Today’s episode features an interview between Matt Trifiro and Mark Thiele, CEO & Founder of Edgevana In this interview, Mark discusses his mission of lowering barriers-to-entry to deploy capability at the edge, the economics of shared infrastructure, and why he says the opportunity associated with the edge could be bigger than the cloud.
Today’s episode features an interview between Matt Trifiro and Mark Thiele, CEO & Founder of Edgevana
Following a 30-year career in IT and a post as the executive director of edge cloud at Ericsson’s Edge Gravity unit, Mark founded Edgevana last year with the goal of disrupting the existing processes for selling and buying co-location, edge, and data center services.
In this interview, Mark discusses his mission of lowering barriers-to-entry to deploy capability at the edge, the economics of shared infrastructure, and why he says the opportunity associated with the edge could be bigger than the cloud.
Key Quotes
“More and more people see the edge as a way to foster engagement with their customers, but also as a way to solve for business problems that people just assumed for years were intractable.”
“Cloud is going to continue to grow...but the edge eventually will be like a combination of enterprise workspace and the iPhone or Android store for applications. And the potential permutations of application and application types that could exist in that kind of future are just way beyond the million-plus apps that you would find in one of the phone stores.”
“I think cloud will play a bigger and bigger part going forward, but it's not because it'll take away from edge.”
“We're entering a world that edge computing is going to open up that I don't think we've ever seen before.”
“There is no way that the infrastructure we currently have at the edge will be enough to accommodate the kind of growth we'll see three to four years from now.”
“The opportunity associated with companies like Vapor IO and Edgevana is how do we actually get more value out of pre-existing footprint, beyond just the capacity of the existing white space floor.”
“I think one of the biggest areas of opportunity is...our ability to recreate life in all of its new forms, not just the remote office, but distributed entertainment in the vehicle, on the road, in the park, at home, whatever the case may be– because once people taste better, there is no going back. I think that that's likely the area where we'll see the most growth and change over the next 18 months.”
Sponsors
Over the Edge is brought to you by the generous sponsorship of Catchpoint, NetFoundry, Ori Industries, Packet, Seagate, Vapor IO, and Zenlayer.
The featured sponsor of this episode of Over the Edge is Zenlayer. Improving user experience doesn't have to be complicated or expensive. Zenlayer helps you lower latency with on-demand edge services in over 150 PoPs around the world. Find out how you can improve your users' experience today at zenlayer.com/edge
Links
Matt Trifiro: [00:00:00] Hi there, this is Matt Trifiro I'm the CMO of edge infrastructure company, vapor IO, and the co chair of the Linux foundation's state of the edge project. And today I'm here with Mark Thiele, CEO and founder at Edgevana. And we're going to talk about Mark's background in technology, the mission of his latest endeavor at Edgevana and all things edge computing. Hey Mark. How are you doing today?
[00:00:20] Mark Thiele: Hey, Matt. I'm doing pretty good. Thanks. So all things being equal and I'm happy to be here. Thanks.
[00:00:25] Matt Trifiro: Yeah. Well, thank you for joining us. So I want, wanna, I want to start, before we talk about edge and Edgevana, I want to start a little bit about your background. how, how did you even get into technology?
[00:00:35] Mark Thiele: Wow. it's, it's kinda like long story I'm filled with romance and danger, but, no, believe it or not, I was bumming around not knowing what I wanted to do. I was actually working as a security guard, taking college classes at a company called Avantech in Santa Clara, California on the corner of Scott and Bowers and the lady that
[00:00:54] used to run
[00:00:55] Matt Trifiro: a firearm.
[00:00:56] Mark Thiele: Nope. Nope, I did have DOD clearance. because [00:01:00] we did some military stuff, but, I, you know, I was taking classes, but I really had no idea what I wanted to do. I mean, I could have been a park
[00:01:06] Matt Trifiro: Where were you? Did you have an engineering background or was the liberal arts? Well, what were you, what were you kind of studying?
[00:01:11] Mark Thiele: I was, I was kind of studying whatever was good at the moment. I really had no clue who I wanted to be or what I wanted to be. You know, I looked at, I looked at my future, like somebody, who looks at, cheesecake factory menu, looks at stuff and goes, Oh sh**, there's so much stuff here.
[00:01:27] I don't know what to pick. Right. And I could like doing a lot of things. And so making a choice for me was actually really hard.
[00:01:36] Matt Trifiro: Well, look at it. Looking at your career, it seems like it's still hard.
