Infrastructure At The Edge

  |  Compiler Team   Edge Computing Linux

Compiler • • Infrastructure At The Edge | Compiler

Infrastructure At The Edge | Compiler

About the episode

Edge computing has existed far longer than many realize. However, the proliferation of 5G networks and the mass adoption of Internet of Things (IoT) devices have made it a cornerstone of modern infrastructure.

John Harris, who manages technology strategy at Panasonic Connect, explains how edge computing devices are quietly running the world around us, from retail stores to space exploration.

Compiler team Red Hat original show

Abonnieren

Subscribe here:

Listen on Apple Podcasts Listen on Spotify Subscribe via RSS Feed

Transkript

I feel a little silly talking about it, but it's a great example. That's old hardware, old architecture, Emily. But they updated it. It's not going to stay the same forever. They can come up with more efficient algorithms for how to get humans to another floor in a building. It seems so obvious. This is Compiler, an original podcast from Red Hat. I'm Emily Bock. And I'm Jennifer Scalf. On this show, we go beyond the buzzwords and jargon and simplify tech topics. This season, we're covering the fundamentals of IT infrastructure. On today's episode, a relatively new kid on the block, edge computing. Edge computing has been around for a lot longer than most people realize, but it wasn't until the age of 5G and the rise of IoT devices where it really took off as a part of infrastructure. And that's why we can't blame people for thinking it's new. It is, I mean, compared to most other parts of IT. I think back over the many experiences I've had in the past, say, eight years with Internet of Things, and even in that time it's gone from point of sales... That's really the main one that I would've thought about before eight years ago, and that goes all the way back, goodness gracious, I think at least 20 some years. I had a friend a long time ago and I really hope he's listening right now because he introduced me to the idea of Internet of Things before it was out there because he put Linux in point of sales systems 20-some years ago. I don't think most people knew about that. Didn't even know Linux even existed. Our cell phones certainly weren't running it. And then I only really got into it and understood what was in those devices because we really don't think of them very often. They sort of just work. Our toaster, our microwave, our cell phone, our point of sales. I mean, it feels like it wasn't all that long ago that screens in planes or in cars were not all that common, and now good luck finding a new one without them. Oh, that's a great example. And even when you go into... Recently, I went into a mall and there's screens all over the mall that you can constantly tap on, and now they'll even do things if you want them to, like connect to your phones and process information. But really what was going on in my head for at least 15 years before I got into the edge computing that I'm in now was literally the point of sales in grocery stores. Blessed self-checkout. Yes. There's so much in a grocery store. I wasn't even thinking about self-checkout yet, but that's a great example. Like the little ad screens and stuff. Oh, I guess gas stations, they're there now. Try getting away from screens. Literally when you're checking out, and even if a human is tapping on a screen, that is an edge device right there. That is processing information and sending different bits back up, which will then possibly or not connect to you and then show you different ads on your phone. So when you're wondering how the phones know that you're into something, it's because there are a heck of a lot of Internet of Things devices out there right now processing extremely quickly at the edge information about you and then feeding you ads. They're all around you. I think that's a great segue into... We've talked a bit about our experiences with them, but let's define for our listeners, what exactly is an edge device? We spoke with someone who has a lot of experience with them. Selfishly, from my perspective, I think that's where the edge is, is where it's touching with a human and it's not wired to the LAN. John Harris manages technology strategy at Panasonic Connect. He's the lead for TOUGHBOOK, a line of laptops designed for people working in field environments. His idea of what the edge means is interesting though. Jennifer, what's your definition of an edge device? Well, now it's very, very different than it was 15-plus years ago. For me personally, literally in my career, I went from helping folks with those devices that were in retail, to the devices that are out in the world on a top of a mountain helping us make phone calls, telco edge computing. Yes. Again, we drive around, we walk around, maybe we hike up into the mountains, we don't think about it, but there are these little tiny computing devices everywhere now that help us make and connect our phones. And they do all this computing right there in these little tiny, very specialized hardware devices out there. And I'm saying devices instead of switches and gateways and routers, et cetera, because it really doesn't matter. All those terms get smushed together now. That's true. We talked a lot about screens, but it's not just screens. No, not at all. It's every single thing that connects the internet and phones together and anywhere that you would have processing happening closer to humans. Yeah. And I think I like what he said there too. It's where it meets the kind of human threshold. It's kind of like where the network stops and the human begins, that's the edge, which is exactly what that means. And to contrast that, I worked for many, many years in data centers, massive data centers. We talked about that a bit on the last podcast. And you walk in and it's just a big humming, overwhelming amount of cables and blinking lights and- The jungle. Yes. So that's more