Elevate the Edge
Elevate the Edge
Edge Cloud Adoption with David Linthicum
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Hosts Jo Peterson and Maribel Lopez discuss Edge Cloud adoption with David Linthicum, Chief Cloud Strategy Office, Deloitte Consulting.
Hello and welcome back to the elevate the edge Podcast. I'm Maribel Lopez and I'm joined here with my co host, Jo Peterson. Hey, Joe.
Jo Peterson:Hey Maribel.
Maribel Lopez:We're really excited as always to be here talking about the edge. And we could think of no better person to kind of tee up the differences of what is or is not happening in the edge. Then David Linthicum, he is the chief cloud strategy officer at Deloitte. He's actually many things. He's a best selling author, speaker, radio, TV, and podcast personality. So, you know, really appreciate you taking the time to be on the show, David. So welcome, David.
David Linthicum:Great to be here. And thanks for having me. You forgot bielawski I think that's the that's the big title thing that's probably more accurate description of what I am. You know,
Maribel Lopez:I was thinking about the beat list geek. And I thought, I don't know if that's quite accurate. I think he's more of an a-list geek. But you know, we will, we'll leave,
David Linthicum:I have proof every con every conference, I speak at it, they always when they call me I gave who turned you down, there's always somebody else a turned down. So in other words, the A list didn't pick up the phone. And so they get to the B list. So it's a I just, I just thought it was it was funny because it normally when they're asking me to speak somebody else's, especially for keynotes, or, you know, more visible parts of the conference, someone else's turn them down. So hey, I'm gonna be less. No biggie, but I am on the beat list.
Maribel Lopez:You know what, the important thing that matters is, no matter where you are on the list, you're on the list. So that's, that's what I'm going with.
David Linthicum:If I drop off the list, it's a bad thing.
Maribel Lopez:It's a bad thing. Well, you know, Joe, and I've been speaking with a lot of individuals about the edge. And we've been talking a lot about the difference between where we originally thought data was going to be collected and analyzed, and how that shifted, you know, there was a lot of this was very similar, you know, you have centralization and decentralization, pendulums that happen quite frequently in it and in business and in life. And the latest stuff that we were looking at was won by Gartner that said, by 2025 70% of the enterprise data was going to be created and processed outside the traditional data center and cloud. And a lot of ways that makes sense to me. But I'm really interested in hearing what your thoughts are on that statement.
David Linthicum:Yeah, I have to remember, there are many types of edge computing, you know, the mobile edge, the cloud edge, device edge, you know, things that are more IoT related or sensor related. I can't imagine now. So they may be having a description of that, which may get to 75% of enterprise generated data that's going to be processed in this way. The reality is that we're becoming a bit more flexible, and how we're moving information and processing into any place where it can be most efficient. So if we are going to move in that direction, it could be by 2025, that we relocated the processing and the data into things that are, you know, like edge computing. So in other words, there could be edge clouds, it could be mobile based systems, you know, all these sorts of things. The difficulty there, it's hard to get enterprises to change that fast. So, you know, it's the ability to kind of motivate them to move in that direction. And it kind of is taken 15 years to get, you know, 25 to 30% penetrated, and people migrating their data and their applications, you know, in the cloud. And if you have another architecture and other thing, they usually take their sweet time and moving. So it's not that they shouldn't, it's not that they shouldn't move, because of the reasons they decided more optimized architecture, cost benefits, things like that things that you get from edge computing, but the fact of the matter is, is they typically don't move fast. So I would say that's a that's a bit of a stretch.
Maribel Lopez:You know, I I love where you're going with this. And oftentimes analysts are aggressive one way or the other, they either think it's going to take too long, or they think it's going to take too short of a timeframe. But I think what they were really the essence of what they're really trying to get at is what you were just describing. And that's that work, you know, basically analytic workload placement, where it makes the most sense. And it doesn't have to be an A or a B. And I think when you say it that way, it's like, Well, duh, that sounds obvious. But there really was a thought and momentum that, oh, we're gonna put everything in the cloud. And then people were like, Oh, we're not gonna do anything in the cloud. And then eventually, they came to this nice sort of healthy balance of what they wanted to do with the cloud. Right. And I think that we're seeing similar trends with the edge.