[00:01:39] Mark Thiele: Well, it is, yeah. I mean, I I'm, I'm all over the map, but to some degree, you know, I think it's, I think it's helped me, in many ways, but yeah. All over the map, but yeah, so, the lady that ran the data center used to come down and wait for her husband to pick her up.
[00:01:53]and, she would sit and talk to me for a few minutes before he showed up. her name was Lorraine and her husband was the name was Ron. And she came down [00:02:00] one day and said, Hey, Mark, what do you think about being a computer operator? And I said, what do you mean? I said, well, you should come apply.
[00:02:06] And I said, well, Lorraine, I don't even know how to spell computer operator. You know, she goes, it's okay. Just come up and apply. And we'll see what happens. You know? So I applied, I got the job a few years later, I became the data center manager, a few years after that HP bought our company. and I had a choice of staying in the data center world, working in HPS.
[00:02:27] Sort of mini slash mainframe data centers, or taking a beginning bottom of the basement job, as a starting PC support person.
[00:02:37] And I wanted to get into client server and wanted to get into what was modern. I didn't want to go in,
[00:02:42] this is 1991. so I decided to do client server. And after doing that for a little while, I was asked to take on a role as help desk manager.
[00:02:51] So I. Okay. I don't know how to spell help desk either, but fine. I'll do that. And I just grew my career path, over the next nine [00:03:00] years with HP. I work for a terrific boss. and, yeah, you know, his, one of his favorite mottos. I mean, he lived by two things that I think were really important. One of them was definitely a bill and Dave thing cause I was at HP and that was open door and he really lived by that open door policy.
[00:03:15]but the other one was, he told me early on, as I was moving through a couple of jobs, he says bark. Work your way out of a job, I'll find you a better one.
[00:03:23] And that's, and that's exactly what I did. Right. So over nine years at HP, I went through six different positions of greater, responsibility until, you know, after starting in 1991, as a PC support tech, not even a college hire PC support tech, I left, managing 400 people in an extended organization across California, Colorado, Atlanta and Boise.
[00:03:47]and, then went to startups. I jumped into the startup modern way. I ox, you know, buy high, sell low. so you know, it was middle of 2000. I joined the networking. [00:04:00] Yeah. Perfect
[00:04:00] Matt Trifiro: six months before the crash.
[00:04:03] Mark Thiele: And so months. And then, so I found myself basically . in the middle of, the worst downturn in Silicon Valley history,
[00:04:16]
[00:04:16]
[00:04:16] and in those last, four and a half, five years, after running. app Sarah, where I started spending a lot more time on edge, working for Absera because the container management and, many of my working assumptions were that containers, microservices, functions as a service, et cetera, might be ideal for many edge workloads.
[00:04:36] So that's when I really started paying a significant amount of attention to what the potential opportunity for edge might be. and frankly, you know, why people would adopt edge and, and what are some of the barriers to entry for a being to make that market grow
[00:04:52]
[00:04:52] Matt Trifiro: yeah. So a few years ago. and, and what, what's been the biggest change in the edge market since, since you first came [00:05:00] on, on your, your radar.
[00:05:02] Mark Thiele: Yeah, I think, more than anything else, it's enthusiasm by a wider portion of the market. so it's more than just enthusiasm, I think. well, I don't think I know that I'm seeing, a lot more, from a pure numbers standpoint of companies looking at, you know, some pure edge solution and, one of the things that has gotten me.
[00:05:22] Most excited, I guess. it's okay to say that, is the fact that several of the recent edge solutions that I've been talking to potential customers about, even when I was at edge is still part of Ericsson. our solutions that I never would have considered. Right. I mean, all of us, you know, we, we all have the same, a cliche is about edge, you know, it's, you know, people always say, Oh, it's, you know, automotive and it's healthcare and, you know, it's gaming and all that stuff.
[00:05:49] Okay. That's great. But you know, give me some real world, what does it really mean? Who's really going to be using it. Who's gonna spend money on it. what does it really need from a latency standpoint? How much data does it create, but what kind of networking [00:06:00] support does it need? you know, that kind of thing.
[00:06:01] Those were, those were areas that were being mostly talked about, but ignored. up until a couple of years ago. And so recently I've seen, solution opportunities or solution requirements, for companies and business models that I never would have even considered. And so that tells me that more and more people, see the edge as a way to, during engagement, with their customers, but also as a way to solve, for business problems that, people just assumed.
[00:06:29] For years were intractable
[00:06:31] Matt Trifiro: could you, walk us through a couple of those examples that you discovered.
