- Speaker #0
My name is Greg Bucco and I'm the Chief Marketing Officer for Universal Automation Network. I'm happy to have today a new session which is mixing the theme of sustainability together with the theme of IT and deploy at high scale because today we will receive in fact both members of Universal Automation with the company GRVN and the company Barbara which are members since a few years now. So today the topic is how to link the UEO technology and how to deploy it at huge scale on a typical use case. So to dive a bit further into that, I am happy to welcome David on stage. David,
- Speaker #1
hi. Hi, Greg. Hi, everybody. How are you?
- Speaker #0
Yeah, good. Thank you. Can you tell us a bit? I hope that you're doing good yourself as well. Can you tell us a bit more about your project and what you've been doing with GRFN? Yeah,
- Speaker #1
of course, of course. Yeah, I will take these next 20 minutes for explaining that and then we can save around another five to ten minutes for open questions. So if you guys in the audience have any questions, feel free to write it on the LinkedIn comments and then we will get the questions and answer. So, my name is David Perón, I'm the co-founder and chief executive officer of BarBara. BarBara is a company headquartered in Spain with operations in six countries. We joined UAO around a year ago, but in this case I'm presenting a project that we've done with another UAO member which is the company GR3N also also named as green technologies uh green is not able to to join today the presentation but we've prepared this this presentation together and i'm speaking on behalf of of the two companies so i will use some slides for this so um uh let's take this slice to to the uh to the audience and i'll start maybe by presenting the two companies for those people who who don't know Green is a company headquartered in Switzerland and they are expressing a very, very innovative process which is called microwave-assisted depolymerization, which is a way from recycling plastic, which is, first of all, much more sustainable and, second, much more efficient also. So, all in all, it's something... really, really necessary for today. On the other side, at the right, you can see Barbara, which is the company that I founded in 2016. And at Barbara, we provide infrastructure to manage and deploy applications and AI models remotely into industrial edge devices. So as you can see, these are very, very different questions, very, very different companies. So one coming from the sustainability world, the other coming from the IT, OT integration world. But both together, we have created a great use case that I'm going to present today. So Green Solution is about recycling plastic in a much sustainable way. So they are able to get any type of plastic and recycle it in a way that the result is directly usable again. with the same components and the same quality that it had in the in the beginning stage. So that allows basically infinite recycling of materials and they can do this reducing the CO2 emission up to 80 percent compared with the traditional crude oil recycling mechanisms. So sustainability which is something that we really need these these days. The challenge that Green has is that they can get into their system, into their manufacturing lines or recycling lines, different types of plastics. So they can get bottles of different materials, different types of blends and so on. And they are able to obtain high purity polymer from any material. So the thing is that they... they are working in a very very a dynamic manufacturing process. It's not a process that is always the same. It really changes depending on the type of materials that they are using. So, Green has created one recycling plant already and their requirements were very, very aligned to what Universal Automation is offering and is promoting. First of all, they needed modular engineering process, so they need to design their control logics in a very modular way. They need to deploy and develop protection rules for their reactive units. And the most important thing, they need to reduce engineering costs during the market scale-up. So for that, they have chosen the UAO runtime architecture, and they are deploying the UAO runtime in six of their process cells of the factory. So what UAO runtime is helping them is to create the logic in a vendor agnostic way. So the runtime can be deployed in any hardware and the runtime can be deployed and programmed and reprogrammed again and again. during the life cycle of the plant and the company. So this allows them to do things in a highly modular way but also in a highly scale way. So the project is not small, it's around 128 monitoring sensors in the plant, so about 500 input outputs in nine control cabinets. So they have nine control cabinets where they deploy an industrial PC and then this industrial PC have the the URL runtime that can be programmed from the IDE. So this is, let's say, their core automation architecture. Now they had a problem and they have a problem because now they are scaling to multiple plans and then they came to Barbara to see how we can help with that problem. But before that... Let me explain with numbers what is this problem. They wanted to create more applications on top of the UEO, like data visualization, predictive analytics, multiple databases like Postgre or InfluxDB. And let's say, let's do some numbers and big disclaimer here, these numbers are just references. These are not real numbers from... from the plants, but I wanted to have these numbers for everybody to really understand what are the scaling challenges that that green was having and why they came to Barbara. So let's say that they have seven devices per plant, one per area. Then let's say that each device has five applications. So maybe the UAE runtime, but maybe the analytics dashboard, the databases and everything. five applications per device. Now let's say that they want to scale this to 10 plans and actually this is the objective within the next years. So with that number that makes a total of 350 applications instances. So seven devices multiplied by five applications per device multiplies per 10 plans it