of the central core. No, in this situation we're definitely talking about the little teeny, tiny devices that are out there, much, much closer to each individual human. Yeah. And I think whether it's moving compute closer to a source or an essential activity or the humans themselves, I imagine the QA testing for a rugged device is, well, rugged, robust, vigorous, let's say. Yes, very vigorous testing. These are devices that have to deal with environments that, well, humans might not enjoy, maybe an owl would enjoy them. Not even kidding. I know folks that go up and climb up those large phone towers. It's a unique workplace hazard. Yeah. Very much. Yeah. They're rugged, I think, because they better be. We externally validate the hardware to a bunch of industry standards, specifically around drop, shock, vibration, temperature, water, and dust. So if you walk around our facility in Osaka, you'll find vibration tables, vibration in humidity, vibration at high temperature, a bunch of chambers that will go down to minus 30 to plus 60, dust chambers, water chambers, all that sort of stuff. We asked John about the challenges and edge device addresses. For Panasonic's TOUGHBOOK, it's kind of in the name. Their customers need hardware that can withstand harsh conditions. We serve multiple industries. We do a bunch of stuff in public safety, fire, EMS, field service, utilities. Ultimately, the use case is pretty much very, very similar. You've got somebody who's outside in the elements. So it might be raining, it might be hot, it might be dusty. So they require a device that can be used outside. Makes sense. Yeah, I think that makes sense. John goes over a lot of different use cases, but it's important also to note that not all edge devices do the same things. There's some necessary customization there. For instance, somebody wants a barcode scanner. Not everybody wants a barcode scanner because it's a very, very specific technology. So what we've started to do is we've started to make some of the things that are that sort of 20% modular. And not modular from the point of view that the end user can swap it out, but at least the IT department can swap it out. So we do a bunch of modules for most of our range now, where if they need a smart card reader, there's a module. They need a DVD drive, there's a module, they want a second SSD, there's a module. So that gives us a little bit of flexibility with regards to being able to satisfy the demand in the market. These TOUGHBOOKS sound amazing. Right? I love the idea of being able to swap out all those different components. I'm thinking of how many times we've had laptops ourselves that we can't do anything with them. They're just one solid device. You have no customization over it. And the fact that they give folks an opportunity to swap things in and out, depending on what they need that will save them time, money. And they don't have to have whatever that extra that laptops come with being able to withstand all these crazy environments. They just have what they need. Yeah. He talked about customization there, not just in the laptop itself, but also in the conditions that it can withstand. There's such a thing as mutually exclusive customization too, which I think that covers. You're often in a forest or a mining site or somewhere with really conditions that are not very fun to live in, bad weather. We've thrown out a bunch of examples there. We heard John talk about DVD drives and we know how and when those devices are often deployed. In situations where remote access isn't possible or just isn't enough to ensure operational success, it almost sounds like edge computing exists in defiance to the modern era. Yes. And that's kind of what I was getting at too, that in the modern era everything seems very simple and put together and you buy your laptop and it's one set device. Now they've got these ways that they can swap them out and test in unique situations. And I can't wait. I'm at the edge of my seat to talk more about the technology that then drives my... I can't help myself because that's where I sit right now within Red Hat. Exactly. At some point we have to talk about what the edge devices can actually do as well versus what they can withstand. When we're talking about in defiance of the modern era, in order to withstand conditions like that, you don't want to include all the bells and whistles that you don't need because that's just more things you have to test and make withstand those conditions. So you want it to be really, really slim down to do exactly what it needs to do and that's kind of it. Exactly. And fortunately, and again, I'm at the edge of my seat. The edge. I know. I can't stop saying the word edge. And I don't know how many times we make jokes about like, "How's the edge? Where's the edge?" There's literally nonstop puns in my world about the edge. Oh, yeah. So I will throw a few more in as we go because we just can't help ourselves. Oh, I never can. Emily is very good at puns. That's my entire humor. So I hope that's okay with everybody. I'm internally a dad. Emily, we haven't said it yet. We're on the edge. I'm on the edge. I am on the edge of my seat. Because it's not just the hardware that you slim down to exactly what you need because we mentioned it in the last episode, but you have to test all of these phases. Not just the hardware up on the side of the mountain where we were hiking, where we climbed up the tower. You don't just test that hardware that's out there. You have to make sure that whatever software you're running can also withstand those conditions because if a little piece breaks on it, can it reroute? Is it load balanced? Da, da, da, da, da. So I want to throw out that question. Is edge computing a modern technology philosophically speaking? And I would argue it is not. That it has been around a very long time, but called the different names. A friend of mine was literally putting Red Hat Linux in point of sale