[00:06:35]Mark Thiele: that I like the most, it's an even name, the fear they will dispel. but basically what they're doing is they're doing high security, low latency data sovereignty to some degree or, or regional sovereignty is, is oftentimes required, 20 millisecond or less. Highly secure, remote access to control systems for critical environments.
[00:06:57] Right? I mean, you and I probably could have talked about different [00:07:00] edge solutions for hours and never come up with something like that.
[00:07:03] Matt Trifiro: And what does it, what does that mean? What does that mean for critical environments? Like what's the, what's the
[00:07:07] actual, like human use case.
[00:07:09] Mark Thiele: Like control systems, chemical facility or control systems for a, a mine or, or control systems for a dam, power facilities, even for government, activities that need to be really secure, but need to be low latency. and, in country, in many cases, And, you know, this is just, again, you know, we all talk about the, in hyperbole about, you know, the benefits and, and, and need for a five G and all that stuff and all that.
[00:07:38] But it's, I think what lens weight to the potential opportunity of the future is when these kind of undercover, you know, dark horse opportunities really start popping up. That's when I start feeling even more excited about the potential.
[00:07:53] Matt Trifiro: Yeah, so that, that critical infrastructure example, it's sort of interesting. Cause you know, one of the, one of the, distinctions that I think is really important to make [00:08:00] when talking to people about edge computing is there's a, there's a flavor of edge computing that looks a lot like low latency on premises computing.
[00:08:09]and then this version of edge computing that, is low latency put over some. Last mile network. so the computing is remote in some sense, potentially even provided by a cloud provider or somebody, it looks like a cloud provider. So, so first of all, how do you, how do you see those two distinction and, and as it applies, and then in this critical infrastructure example, which of those, or a combination of both is, is the way to deliver a solution that satisfies that customer's requirements.
[00:08:35] Mark Thiele: Yeah, I think for the most part in the solutions, a specific solution I was referring to, it's more about, distributed infrastructure access, you know, with, with common app. Accessment that ologies and, and, so if a company has facilities in Africa and Asia and the middle East, they can go to locations that are.
[00:08:53]appropriately sovereign, on this provider, in any of those locations, right. And use common [00:09:00] infrastructure and, solve for the problem of low latency and security, whether or not the destination has its own compute infrastructure available on campus. And I see that as, you know, truly fitting into what I would describe as, as the edge market, whereas on premises, edge of education of AWS cloud or something like that through something like outpost, which is tethered, to me is just, you know, on premises it with a collar on it.
[00:09:27]so yes, you could call it edge if you want. And I don't care if you want to call it edge, but specifically when I'm thinking about, edge and the edge marketplace, I'm thinking about companies that are delivering either the capability to run applications or companies that are delivering applications to a broad and diverse customer set, over a wide area, or even across the globe.
[00:09:49] Matt Trifiro: Right. Yeah. That makes it, that makes a lot of sense to me. And, you know, you talk about this, this, these changes and these sort of inflection points in the market that you've sensed that. I imagine that, I mean, you've just [00:10:00] founded a new company Ivana earlier this year, and I imagine that some of these, these insights and inflection points you're seeing drove you to start at Fanta.
[00:10:09] Can you tell us a little bit about. what precipitated edge Ivana and who your other cofounders are, are, and what problem you're trying to solve.
[00:10:17] Mark Thiele: Yeah. So the, I'll start, when at the problem, actually, no, I'll start with what I see in the industry and then why I see that as a problem, and then we'll go into the rest of, Edgevana. But what I'm, what I'm seeing in the industry is that, and I've written about this a couple of times and you and I maybe have talked about it at some point in the past, but, one of the problems that I see for edge deployment is that.
[00:10:40]there are, there's a wide gap between what people assume is even available and the variety of options. That are delivered through different APIs, different languages, different technology, solutions for solving, either some narrow niche, [00:11:00] a job opportunity, or somebody's idea of solving for abroad, edge based opportunity.
[00:11:06] And so this, it, it makes the barrier to entry for companies trying to build their own edge solutions or deploy globally. I'm a lot higher, right? It's like, why would Mark and that go to the car dealership and ask for a car with, wings on it? Right. I mean, unless it was 1959, we wouldn't ask for a car with wings on it because we assume that cars don't have wings.
[00:11:31] Matt Trifiro: Well, unless it flies, I take the car that flies with
[00:11:34] Mark Thiele: Well, there you go. And you know, so truth is that's starting to happen now. but yeah, generically speaking, people don't ask for things that they don't think they can get. And so many people are either building in secret or they're trying to solve for, you know, customer connection issues or data integrity issues or localized, IOT issues, et cetera, in their own unique ways.