makes 350 application instances now There is a very interesting statistic here that we got from the Puppets History of DevOps report, which says that any application requires an average of 10 software updates per year. And I know that this was not very common in the previous automation world, where things used to be very static, but the world today is changing. So there's more cybersecurity issues, so you need to update. Sometimes because of cybersecurity, there's a lot of competition between companies and you need to create new features continuously. So you need to be productive. So now the automation world is changing and UAO is promoting this kind of dynamic environment into the software. So let's say you want to do 10 software updates. So if you want to do 10 software updates per year over 350 application instances. that means you need to do 3,500 update events. And with a conservative cost of 100 euros per update. in engineering hours. That means that the infrastructure cost of a deployment like this that we are seeing here is around 300k euros for software maintenance. So this might be a good number to justify the investment on a project that facilitates this orchestration and maintenance of the software. But there is also an interesting study, which is that 5% of human-made updates can introduce errors, and 40% of those errors can result in a downtime or service degradation. So, in Green's example, let's say that with 3,500 updates events in a year, probably more than 50% can have a small mistake. that can result in service degradation or downtimes. So the hidden cost could be even higher than 5 million. So if you want to manage all these remote devices and all these remote applications of many devices in many plants manually, you will probably have a hidden cost which is millions. So this is just an example, but this is just to exemplify the reason why green came to barbara and asked for help because indeed what we do at barbara is to enable companies to remotely deploy run and manage any application at any edge device and we do it with the same simplicity that today you can find in in cloud environments so cloud environments are very simple to use so you can go to the cloud upload applications configure applications so we can do exactly the same but in industrial computers that are distributed in the plants. So this is the architecture of Barbara but in the sake of time I will skip this slide and go directly to what we've done in the project. So in the project we have installed our firmware in the industrial PCs which in this case for the for the first testing and the first plan that we are doing together with Green. are ASRock industrial IPCs, in particular the model IEP 5000G. So here we have installed our operating system which is called BarbaraOS. And from BarbaraOS you can manage remotely from the management panel which is here at the top, you can manage the applications that are distributed in those industrial PCs. So we have put one device per cell, per manufacturing cell, and then we have put another devices, in this case virtual devices, that have also the edge engine from where we can deploy remotely the other applications, right? So all this is managed from a centralized platform which is called Barbara Panel and I will show Barbara Panel. in a few minutes so you will be able to see it. But the thing is that using this orchestration platform, first of all, the runtime development environment, which is in this case Schneider EcoStruxure Automation Expert, this piece here, through this system it can connect remotely to any runtime no matter where it is in the world. If you have 1,000 or 10,000 industrial PCs with the runtime distributed all over the world, using the building capabilities of Barbara, the development environment, in this case EAE, can control and deploy logics into those runtimes remotely. Second, from Barbara's management system, as you will see later on when I show the system, you are able to remote manage and access to any application in the whole architecture. So you can deploy the runtime in thousands of devices. You can deploy analytic application databases in hundreds of devices just with one click. Barbara is also offering a marketplace of IToT advanced analytics applications, so we have runtimes. from different vendors that you can deploy, but you have also other applications that can be deployed next to the to the runtimes. You can deploy applications and changes into the devices using continuous integration and continuous deployment pipelines. That means that you can automate all the all the deployments and this is very related to what I was saying about the errors before so if you need to do every deployment manually end of the day you will probably have 5% of errors leading to downtimes and so on. So with the CI-CD pipelines you can automate those deployments and do it in a very automated way. Once you have tested it in the laboratory then you can deploy applications and changes in hundreds of devices using automation tools. One interesting thing that is also offered by Barbaros Orchestration Platform is the redundancy for containerized applications. So in this case we are deploying the runtime in a container and using our orchestration capabilities, you can deploy that container with the runtime or with analytics on the base in what we call a swarm, which is basically clusters of devices. and with this we are able to provide absolutely zero availability, zero downtime sorry, and high availability. And last but not least, all this deployment and all this architecture is secure by design and is following the recommendation of IEC 62443. So it's highly secure and highly compliant with the norms that are being required in this case by Green, but probably by many other... similar companies and manufacturing companies. So let me show very quickly the platform. So this is the orchestration panel that I was describing. So from here you can basically control remotely the different devices that are in the footprint, right? So when I say devices, I say I'm referring to multi-vendor architecture, including vendors of many types. So as you know, UAO's purpose is to be able to offer the same user