David Linthicum:Yeah, I mean, it's, it's all about architectural optimization. And I talked about this in my InfoWorld blog a lot. And the fact of the matter is, it's not necessarily moving to edge computing or cloud computing or anything between arrays and keeping things on existing legacy systems is, in other words, what's going to provide the most efficiency Need to bring the most value back to the business. And that's something that we're not asking ourselves, you gotta remember. Typically, whatever we do, wherever we move application workloads and data is going to work, you can always make it work with enough time and money. The reality is we don't have an unlimited amount of time and money. And also, we're moving resources away from things that are important to deal with these sorts of architectural complexities that we're setting up. So at the end of the day, and I think this is a primary focus, where it should be focused, we're looking to optimize the system. So getting to a particular architecture pattern, that's as close to being 100% option, it's never gonna be all up 100% optimized, as close to being as optimized as you can get, and then be able to figure out the value metrics. Now it comes back, whether it's edge computing, cloud computing is typically going to be a mix of things, you know, multicloud, complex, distributed computing system that we're building on, as long as we can prove the value then then it's in its right place, we should really kind of care less what technology we're leveraging.
Maribel Lopez:I think you've spoken to this, but I'll bring it up anyway. And Joe found this interesting stat from Grandview research, which is talking about the compound annual growth rate for edge that, you know, 38 Oddish percent and that the edge growth rate is going to surpass the cloud growth. Great. That basically said the edge growth rate will be about 15 to 20%. Do you agree with that projection? Or does that projection even matter? Because it just really about wherever the workload should be?
David Linthicum:Yeah, it doesn't matter to me. I mean, I think what they're doing is they're looking at the, the acceleration of the market, and obviously, the momentum is on the edge sign clouds been around for 15, you know, 20 years, 30 years be count SAS and some of the applications delivery networks we had in the past. And so it's an old timey model, it's been around that, you know, people are just getting into today, you think about it took us a long time to move in that direction, and get the amount of penetration that we have. So edge based systems, including everything that's you know, mobile edge and cloud edge, and you know, all these things that kind of grow stealthy around you, people don't realize that they're there until they get well that's an edge based system you're holding in your hand, it's a phone, it's a car, it's a motorcycle you driving around. So I think that's what they're what they're counting on. And I think that's probably close to what we're going to see. So if they're going to surpass the growth rate is going to be the acceleration of the rate, you remember that, suddenly, the growth rate of cloud computing has been around for a long time, it's going to grow at a relatively stable pace. But now it's really interesting to edge based systems, which is going to grow on an accelerated pace. Because guess what, it's new technologies, the adoption is going to be a bit more aggressive. And also, by the way, it's cheaper to get into, I mean, Raspberry Pi's are cheaper than an actual Raspberry Pi. So we don't have to buy big big cloud infrastructures, and things like that we can get into Edge based systems in very, with very small amounts of money. So hardware costs have gone down, if you looked at hard drive, disk storage costs have gone way down in the last 10 years. So this is really kind of a viable, lower cost option in many instances. So that's going to drive the growth of it,
Maribel Lopez:you know, now wanting some random, or Raspberry Pi, but not the type that you have to connect and program. The other type would be really good right about now. And I did have to laugh, because I remember, way back in the day, we were talking about web hosting in application hosting, we had, you know, HSPs, which then became SAS, the web hosting kind of moved into being more like Cloud, and we have IoT and m to m. And that's more like edge right now. And it's just interesting to see how the different curves move over time, Jim, over to you?