[00:11:54] And so I wanna, I want to try to find a way to reduce that barrier to entry. And, I thought [00:12:00] one of the best ways to do that was simplify access to, okay. And to available capacity and capability from networking and from data center standpoint. And even just to some degree, just skills and, and deployment capabilities.
[00:12:13] For customers so that, well, it's not exactly a vending machine where Mark or Matt can go up and, and put a quarter in and get a slice of that. But, it's pretty close to that, right? Is that you can go in and define what you need search for it, identify it and select it. And even if you're buying it from a dozen different companies, make it look like it's one.
[00:12:31] And so again, lower that barrier to entry while using. in many cases using resources that are already available maximizing the, what's already been constructed and, and deployed in cities everywhere.
[00:12:42] Matt Trifiro: So if I could. so I've, I've seen it. Yeah. Some of the public, you know, in case our listeners, aren't aware, at least the time that we're recording this, it may change by the time this episode is released. Advani is still, somewhat stealthy. and so you, you may be a little circumspect in, in the specifics, about you answering my [00:13:00] questions.
[00:13:00] But, it sounds to me that a little bit of what you're doing is you're providing an abstraction layer for. The business, as opposed to like a, an abstraction layer for developers,
[00:13:13] Mark Thiele: well, it is, it definitely is more of an abstraction layer for the business and, and for the, the suppliers to that, would be marketplace.
[00:13:22] Matt Trifiro: Right. So aggregating a larger source of, of potential customers, on
[00:13:26] behalf of these suppliers, large and small. Yeah. It's pretty
[00:13:29] Mark Thiele: That's the basic, the basic idea is that,
[00:13:33] you know, you could build the best data center in the world in, Austin, Texas. And, it may in fact be the best data center in the world, but if your customer needs data centers in five locations or more, they're unlikely to spend a bunch of time negotiating with you and the 20 other companies, that reside in any, or all of those five cities in question, in order to get the five best data center operators from those five cities, they're more likely to say what's the 80 20 rule of data center offerings.
[00:13:59][00:14:00] which, which two companies or three companies have data centers in, in, in four or more of those cities, let's negotiate with them and solve this with one contract.
[00:14:09] Right.
[00:14:09] Matt Trifiro: Right. And then, and I think that's the key. It's not, not that you're providing a directory, you're not Yelp for edge infrastructure. You're actually looking to provide kind of a, a uniform, one point of contact, from which I could. Yeah. That's there any, are there any companies in sort of the traditional cloud market that have done something like this that you can think of?
[00:14:29]Mark Thiele: there are some companies that, I wouldn't say in the cloud market, but there are some companies who have, tried to simplify access to both cloud services and, data center or networking services. and they are good services and many of them are services that if I was still buying co-location space for myself, I might use, in the past.
[00:14:47] But, the things that I believe they miss is that they still yeah. Are heavily reliant on traditional, in the agent and broker model. And, the, these things Institute a lot of [00:15:00] delay in between identification and actual make of a sale, and that's bad for the customer and it's bad for the seller.
[00:15:06]and it's also true that, you know, when I talked to data centers around the world and I've been talking to me in the last two months, specifically, you know, every single one I'm going. Yeah, I did my concern that feel that they're becoming more and more. Though that in five, six, seven of these different kinds of companies that do that service, that, that we just referenced, like they're ever getting in solid leads for many of them generally speaking, you know, my information tells me that those leads are just going to brokers and those brokers are going to the big three, cuz nobody gets fired for picking one of the big three
[00:15:39]Matt Trifiro: Yeah. Yeah. And it sounds like you're, you're looking to apply a lot of software automation, right. You know, the, the, the joke is well, and we actually see this because my company, buys a lot of dark fiber. You know, you go onto a dashboard and, you know, an order some dark fiber and what actually happens on the backend half the time is like eight [00:16:00] emails get generated and a bunch of people have to like go, go, you know, plug in cables.
[00:16:06] so, so I do, I do. So do you imagine, so you play software automation, you drive some of the costs out of it, but also you start respecting some of the time to market requirements and speed requirements of a lot of these new businesses. But I'm also wondering, you know, one of my hypotheses is that.
[00:16:23] Because capacity at the edge is so constrained. And I imagined, once the momentum. kicks into high gear. The demand is going to app strip the availability that what we'll start to see at the edge, whether it's a network slice or a piece of storage, or, you know, some compute will be real time auctions.
