Episode 268’s Sponsor: Silicon Labs
Silicon Labs, a leader in secure, intelligent wireless technology has launched their 2023 Tech Talk schedule. This year’s Tech Talks include a dedicated technology series for Matter, Wi-Fi, Bluetooth, and LPWAN in order to help you build the development skills needed to deliver cutting edge IoT products. Join Silicon Labs experts, industry leaders for these one-hour, live virtual trainings created for developers by developers. Accelerate your device development today by registering at silabs.com.
What is Edge AI? And why is everyone talking about it? David Purón, CEO and co-founder of Barbara, joins Ryan Chacon on the IoT For All Podcast to discuss Edge AI and edge computing in IoT. They explore Edge AI use cases, scaling edge computing projects, the challenges of Edge AI and edge computing solutions, Barbara’s role in edge computing, and the future of Edge AI and edge computing.
David Purón is CEO and co-founder of Barbara. He is a seasoned engineer with 20+ years of experience in management and executive positions and considered one of the top-level entrepreneurs in Spain. In 2022, David was chosen as one of the top 100 creative entrepreneurs by Forbes Spain. He started his career as a software developer and international standards delegate in the leading Spanish telco carrier Telefonica. In 2009, he moved to the device manufacturer side, working in big companies – such as Huawei – and startups – such as Geeksphone. Since 2016, he’s steered Barbara, the Cybersecure Industrial Edge Platform designed to connect, deploy, and scale artificial intelligence and industrial automation applications in thousands of distributed Edge Nodes.
Interested in connecting with David? Reach out on LinkedIn!
Barbara is the Cybersecure Industrial Edge Platform, designed to implement automated decision-making within critical industrial processes. It connects, deploys, and orchestrates AI and ML based applications at the Edge. Barbara’s architecture, distributed across thousands of computing nodes, allows companies to communicate and virtualize any industrial element and then operate it through real-time artificial intelligence applications. Developed with cybersecurity by design, Barbara is compatible with Edge applications to optimize industrial processes and asset management. It is the perfect enabler to address the industry’s biggest challenges and accelerate industrial intelligence. Barbara’s Industrial Edge Platform is a powerful tool that can help organizations simplify and accelerate their Edge App deployments, easily building, orchestrating, and maintaining container-based or native applications across thousands of distributed edge nodes.
Key Questions and Topics from this Episode:
(02:13) What is Edge AI?
(05:25) Edge AI use cases
(15:37) How Barbara uses edge computing
(19:27) Learn more and follow up
– [Ryan] Hello everyone, and welcome to another episode of the IoT For All podcast, I’m Ryan Chacon, and on today’s episode you’re going to learn why everybody is talking about Edge AI. With me today is David Puron the CEO and co-founder of Barbara. They are a company that is the cyber secure industrial edge platform. Great conversation with David. I think we will get a lot of value out of this. Please feel free to subscribe to our channel if you have not done so already, like this video and hit that bell icon to get the latest episodes as soon as they are out. Alright, before we get into it we have a quick word from our sponsor. Silicon Labs, a leader in secure intelligent wireless technology has launched their 2023 tech talk schedule. This year’s tech talks include dedicated technology series for matter, wifi, Bluetooth, and LPYN in order to help you build the development skills needed to deliver cutting edge IoT products. Join Silicon Labs experts, industry leaders for these one hour live virtual trainings created for developers, by developers. Accelerate your device development today by registering at silabs.com. That’s the letter s, the letter i l a b s .com. Welcome David to the IoT For All podcast. Thanks for being here this week.
– [David] Yeah, thank you. Thanks to you.
– [Ryan] Absolutely. Yeah, it’s great to have you. Exciting conversation I know we have planned here today, but I wanted to kick this off by having you give a quick introduction about yourself and the company to our audience.
– Sure, yeah. My name is David Puron. I am a co-founder and CEO of Barbara. We’ll have a lot of time to discuss what we do, but in nutshell, we provide an industrial cyber secure edge computing platform. About myself, for a lack of a better word, I consider myself a serial entrepreneur. This is my fifth company. My previous company, Black Phone. We were dedicated to selling private and secure mobile phones, and we had an exit. We sold the company to a US firm back in 2016 and after that I founded Barbara and I’ve been in Barbara for my last five years.