devices, now I did the math, 26 years ago. Oh, man. So for those of us in the industry a very long time, we can think of things, specialized versions of Linux, for example, from a long time ago that now, looking back, we say, "Oh, that was customized for a tiny device much closer to the users." And a lot of that stuff, I mean, it just runs forever. But now we're contradicting our last episode where we're like, "Hey, guys, update your hardware and your software." I'm playing with that. I'm playing with that right now. How do we talk about that, yeah, a lot of these things, once they're placed, you don't move them. Yeah. If you put it at the top of the mountain, how often are you really going to want to go all the way to the top of the mountain to update or replace it or whatever you need to do? So part of that slimming down is also slimming down the maintenance aspect. I think there's two angles kind of to it when we're talking about modern technology, because due to its nature being far away, hard to access, hard to update, it does have to a little bit exist outside of time and slimmed down from the top of the line that you typically will see from quote/unquote modern technology. But I would pose it is modern technology that allows edge devices to exist, how they do now with advanced testing and advanced resiliency that we can offer it, but it also can modernize technology that isn't yet. So while in and of itself maybe isn't what you would think of as modern technology, but it does enable it and it does require it. The other aspect of that, is there a perfect edge device? Our guest weighs in after the break. We left off with a question, is there such a thing as a perfect edge device or one that can handle every situation? Are we the perfect edge device? No, because our device serves a specific purpose and there might be instances where our device is completely inappropriate. So you might have an industrial process that's running on a PLC and it might just need a very, very small black box that plugs in and it's got LAN and a little bit of compute, little bit storage, but the rest of it is being done somewhere else. Then there's the cost element as well. So no, there isn't a perfect edge device. I think that makes a lot of sense. We talked a little bit before about mutually exclusive customization. I think of it as a difference between a spaceship and a submarine. There's overlap in the things they need to do, like be airtight and keep pressure, but you cannot put a spaceship in the ocean and expect good things to happen. Oh, I love that analogy. And then I'm going to flip it because I always am a glass-half-full kind of person. There is a perfect edge device for a particular workflow, workload for a particular need, or at least we're trying to strive for the perfect device. So the hardware and the software. In my industry right now in telco, we want you to be able to make the fastest phone call. We want your internet to be the fastest internet on the planet. So we are going to constantly push as hard as we can on the hardware and the software to make you that perfect device, to make your service provider, your carrier, the perfect device. Is it always going to be? No. And are they all going to be customizable and not exchangeable? I was about to say... not exchangeable. That's why there is no perfect device, but I want to do a little bit of a glass half full on there. No, I think that's exactly right. There is no perfect one-size-fits-all edge device, but there can be one perfectly adapted to your need. All right. How does edge computing relate to infrastructure, besides being a physical part of a system? I think the answer lies in the older technology that we still use. The train line that I use to go to the office, they've got trains that are older than me. So from a technology point of view, I can guarantee you that that's not got microprocessors or anything. You can see it on the user interface. But then if you look at trains that are probably 30 year old that might be PLC-enabled, they'll have a serial port. So the engineer needs to be able to plug a serial port in it to be able to update the firmware, download statistical information about the equipment that's been used in the device. They might be pulling out runtime information from a motor, how many hours does it run? They might be looking at efficiency statistics from that particular motor, and that's all [inaudible 00:16:53]. Yep. Edge devices can be used to modernize older technology. I love that he brought up trains. We all get on a path, a metro, a train. Hopefully we have that available to us close to us and it is amazing what they do with that older hardware and much older technology. Yeah. And he brought up something that I hadn't considered too. There's the add-an-edge device to modernize it, make it smart or connect it to the internet, whatever, but there's also the collect data aspect, and that's something I hadn't thought about in the edge device world either. Oh, Emily, I have to interrupt for just a second. We are also on the verge. We just had one of the most exciting experiences in human history, Artemis. Artemis is a great example. The shuttles, the previous devices... Devices, oh my gosh, ships. Oh, yeah. The original one ran with the computing capability of a Nokia phone, right? Yes. We've come a long way. I love this story so much. It gives me absolute chills. So yes, we're talking about trains and that is extremely important. We're talking about spaceships going underwater and we're talking about spaceships that just looped us around the moon. Exactly. And I think John here has another helpful example too from early in his career. When I left school at 16 years old, I went into the elevator business. And that's a really, really similar business because things that they're selling today are all microprocessor-based, but things that they were selling 30 years ago, all relays and switches, because technology has