platform for enabling innovation at speed. That's what we are kind of trying to achieve. We are using edge technology, cloud technology, trying to build an whole app ecosystem where we can get this apps coming seamlessly faster. And that's where the IT infrastructure is going to be the key for us to kind of build on, right. And that's where we are using open source. We are partnering with Red Hat. We're trying to get stuff into it.

>> Yeah, let's talk about the open source piece. We're here, it's the biggest open source conference in Europe, obviously matters. How does open source software help enable your innovation? Anders, I'm going to go to you on that. Yeah.

>> If you want to be innovative, we have to use latest technologies. That's where the innovation happens. And of course that happens in the open source world here. So what we are need to do is to bring that into the industrial OT world, but at the edge. And to do so, we need to bring the whole technology stack in there, including operating system like Linux operating system, Kubernetes, all of the plugins that goes with that, all of the extensions, and make that more like a embedded operating system for applications at the edge. We kind of think as Kubernetes is really the operating system in a way, because that's where our applications travel, provides all the infrastructure. So we package that up and of course open source has challenges too. Like there's lots of different developers and communities and when we ship something embedded in our product, we have to also take responsibility for that. We have a kind of warranties, we have to support our customers. So that's where Red Hat comes in as a kind of a supplier that we can say like, we don't know really how to fix box and all the plugins, that's kind of their job. Our job is to put this together and bring it into this space where we can add applications.

>> Yeah.

>> Go for it.

>> So I mean, you're using some of the specific technology from Red Hat as well, like microshift and their device edge. How is that playing a role in what you're doing as well?

>> Yeah, so we realized quite early that Kubernetes is the way to go, but that was a big thing in the cloud back then. And we need something where we can scale to where we deploy. We deploy from data center in the plant, like small servers into kind of maybe single computers and then industrial pieces in the field. And they are not very powerful. So that means like there is no one size fits all.

>> So right.

>> So to deal with that, we needed something which could be Kubernetes at the small range, but can also move up in the stack and become bigger. That's where we worked with Red Hat to use the Red Device Edge and get Micro shift as like kind of a, let's say more lightweight kind of embedded Kubernetes that we can use. Today we are done, we can build an embedded version of this. We can install from bare metal to running Kubernetes in five minutes, which is kind of awesome.

>> Whoa.

>> Five minutes. Yeah, that's good.

>> Not casual data drop there. That's impressive.

>> That's the micro shift. As I said, we have a resource constraint environment. We cannot blow up (indistinct) with the multi-node clusters.

>> That's true. >> In this space. So we needed really something which is low footprint and that's what we worked with Red Hat to get. And we have microshift Red Hat devices now. We are happy working with that and it's what we are trying to go to the market with.

>> And I would expect another piece of it is the fact that the security aspects of it and the fact that you can have it at the edge and disconnected in a disconnected manner and things.

>> And scalable depending on what happens.

>> Yeah. >> Yeah.

>> We did that option as well. We have a offering which is completely disconnected as well. That means if customer has very much concern to not have an internet connectivity reaching up to that level, then we have to operate this disconnected locally managed, and it has its own cost of course, but it's something we have to do for the kind of market we sell.

>> Yeah, that's interesting. Okay, so lots of exciting stuff happening. You're working on some of the biggest projects on the planet literally, with some of the scale that you have, which is amazing. What's next?

>> I mean we are here here, but to talk about AI, AI and. -

>> We haven't said that. I know it was coming, but yeah, keep going.

>> So I mean that just I guess proves that if we want innovation, and we have a platform like this, and the AI comes in as a way to just use it on the edge, we have GPUs that we can attach, we can have most containerized, we can have serving APIs on the edge. We have the data. So of course we put this together and we will I guess find out what comes next. I hope a lot of cool stuff.

>> Yeah.

>> Yes (laughs)

>> Sure, I mean that's the kind of next step once we get this edge out, which is a platform which allows the application to run and the application is not just like plain application. You can have this AI models running on the edge as well. And that means that we can enable use cases which can predict stuff. And we are working on one of such use cases for our control system, like augmented operator. We can assist operators, operator can have a prediction that okay, the plant is going to fail or certain equipment is going to blow in six months, three months, looking at the patterns. So those kind of use cases also, we will enable on the edge as next steps for sure.

>> Yeah. Which everybody, that's the old school ML type of AI, which again, it's not the gen AI that everybody talks about, but that seems like it has more value to your customers, especially 'cause if a plant has to go down, that's huge cost to them. So it would seem that bringing that to the edge could help massively.

>> Yeah, I think we haven't seen all the use cases yet, like what can be done with this? But it looks very promising and just kind of bringing it into the environment where we can use that without having to bother with everything else around. Like that also makes these use cases easier to implement, faster to realize.

>> Yeah.

>> Yeah. So one of the challenges which we want solve, we want to bring the app ecosystem as we said, right? And how easy to bring the application because we are a very huge company operating in more than a hundred countries. There are business units which is selling it to customers. They have ideas, they know what they can do for their customer to solve certain pain points, but they do not have the ability to convert their ideas into an application and deploy on something, right? That is what we want to empower them to, let's say you got an idea, we have an app ecosystem here. You just build your app so easily, deploy it on the edge and you are there and you deliver the value for your customer. That's it. So that's what.

>> That's also why we went Kubernetes stack. If you want to go to something like a mobile phone with apps, you can basically click, install, and then it works. We need to have a very well defined way to package application, move it somewhere and apply it and update it. That's also what we get with this cloud native world that it enables these things to happen.

>> Yeah, makes sense.