– [Ryan] Fantastic. Yeah, I’ve come across Barbara a number of times in my time in IoT over the last few years so it’s great to finally have the opportunity to speak. I know our conversation today we wanted to kind of talk about edge computing, edge AI and kind of topics around that which we haven’t covered a lot lately, especially on the AI side. So let’s kick this off by just telling our audience, what is Edge AI and why is it becoming such a popular topic right now?
– [David] Well, edge computing no surprise, sorry edge AI no surprise is the combination of edge computing and artificial intelligence which are two terms that by separate they are, you know they are buzz words and everybody’s talking about them but when you combine them together, it makes a lot of sense. Right. Let’s start defining what is edge computing. I guess everybody more or less knows that edge computing is providing data storage and data computation very close to where the data is being produced. In this case very close to the IoT devices. So the IoT devices, traditionally has been connected to the cloud to make computing. Now, for reasons that I will give you later they are being connected to services that are closer, to those devices. This is edge computing, right? So these devices are generating a lot of data, temperature, position data, whatever. And then this data, when you when you process this data with artificial intelligence you can basically change the whole paradigm of the productivity of a company. So when you have a lot of data and you process it with artificial intelligence meaning algorithms that are simulated in the human brain these two things combine it together edge and AI is revolutionizing the productivity of many companies. Why the moment is now? Well, because of three factors. First, because there are 13 billion of IoT devices. So now we have a lot of data, and if you when you have a lot of data using AI, you can get a lot for a lot of insights, a lot of predictions. So first you have a lot of data. Second, the AI technologies are now democratized. So almost everybody, quote unquote can can program an an AI algorithm today. So it’s becoming easier and easier. And also in terms of the edge computing, the third factor is that the edge hardware with devices like the Jetson nano the Raspberry PIs, is becoming very powerful, right? So you have right, very small devices that can receive data can process data with AI. So this combination, a lot of data simplicity on the AI developments and the edge computing evolution is a combination that is making this world to explore today.
– [Ryan] Fantastic. Yeah, it’s been interesting just to talk about AI in general of how popular it’s becoming and what IoT is doing to really drive AI. I mean, AI needs the data. Without the data the AI models really are not as valuable. So IoT is that area to be able to collect data in ways we weren’t able to do before. And then obviously the edge computing side is becoming incredibly popular with what it’s kind of providing from a value standpoint for a lot of these solutions. So let me ask just to kind of bring it full circle for our audience here. Can you talk about some of the use cases you’re seeing when it comes to edge AI, edge computing, and I know there’s kind of been at least one of the conversations I had recently I’d love to hear your thoughts on it is as it applies to more of the industrial side of things. So where is edge computing use cases? What are some edge computing use cases say and edge AI kind of playing in a role in there as well?
– [David] Yeah, so you made a very good point. So IoT itself with AI makes a lot of sense, right? But as I was saying, in the past five, six years all the IoT AI processing has been done in the cloud. So this is already done. There is a lot of companies, financial companies or retail companies, media companies that are doing AI with IoT data in the cloud, right? What happens with the industrial companies is that cloud in many cases doesn’t fit maybe because of the latency, they need to respond very quick in some algorithms, maybe because of the fermentation of the data, but the most important thing is the privacy and the security. So industrial companies, in many cases they basically cannot put data in the cloud because of the regulations, right? So for these companies doing things at the edge where the data doesn’t need to leave their premises makes a lot of sense. So this is why edge computing and AI is becoming more and more prominent in industrial companies. So that’s that’s very true.
– [Ryan] Absolutely. Yeah. That’s fantastic. So let me ask you, if I’m a company out there listening to this and bringing edge computing into my solution or the company I’m working with is there’s an edge computing component to it. How can companies really look at that and look at for opportunities to scale it and to provide more opportunities for edge computing to be utilized in and potentially other areas or just at scale for that individual solution?