changed. Technology has become more efficient and more usable and the function has become more... Function and features have changed. So elevators used to just go up and down, and now they do things like finding the closest elevator to that particular floor or they might like to run up or run down, depending on the efficiency of the motor. Or it might decide that it's going to serve the top floor quicker because that's where the C-suite is, rather than serving the fifth floor where the canteen is. So these things change as time goes on. And just because that piece of technology's changed, doesn't necessarily mean that there's no need to service it anymore because it's expensive and our world is complicated. There's equipment out in the world that's been there for 50, 60 years. If it's not broken, why replace it? Absolutely. This is such a common situation, especially in industries that are perfect use cases for the edge, like we've just mentioned trains and spaceships. I still have chills. I'm still so excited for the world that we were able to do that very recently on the Artemis. You have all this older machinery, that's newer machinery, but not that new. They have to test that. I mean, oh my gosh, of all conditions in the world, you guys, we really should be talking about Artemis. It is hard to get harsher environments than the vacuum of space, yes. Exactly. Oh my gosh, it's so exciting. But there's so much older machinery also and we really don't think about it. We walk down the street, drive down the street, get in the train. You're looking at machinery that has been running for so, so, so long. It very much is, if it is not broken, why are we going to replace that? It's expensive. Exactly. And you can enhance it without having to replace the entire thing. And I think that has a little bit of an eye towards sustainability and efficiency as well. I like the elevator example of you don't have to rip out and rebuild whole elevator systems. You can just add some intelligence to them and then use the infrastructure, pardon the pun, that's already there. And I have to say, it's so fun. We did that in one of our buildings recently where there used to be... It's a great example. I have to give this great example really quick because it is an elevator example, where you used to get into the elevator and press the button inside the elevator for what floor you want to go to. And John alluded to this. You used to press the button and it would take you and that's fine. And now you have to stand outside and press the screen for where you want to go, so that it does exactly what he was talking about. It takes you faster to whatever floor. It's like load-balancing humans and which floor they're going to go to. This is about 10-plus years old now that it's done this. I feel a little silly talking about it, but it's a great example. That's old hardware, old architecture, Emily, but they updated it. It's not going to stay the same forever. They can come up with more efficient algorithms for how to get humans to another floor in a building. It seems so obvious. Red Hat tower literally has that. If you've ever been stuck there during a fire drill or something, and then everybody floods out and stands in the street and then everybody has to go back up the elevators, having that kind of efficient system in place means you get back a lot faster than if everyone there is individually pushing the buttons and you have one or two people in each car. It's a way to unlock a lot of value from existing infrastructure that it doesn't already have without rebuilding it from scratch. I really wish they would do that to all elevators because now when I get to one of those elevators that doesn't have that, I'm equally as confused. Yeah. It's so true. It's like trying to use a hotel shower. It's like they're all different and you never know what's going on. Elevators are the same situation. It's like, "Wait, where's the button on this one?" Things you don't think about every day. It's so true. It's so true. Okay. So I want to go back and recap what we talked about today, because we talked a lot about edge devices. So first, edge computing is a growing part of IT infrastructure, but it's definitely a specialized aspect. We talked about it's not one-size-fits-all, but it is extremely customizable. They cater to every scenario. Edge infrastructure speaks to the exact environments where it's deployed. And edge computing is a standalone technology and it teaches us that not every technology is easily interchanged or replaced. Sometimes the shiny new thing is there to make the old thing do its job better. Exactly. So now we'll turn it back over to all of our listeners to hit us up on social media at Red Hat and use the hashtag Compiler Podcast. And that does it for this episode of Compiler. This episode was written by Kim Huang. Thank you to our guest, John Harris. Compiler is produced by the team at Red Hat, with technical support from Molly Brock. If you like today's episode, follow and review our show on your platform of choice. Until next time.

More about edge computing

Edge Computing geht über das traditionelle Rechenzentrum hinaus und bringt Services näher an den Ort, an dem sie benötigt werden. Es bietet Vorteile im Hinblick auf Latenz, Echtzeit-Performance und das gesamte Kundenerlebnis.

Blogbeitrag

Neuerungen in Red Hat Enterprise Linux 9 für Edge Computing

18. Mai 2022
Edge Computing
Blogbeitrag

Industrie 4.0 und OT-Transformation: Technologien, Vorteile und Herausforderungen

23. Februar 2022
Edge Computing

About the show

Compiler

Do you want to stay on top of tech, but find you’re short on time? Compiler presents perspectives, topics, and insights from the industry—free from jargon and judgment. We want to discover where technology is headed beyond the headlines, and create a place for new IT professionals to learn, grow, and thrive. If you are enjoying the show, let us know, and use #CompilerPodcast to share our episodes.