Jo Peterson:Well, you know, I gotta love the marketing Lafi stuff, right? I read this thing the other day, and it said, edge is the new cloud. Like, you know. And I, David, as I think about this, I one of the myths is that edge is going to somehow displace cloud computing. But Maribel and I have talked about this, we both think they're complementary technologies, what do you think?
David Linthicum:Well, you have to be on the edge of something. So you have to be on the edge of a cloud. And normally, and I looked into the market last year, in terms of the growth of the edge vendors out there, and I found out that the primary technology providers that are cashing in on edge computing are the public cloud providers, you know, it's AWS, Microsoft and Google that are providing access to these various systems. They wouldn't be they wouldn't promote it if it wasn't complimentary for what they're looking to do. And so the reality is, we deal with architecture through tears. And so in many instances, we're going to play some processing and even some AI capabilities, some analytical capability, some storage capabilities on edge based systems could be an edge based device or a complete edge based cloud or whatever you need that's closer to the location where it's where it's can do the most good and have the lowest latency and also run with a less cost. But we're always going to need these back end systems that doozy Huge honkin processes in store petabytes and petabytes of information that's very difficult to do with an on premise device or very difficult to deal with an edge base computer, or certainly, you know, a cell phone or something like that we hold in our hands or in our vehicles. So it's gonna be a paired, it's going to be a paired value relationship. So the reality is like, when I put systems together, I look for things that can be placed at the edge for the reasons that we're talking about here, performance and cost, viability, things like that. But in many instances, you those are two heavyweight processes, you have to put them on the back end systems, they can process faster back there. And the reality is if these things working together, and different tiers are able to do much more, so it's apparent value, and edge is going to need cloud clouds gonna need edge. So we might as well just move on to it's another architectural option at the end of the day.
Jo Peterson:Yeah, it is. And that's a great answer. And speaking about paired value, I've read that you can get into a situation where you run the risk that your data capture will outpace your ability to drive business insight at the edge. So with that in mind, with the end in mind, why is it really important to have a an edge strategy with a cohesive plan for managing the data and the infrastructure? Talk to us about that?
David Linthicum:Yeah, it's ultimately you have to have an edge strategy, because again, it's an architectural option, it's a reason we can save money and looking at various technologies. You know, that book I'm writing, and, you know, looking to an insider's guide to cloud computing, spoiler alert, but I do a lot of stuff and looking at the viability of different architectural options, not necessarily looking at Cloud only based solutions, or edge only based solutions, things like that. So you have to look at the volume of data, how you're managing data, how you're processing data, and insights, or observability, you want to have that data, you want to sink back into immediately to the source of the data. So in other words, a jet engine, for example, your ability to have an insight that's on fire, and you have to pull a fire extinguisher, so you don't crash the airplane, you're gonna want to have those things on the edge based systems that are connected to a ground based system, because it's not that hard to determine if something's on fire, and then take an action on that. So again, it becomes a balancing. In other words, your ability to put the right amounts of information processing, looking at what you're looking to do, we're looking at your future uses of data. And then really kind of managing information to the point question you had is something that's centralized, so have a centralized control plane that runs data that controls and deals with data observability, and governance and management across the edge based systems and the cloud based systems. And by the way, extend that to your legacy systems, traditional systems, all these sorts of things. The reality is that we can't start segmenting data all over the place like we've been doing for the last 30 years and expect that a results, we have to have some sort of a central control plane that's able to leverage the data where it exists, manage it, where it exists in the data is hosted where it exists because of some business reason. And that's because it's the more optimized place to be sorry for the complex answer.
Jo Peterson:I'm glad you did it. And I've had the pleasure of reading a couple chapters of your book. So please plug the book again, in case everybody didn't hear the title
David Linthicum:Insider's Guide to cloud computing secrets, someone will tell you, I love that I love working title.