– [David] Well, the first thing that we need to discuss is the different type of, of edges, right? So everybody talks about edge computing in general but I love a picture from Garner that is a pyramid where they put the different edges, right? And starting from the top they call a local data center is edge computing, right? If you have a local data center in your city that’s edge computing then maybe a data center in your company, right? In your, in your rack, center rack it’s also it’s computing. Then if you go to small gateways that are connecting directly to the sensors, it’s edge computing but it goes even inside the sensors, right? You can do processing inside the sensor and this is the, the so-called thin edge or far edge, right? So the first thing a company needs to to understand is that there are different types of edge computing and if you are a big company like a big industrial company, what you need to do is to think in those edges as an infrastructure. Many companies are just doing a small different pilots in utilizing edge devices but the edge computing strategy needs to be wide across the whole company, right? So it needs to be the CIO or in the, or in the CSO agenda, and they need to build this different layer, edge infrastructure. And once you have the infrastructure you need to start thinking in the use cases. And for thinking in the use cases you need to think what are the use cases that are driven by either privacy or latency or scalability of the data. So if you think in one AI algorithm and and it has requirements on these three in one of these three areas, then put it on the edge and then you will save a lot of money by putting it on the edge. So this will be my recommendation. First of all, build infrastructure, and second think of the use cases.
– [Ryan] Fantastic. No, that’s great advice. What are some of the challenges that you’re seeing on this side of things? So with the edge AI, edge computing projects, when it comes to the deployment of them, are there challenges that you’re kind of seeing come up more often than not?
– [David] Yeah, I will say the the biggest challenge that we’re facing, this is not exclusive to to edge. I think it’s very related to IoT and I’m sure that everybody has identified this in many of the, of the conversation that you had and it’s about the standards and fragmentation. So every project is, is very different. I mean, you, you’re doing a project for water utility and then you go for a project for an energy utility that it could be the same but at the end of the day, there’s different devices different protocols, different hardware requirements. So at the end of the day, scaling doing edge computing as a company for Barbara it has the challenge of fermentation. We are addressing this challenge in many, many ways, but it’s definitely challenging. So the number of different protocols different devices, different staff, different applications different technology, different language everything is quite fragmented yet I guess it’s a matter of maturity as well.
– [Ryan] Yeah, I agree with you and I’ve been talking to companies about different levels of standardization in different industries. I know in the industrial space, there’s at times can feel like a lack of standardization. What, what are, how is that challenge kind of playing a role and, and how are you approaching it or what’s being done to kind of overcome it?
– [David] Well, we are, we’re trying to push for for standards and we’re participating in in industry alliances and trying to to drive the market to, to specific standards. Like for example, in industrial world we are pushing for a standard called OPCUA. So we’re trying to build as much as we can using OPCUA in terms of connectivity between devices and between edge and cloud. We are pushing for MQTT or MQT, and we are building the connectors. We are pushing for open source some of the developments. And we are also participating in some industry initiatives to try to create this alignment in the industry. And, as I said, it’s probably a matter of time.
– [Ryan] Yeah, absolutely. When, when you’re kind of working with companies right now, how are, what’s the need look like in a lot of the solutions in, I guess, industries that you’re involved in for kind of open inoperable solutions? How are, you know, how are the, how’s the landscape of IoT devices playing into that, enabling that or even maybe being a challenge in that kind of regard? What are you seeing there?
– [David] Well, we are working with, mainly with the department of the CDO, of the chief data officer. And many of these companies have gone already through a long period of sensitization and IoT device deployment. So as I said before, there are already a lot of IoT devices in many of these companies, a lot. They have sensors everywhere. They have, they have industrial equipment that is connected they have activators, they have all this equipment. Now, the challenge is this equipment is normally in what is called the OT space the operational technology space. And then you need to connect this equipment with the IT space, with the server space right? And, and these are different departments with very different cultures. Like for example, the OT people is very used to to have no connectivity at all having isolated network. And now when they have to connect the sensor to an edge computing platform or a cloud, it doesn’t matter. That interoperability is very challenging but not specifically because of the technology’s more because of the culture of the people. OT and IT integration is a big topic these days and it’s very challenging.
– [Ryan] Yeah. And one thing I wanted to ask you kind of unrelated to what we’ve just been talking about but when we talk about edge in general, there’s kind of like a spectrum of edge and where companies play a role. Where do you all fit kind of in that?