experience to developers and infrastructure managers, no matter what is the hardware behind. So you can use hardware from Advantech, you can use hardware from Snader, you can use hardware from ASRock, from Lanner, from any industrial manufacturer. and control it from this panel, right? So once it is in this panel and once it is controllable, you can select the devices where you want to deploy those applications and I can basically deploy the application. So I can select the application, for example, let's say I want to deploy the Snaders of the Packed Runtime, which is a UAO compliant runtime. I can select the version and I can send configurations. to all the devices in the field. So if I click next here, I'm basically deploying the runtime in all these devices. And once the runtime is there, I can enter here and see the logs of the applications. I can see how they are working and so on. I can do the changes. I can automate the changes and so on and so forth. So all in all The idea is to, as Greg was explaining, to be able to scale a project and pass it from one plant to multiple plants. So this is how the green project will look in the future with a lot of devices, a lot of runtimes, a lot of applications, instances. All of them control from a single pane of glass, which is Barbara Panel, which is our orchestration platform. So. Our orchestration platform is integrated with the URL runtime so that it can deploy and configure remotely the runtime but also with other applications like analytics, dashboards, databases, logic programmers like Node-RED and many other things. So with that basically Green and Barbara We are trying to define a kind of blueprint for scale projects into multiple plans, global plans. for data-driven manufacturing. So I think we are very pioneer. So the use of UAO runtime has been E for green in order to let them deploy control logics very quickly in a vendor agnostic way and future proof for the future modifications but also the seamless integration with with Barbaras orchestration is providing the scale to make this project a reality, not only for one plant, but for multiple plants all over the world. So with that, let me stop here and save these five minutes for questions, in case there is any question, Greg.
- Speaker #0
Yeah, thank you, David, for this very interesting presentation. Yeah, indeed, you see it with your calculation how easy it is to like let's say, save at the end a lot of money due to either deployment on new scale or avoiding the human mistake or being able to react to that. Actually, that's more the case. And well, following that, I see that people have been quite active because I see a few questions coming. One question from Grant, is Barbara offering involved with expanded edge orchestration capabilities. and following the evolving Margo standards?
- Speaker #1
Yeah, we're definitely supervising the Margo standard. We're not actively participating yet, but we probably plan to participate very soon. And once the standard is out, we plan to follow it and do the adaptation of our interfaces to Margo, which are, I will say, not very different to what we do as of today. But yeah, Margo is a great initiative. because one important thing that companies need is interoperability. So many companies will be looking probably for an orchestration platform that is compatible with orchestration engines on the devices and also having different options for different vendors for the same problem. So yeah, Margo is a great initiative and we are following that and probably will be contributing very soon.
- Speaker #0
Okay, great. Great. Next two questions coming from Saad. First one, Barbara, OS is based on which Linux distribution?
- Speaker #1
Barbara, OS is a Linux distribution itself. It's built from the kernel and it's built with the IEC 6443 security requirements. So this is our core operating system and this is the one that offers us or our customers the possibility to be compliant and very lean with with severe security requirements but we can deploy on other operating systems i mean we can deploy our orchestration on top of Yocto and also other operating systems so our customers are not tied to to Barbaro OS but when it comes to provide industrial security we always recommend Barbaro OS.
- Speaker #0
Great, great. Next question from Saad. The back end of the Barbara panel, can it be hosted on premise or in the cloud?
- Speaker #1
It can be both. So in the case of green, this is hosted in the cloud. Actually, all the communications are outgoing, all are secure with single certificates and everything. But it is connected to the cloud in a secure way. But the Barbara panel can of course be... be deployed also in a private cloud or in a server, in an on-premise server, which is actually what many customers require due to the highest level of cybersecurity and their compliance. So, Barbara Panel can be hosted in the customer premises and then the communication between the nodes and the orchestrator will be always internal.
- Speaker #0
Okay, good, great. We have the next question, which is, what are the core skills required to manage and implement this type of system?
- Speaker #1
Good question. So, well, for the implementation, we normally work through system integrators, which are the ones who implement the service in the plants and everything. So that is, let's say, kind of a managed service. for managing the system as you have seen in the very very short demo that I showed, there is no need to have any software skill and this is important because many orchestration platforms they require huge software skill like command line interfaces, git repos and many things that unfortunately the industrial world is not very used to. So we've put a lot of effort on making the user experience very simple and very lean so that you don't require software skills or anything. Everything is done through graphical interfaces. And by the way, this is also accessible through an API. So if you don't want to use the graphical interface and you have that software skills, you can manage this and automate things through API. But all in all, it's quite, let's say, industrial user experience with no need of strong IT skills.