>> Yeah, yeah, it makes sense. And it's so important. I mean, and everyone's looking for that easy button. People and especially, I mean I love that we talked about edge so much. They want it where they are, not just exactly in the data center or on, you know, it's such a different.

>> The data is sensitive,

like in our industry. >> Extremely. They ship it somewhere else, it'll depend on something else. That is not really desired.

>> No, and it slows down the latency. You know, we're not going to be able to make the real time decision making, which is where you need it, right?

>> Yeah. I mean we have seen this constraints clearly the latency and security.

>> Yeah.

>> Which does not allow or let customer does not allow the customer's data to leave the plant. But then we do more stuff on the edge. But the use cases where certain data can come to the cloud, we can deploy in hybrid situation, right. To have let high latency requirement, load that low latency requirement, applications running on the edge and the things which we need to have the fleet level monitoring and all those things, which is just simple dashboards and all can run on the cloud, right? So if it's something customer by customer looking at their use case application needs, that's how we decide and try to serve our use cases.

>> Alright. Wow. We could talk about this all day. I have one final question for you. Now my mind's going a ton of different directions. What do you hope when we have your fabulous selves back at the next Kubecon when you're here for the second time, what do you hope you can say then that you can't say today? Anders, I'm going to start with you.

>> Oh God, that's a hard question. I hope you can say that this is now out in the field. Like, and here's what we are running and here's what you can do with this and here's how you can do this. If you are like an application develop, want to run on the industrial edge, here's how you do that. You want to open up this ecosystem as well for not just ourself.

>> Yeah.

>> Yeah. I mean we would like to okay, demonstrate maybe that how our app ecosystem is working on this low footprint Red Hat device edge. How we are able to kind of deploy this to our customer. It's a good story. And also show this how easy it is, right? Because that is such a key part, which is critical for us as well. I mean, it's so complex to do it and it doesn't fly, right? So we need to make it as easy as possible to bring an app on the edge. And I'm hoping to use as much of open source Kubernetes, whatever we get out of here. And to enable to our, I would say development teams or customers who can do that.

>> You know, that's really what it's all about. I mean, with Kubernetes it was all about decreasing complexity with AI, same thing. And at the edge it's even more imperative given the constraints. You were both fantastic guests. I am so grateful that you joined us and that ABB is at the show. Great partner of Red Hat. Rob, it's always a pleasure sitting on the desk with you. It doesn't even feel like work. And thank all of you from tuning in wherever you are around the world. My name's Savannah Peterson. We're here in Paris at KubeCon Cloudnativecon live from the fabulous country of France. You're watching the Cube, the leading source for enterprise tech news.

Vikas Maurya & Anders Trosten | ABB

>> This is awesome. I mean, it's been so great. The energy has been wonderful.

>> It's buzzing here.

>> It's totally buzzing. And I think that, you know, AI and everything and automation and how you're going to simplify operations and I think this is a great way.

>> What an interesting lead-in you just said there for me.

>> Just segue into our great guests here.

>> We have two fabulous guests from ABB. We have Anders and Vikas. Vikas, thank you so much for being here. How you guys doing?

>> We are good. Awesome.

>> Glad to be here. Enjoying here.

>> Yeah, I can feel your energy. It's buzzing off the floor. I can imagine. Is this a big show for you?

>> Yes. I would say it is a quite a big event. We are here more or less the first time, so that's also very exciting. First time.

>> In Kubecon Paris, yes.

>> Oh my gosh. Wow, that's exciting.

>> Yeah.

>> Well welcome. We're happy you're here. And you came all the way from India.

>> Sorry.

>> You came all the way from India.

>> Yes.

>> Yeah, wow. It's quite a journey. So talking of the conversation process automation, give us a rundown on ABB, just in case the audience doesn't know what's going on here.

>> And how does that compare to the world of IT automation and what are the similarities or differences?

>> But like transportation and stuff like that as well.

>> Right.

>> There's no undo in that world, right? So that is kind of.

>> There's no undo yeah, no command Z on that. Yeah. Control Z, yeah.

>> Yep.

>> Do they know what problem they're solving for? Or do they just want things to be more efficient?

>> Yeah, so maybe I can.

>> Yeah, go for it.

>> Yeah.

>> I was going to say competing priorities there a bit.

>> Colliding. Yeah, yeah, yeah.

>> Yes really, right? That's what's going on.

>> So that's the tricky thing we are trying to deal with.

>> Yeah, you're almost like moderators in that space, I bet. Yeah.

>> Is that where the ABB ability, EdEdgenius comes in and did I pronounce that right?

>> Exactly.

>> Yeah.

>> Go for it.

>> So I mean, you're using some of the specific technology from Red Hat as well, like microshift and their device edge. How is that playing a role in what you're doing as well?

>> So right.

>> Whoa.

>> Five minutes. Yeah, that's good.

>> Not casual data drop there. That's impressive.

>> That's the micro shift. As I said, we have a resource constraint environment. We cannot blow up (indistinct) with the multi-node clusters.

>> And I would expect another piece of it is the fact that the security aspects of it and the fact that you can have it at the edge and disconnected in a disconnected manner and things.

>> And scalable depending on what happens.

>> Yeah. >> Yeah.

>> I mean we are here here, but to talk about AI, AI and. -

>> We haven't said that. I know it was coming, but yeah, keep going.

>> Yeah.

>> Yes (laughs)

>> Yeah.

>> Yeah, makes sense.

>> The data is sensitive,

like in our industry. >> Extremely. They ship it somewhere else, it'll depend on something else. That is not really desired.

>> No, and it slows down the latency. You know, we're not going to be able to make the real time decision making, which is where you need it, right?

>> Yeah. I mean we have seen this constraints clearly the latency and security.

>> Yeah.