Maribel Lopez:So David, we're hearing a lot from different technology vendors, about creating, you just mentioned this discussion about managing and orchestrating and managing from the edge and into the cloud. And we're starting to hear about people trying to create systems for that. Is that something that you see being a coming trend over the next year, this multi cloud or edge through cloud management stack so that you can do security and governance and other things? Where is that just lovely Nirvana pipe dream, that will never happen?
David Linthicum:It's, it's right now it's a pipe dream that certain technologies exist. And you can certainly make it exist on PowerPoint, because I've done it a few times. But the ability to have this logical abstraction layer that sits across your edge based systems, your legacy systems, your multi cloud, your public cloud, your private clouds, and also industry clouds, and you know, everything else we're looking to build. And we're doing that for a few reasons. Number one, it's if we we put native security systems and operating systems on each of those devices. Each of those platforms, including edge based devices, we're gonna have several security systems, several governance systems, several operating operation systems, things like that, it's just gonna become too darn complex for human beings to figure out I mean, if you look at some of the security breaches that we had, every day, I'm one of those is always traceable back to systems that are overly complex for the training and humans that are maintaining them, and therefore they make a misconfiguration they make a mistake, and then suddenly, it's breached. So your ability to eliminate redundancy by doing everything down natively within the particular platform, cloud edge legacy and putting it up at a centralized layer. So when he talked about the data control plane that it would be an instance of that we're managing data centrally, even though it's physically is stored in different formats, different platforms. We're dealing with it through a virtualization and abstraction layer. So it's simplified in terms of how we're dealing with it. So we may deal with the native interfaces. But we're not having human beings dealing directly with that, well, it's doing the same thing for operations, doing the same thing for governance and doing the same thing for Finn ops and the other cert 30 services that have to be into these things in order for them to live a healthy life, we're just talking about when they get to, to this centralized cross cloud, you know, it's not just cloud, you know, cross platform layer, we're removing some of the redundancies out of the building, the various systems were under, we're trying to mediate the complexity of building these various architectures. And you think about this is going to be the key thing, it's not that edge computing has huge amount of holes, when it's able to do it works just great. It's the operational complexity, that's an outcome of this, because we ultimately have to throw the keys to a company that has to maintain human beings are looking to maintain the system. So your ability to have to reduce the operational complexity through these cross platform, systems, security, things like that is absolutely critical to success. So we have to figure out how that is. And so, you know, people like me, and other people are calling it super cloud and meta cloud, whatever, the Supercalifragilistic, expialidocious, whatever they want to call it. And it doesn't really matter, as long as you understand that we're trying to reduce redundancy by doing everything in a native way. And just get out of the insanity of doing that that won't scale. So let's figure something else out.
Jo Peterson:So speaking of native ways, I'm going to give you a really long question here. And you're it's all fair and good if you give me a really long answer. And maybe we'll have a short one, I don't know. As organizations plan their approach to leveraging edge computing, there's a couple of schools of thought. So the first is known as cloud out where you take what you're doing in the cloud and move it to the edge. And the idea here is to improve performance and discreet and decrease latency. And the second approach is being referred to as edge M. And in this approach, you probably are familiar with its systems and applications are native, they're built for the edge. And that allows the organization, of course, to bring data into the cloud as needed. Which of these approaches are you seeing more of in the marketplace,
David Linthicum:I see more cloud out. But that should be more edge in as more cloud out because it's easy to do lift and shift and you know, take, you know, you know, take LAMP stack stuff off existing cloud stuff and put it on an edge based computing edge based platform, whether that's a big true server, or it's an actual device, we can find ways in which we can run those things. The problem is if the developers don't have an understanding of the platform that they're building their applications on, it's not necessarily going to be optimized, it's not necessarily going to perform to the expectations of the users or loving the system. So once we get we get that thing, well, it works well, it does work. But it's, you know, it costs 150% more to operate that thing, it's going to take more hardware, it's going to take more maintenance. And that's because you didn't build it for the platform, it was moved to that platform as an afterthought, if I get to understand why cloud out is. So the engine is kind of a more of a an architecture is architecture. Because we're building the thing purpose built to run on a particular device to running on a particular server. In other words, we have an understanding what that platform is capable of doing and ways in which it stores data, the way in which it deals with security and governance and operational type systems. And by doing that, we're going to have a far superior solution, because it's something that should be able to scale. But the thing is, we're engineering it for its particular use case, we're not just engineering it to put it in one place, and then moving into another I think that's just bad, bad computing.