– [David] Well, as I said, if you go from the far edge, sorry, from the close edge, which is the edge that is close to the cloud. You have the data centers, and then if you go downstream then you go to the far edge which is the edge that is very close to the sensors and actuators. At Barbara, we work, especially at the far edge. Some people also call it thin edge. Why we do it, because is where the challenges happen. So one of the big challenges that we face is the cybersecurity. So if you have, the farther that the devices are the more complex it is to secure because they are normally unattended devices devices that is very complex sometimes to update or upgrade. And this is where we are specialists in the far edge and this is where we bring more value. The close edge or the big edge or the thick edge is more for players like, you know, the the typical upscalers like the AWS, Google and these people who are going down a little bit on the stack, but they won’t ever go that far in the stack because that requires a lot of expertise in firmware, in devices, CPUs IoT expertise that those companies don’t have. But the challenges like Barbara or some of our fellow competitors are experts in that. Yeah.
– [Ryan] Yeah. Are there any exciting or interesting use cases you all have going on that are worth talking about? Especially from, you know, being at the far edge?
– [David] Yeah, a lot, a lot. We’re working, you know, we are working with industrial companies mainly because as I said, these are the companies that are more prominent to use the edge. And in particular, we are working with industrial companies that have a very diverse and dispersed set of assets. We are working in, for example, in the utilities world or in the water treatment plant, for example in the water treatment plant, we’re working with a big company called Acciona here in Spain but the company is global, optimizing with edge AI the amount of chemicals used for the water treatment. This was before a very manual process. So an operator had to go and measure the water and guess the chemicals that they had to to buy for the next day. And they were really spending much more money than what they need. There were a lot of extra chemicals every day, right. And they couldn’t sell it in the market. So now with edge AI, all the sensors are calculating the amount of chemicals that that needs to be bought for the next period. And they are saving huge amount of money. Huge amount of money. Yeah. And we’d also, for example, working in utilities in electric utilities detecting flow. And for this, we are using a very unique approach which is edge mesh networking. So the different edges are talking about themselves to detect fraud patterns in the electric network and raising alarm and this is something that normally was done by by sending people out to the service stations and the people’s houses. And that was really expensive. And now this has been automated and this has been automated with artificial intelligence. So it is learning continuously is doing better and better like a human will do it. So, and also saving huge amount of money for the utility.
– [Ryan] Fantastic. Now let me ask you before I let you go here, where do you kind of see the future of edge AI, edge computing kind of from your perspective going and, you know what are you most excited about happening over the next, you know, next year or so?
– [David] Yeah. Well, so edge AI has been situated by Garner as the breakthrough technology for this year. So it’s happening now. It’s not, it won’t happen in two or three years. It’s happening now and it’s the technology with more impact and the recent years after cloud computing according garner. So it’s definitely something that is happening now. So I’m really excited about seeing how companies are industrial companies are hiding those data science those chief data offices, those profiles which are really new to them and see how these people is gonna create huge impact in the in the accounting of those companies. So I’m excited in the next one or two years to see use cases as the one that I was describing where there is a huge impact on what we’ve done. So we’re not anymore in the year of pilots we’re now in the of projects and of enhancing productivity in the companies.
– [Ryan] Yeah. That’s fantastic. Yeah, it sounds like a lot of exciting stuff’s going on in the space. I mean, edge computing has definitely grown a ton. People are now starting to really see the benefits of everything. And as you bring in the AI component it’s fantastic to just think about the value that can be provided for sure. For our audience who’s listening to this and wants to follow up, learn more about Barbara more about what you have going on, follow up with any questions what’s the best way they can do that?
– [David] Well, they can contact through our website directly and yeah, we are very happy to create an ecosystem of of people willing to do things on edge AI. So any company providing AI algorithms, system integrators, developers, anyone who is interested in the subject just go ahead and go to our website, www.barbaraiot.com. They can contact us there.
– [Ryan] Fantastic. Well, David, thank you so much for taking the time to be on the podcast. Really appreciate it and excited to get this out to our audience.
– [David] Yeah, thank you Ryan. Happy to be here and hope this has been being interesting for everybody. Thanks a lot.