- Speaker #0
So really open and really easy to use for everyone in the industrial domain. Okay. Yeah.
- Speaker #1
Yes, as well as providing the requirements that are required by the industrial customers, like high availability, zero touch and everything.
- Speaker #0
Yeah, makes sense. Yeah. Another question from Sartre. The comms between Barbara Panel Backend and Barbara OS to work even when a firewall is in place?
- Speaker #1
Yes. So as I said, the the communications are outgoing always from the os to to the orchestration panel and they are secure and it requires to open uh ports but it requires to open outgoing ports from the ot to the it so normally the os is installed in level two level three while the will the panel can be either in cloud or or the corporate network in level four uh so the the firewall rule that you need is normally um outgoing and it's just a couple of ports open. And also the OS is prepared to communicate using VLANs, proxies and the typical complex network configurations that we find in those type of networks. But in general terms we don't find a lot of reluctance or resistance during the deployments for the communication between the OS and the panel. It's quite simple, always outgoing. and capable to connect through VLANs and proxies.
- Speaker #0
Okay, okay, good. And for me, another question, but this time, let's say more linked as well to the 649 and to the UA technology. So the runtime, is it like configured through the orchestration panel? Or do we need like IDE at all? Or how is it working?
- Speaker #1
You need the IDE. So the IDE is still there for actually programming the actual controls and routine. So what you can control and configure from the orchestration platform is the runtime configuration itself. So the networking of the runtime, the versioning of the runtime, that is the configuration of the runtime application itself. Then for deploying the the control logics and everything, you work also through the IDE of your preference.
- Speaker #0
So I select my IDE, I make my application. This application, I need to deploy it on one or thousands of IPCs. Then I use the Barbara, let's say, platform and Barbara OS to deploy that same application to the thousand IPC I have in the underground, basically.
- Speaker #1
That's correct. that's correct so um you first deploy remotely the the runtime itself i mean you install your 700 computers and then you deploy the runtime first then you go to your ide uh program the the logic and then deploy the logic to the 700 devices um using the orchestration platform no matter where they are i mean if they are in different local networks and whatever the orchestration platform is the one providing that connectivity layer and that um
- Speaker #0
that management layer yeah good great great david thank you i think we come toward the end of of of our sessions i see no no no other questions so uh oh there is maybe one question coming just now let's see if we we can take it one may keep updates to notes uh what fell safes exists if someone goes wrong with the new application? How do you recover from a change? That's a very good question. So we do have a very important feature, which is called network rollback. Because if you do the change to the network itself or to the networking, that's a critical change. Because if you lose network, you don't have orchestration again, and then you need to send people to the field. So first thing, we have the network rollback, automatic rollback feature, which is if the network is working bad, chains breaks the communication then automatically falls back to the previous version. That is with regards to the network. With regards to the application configuration itself, what we offer is high availability clusters where you can put applications not in single nodes but in clusters and what we always recommend is to do the applications in clusters and then maybe update two instances of the of the cluster and leave one with the with the old one and then once you you've done the the the testing and whatever then start rolling back the the rest but the high availability feature is very useful first of all because you have zero downtime on the on the changes but also because you have the possibility of doing these rollbacks in case you you fail but they automatic rollback is only done in the case of network failures.
- Speaker #1
Yeah, okay. Okay, good. Thank you for that. One last question before we close. Is it possible to enhance the Barbara panel to even deal with downloading and deployment of control applications and its configuration if, for example, a construction automation expert provides an headless API interface to do that?
- Speaker #0
Yes, it's possible. I know both Barbara and EAE provide API interfaces, so that type of integrations are quite common in the projects. There is not yet an integration of this type, but it can be done ad hoc, and I'm sure that eventually it will happen at the product level and will be embedded in the products. but as of today it's possible but it requires that API to API integration, but of course it can be done.
- Speaker #1
Okay, good. Great. Great. Then let's go back here. Thank you, David, for this very interesting live that we had about your technology at Barbara together on the use case of GR3N. So kind of dealing with IT huge scale deployment together with... the sustainability which is given by Geo3M. So for sure, you will come back to us when you have some new updates to share with the community. We are always happy to have you and have the Barbara team all together to share what you are doing on the market, guys. And for the viewers, thank you for watching us and please stay tuned because we'll come in the next month with a new live series. Stay tuned to see what is happening, which use cases are going on together with the universalautomation.org technology. So thank you to you all and see you soon. Bye bye.
- Speaker #0
Bye. Thank you.