Jo Peterson:That was great answer. Thank you.
Maribel Lopez:I love it. That's just bad computing. And I think we should rename the podcast, the Supercalifragilistic expialidocious edge computing cast. that'll fit nicely in a Twitter handle, I'm sure. But I really hear you and respect some of the challenges around that. So one of the things I wanted to ask you is if you could share a couple of vertical specific use cases that you're seeing that are interesting to you today.
David Linthicum:Yeah, I just wrote about one today in InfoWorld. And that's the automobile industry. And in other words, the automobiles becoming these huge edge computing systems and platforms unto themselves. So you know, it used to be we had different systems that sit if you sit in your car, you have, you know, your play system that you use to encode text messages and play podcasts and things like that, and then your safety systems and your cooperation systems and but these cars are really becoming huge, huge smartphones unto themselves with one basic operating system. So the one operating system controls all the things the entertainment system, the safety system, the operation system, that's where we're moving to. So if you think about that, that is kind of the massive use case for edge computing. And the fact of the matter is we have to operate that in a connected and vice versa and a disconnected manner because you know, selling cellular networks aren't always dependable, and do so in such a way where it's able to provide automated driving systems, the ability to have self driving capabilities. And that's really kind of life changing, we get right down to it, your your ability to be more safe in your car, than you are in your house. And right now, if you go on your car, you're hugely in danger, you know, versus doing any other kinds of transportation. And I think that that is been moving for a long period of time, I built prototypes of that in the 90s, when I was a young software engineer, younger. And ultimately, it's, it's, you know, there's huge amount of value that's going to be put there, all the cloud providers are aiming for that space. And if you think about it, they're aiming to build and own an edge based system, an edge based platform, that's going to be purpose built for an automobile. And so we may see them for certain specific brands of automobiles, but I think we may see operating systems are able to span brands. So in other words, we're gonna get a truck that's running this particular version of an operating system that handles all these various systems. And we know it's going to work better than these various, you know, systems that are cobbled together and, you know, work independently, one to another aren't necessarily linked and don't add a tremendous amount of value. So it's hugely exciting for me. And I think that, you know, we're going to be talking more about this over the next five years.
Maribel Lopez:Absolutely agree. I mean, the cloud is interesting. And then you look at automotive and you can say, well, people talk about a car being a data center, right? Driving data center, so but it's the it's the ultimate edge processing vehicle when you think of it. Yep.
Jo Peterson:I think the last thing we have for you, David is a fun fact.
David Linthicum:And I brought one all right, we're ready. A cry crocodile cannot stick out its tongue. I did not know that. Hmm. So they'll never offend you. If you get a pet crocodile. Or whatever. You know, Nanny nanny booboo, that kind of stuff. But I was, does an alligator stick out at Sun? Or is this just a crocodile? And I think about crocodiles are more primitive than alligators. So maybe that's just a crocodile thing. So not that I want to run into a crocodile to test out whether he can stick his tongue out or not. But I thought that was a fun fact. I didn't know that. That is cute.
Jo Peterson:And so non tech related, which we love. Well, thank you for being our guest today. David. We really appreciate the time and you know, fun and insightful as always,
David Linthicum:you guys do a good job and asked me back anytime. I appreciate you being on I appreciate being on your show.
Maribel Lopez:Thanks, David. And we'll put links to all of your interesting work in the show notes and hope to talk to you again soon. You got it.