[00:16:47] And I'm wondering if you've thought, into that domain where your customers are buying global infrastructure or leasing global infrastructure, almost on a real time basis with like automated trading.
[00:16:58] Mark Thiele: Yeah, no, I'm, [00:17:00] you're you're right on, man. I mean, I know you, you know, you guys are a hugely smart group, and so it doesn't surprise me that you'd be suggesting that. but those are some of the things that we're considering as part of the platform when you think about, cause I mean, just logically, right?
[00:17:13] If we looked four or five years ahead, how do we solve for so many of the problems that we face in the world today relative to available space, cost of computer, trying to be more sustainable. Et cetera, et cetera, et cetera. The best way to do that is to eventually make the entire landscape look more like the combination of Airbnb and yeah, just in time compute resource, regardless of who positioned the resource there originally.
[00:17:35] Right. And I realized that it sounds simple and it's not, but in the longterm, I believe that more and more of the market, based on application design characteristics, if nothing else. Along with, the capability to do policy driven workloads against latency, et cetera, et cetera, meaning that people can, you know, take a slice of compute at five milliseconds in a thousand cities [00:18:00] around the world
[00:18:00] Matt Trifiro: Yeah.
[00:18:01] Mark Thiele: it for an hour and be done and never have to do it
[00:18:03] right.
[00:18:04] Matt Trifiro: yeah. And where. Where in, you know, the, the world of, of, of today where an application developer is, you know, managing workloads in say the United States and two locations, U S Western USC, for example, that that's a very different prospect. When you get to a hundred or a thousand locations, you can't just.
[00:18:24] Do that in your head or in a spreadsheet, you actually need automation tools. and, and, and real time cost might be changing, right? So I might be a streaming game developer, and I'm like, I'm going to have a, you know, a competition. and so I want, as you said, I want, I want a bunch of, five millisecond, one gigabit per second network slices in a thousand locations, now, and also, maybe I'm only willing to spend this much, so go try to get me the best.
[00:18:52] Mark Thiele: That's right. I mean, you know, we, we, we all have a, an 80 20 rule or we should for almost everything. Right. [00:19:00] that's right. Yeah. Obviously we don't want to pick the 80 20 rule. Okay. Well, 80, 20 rule. Who's the best doctor. No, I'd rather go with the 99.
[00:19:09] Matt Trifiro: It's okay. If it's too expensive to deploy the airbag, we'll just wait.
[00:19:12] Mark Thiele: That's right. That's right. But, you know, when it comes to, applications, et cetera, we all make trade offs all the time. Right. And, and almost every application environment that I've ever built infrastructure for in my entire career made pragmatic decisions around cost versus performance metrics, et cetera, scalability, et cetera, et cetera.
[00:19:30] And so, I don't, I see, the future of edge. the ability to deploy on heterogeneous hardware to be able to traverse networks, you know, some of the things that you guys are even working on, amazing in that space. To, to really be kind of the future that allows for a workload to, you know, sort of like, you know, think about the metal from, the Terminator guy, and you know, a little bit of metal is leftover and yeah.
[00:19:55] And it just squeezes in, and it fills in the, where it's supposed to fill in, in the body. and I, [00:20:00] you know, I really see in the future. That more and more, application opportunities at the edge, especially when it becomes more of a marketplace, for people to deliver workloads to not just because you have a building in the city, but because you're, you're an external party benefiting from the information that is being created by that building in the city, et cetera, et cetera, that this, this channel that's too.
[00:20:22] So literally drop. Or ounces or pounds where or compute requirement wherever there happens to be a hole that fits,
[00:20:34] Matt Trifiro: Yeah. Yeah. That's, that's really interesting, Mark. so in, can you tell us,
[00:20:42] Mark Thiele: Yeah. I mean, realistically we're targeting, what I would guess is the kind of the top 15% of enterprises companies that approach a very wide global audience of, to some degree. but may not all have the kind of, localized footprint that, you [00:21:00] know, a Walmart or a McDonald's or somebody like that has, and still needed to be able to reach partners or customers or et cetera, et cetera.
[00:21:07]and also, suppliers, right. cloud suppliers, et cetera, that, CDN suppliers that are trying to expand networks and get access and, and, and shrink their time to value and minimize their overhead associated with contract management, identify, you know, better. Or more automation or better sustainability, specific regional plays, et cetera, et cetera that have relatively complex needs around secondary and tertiary markets, global markets,
[00:21:37]et cetera, are really our best targets.
[00:21:40] Matt Trifiro: Yeah, that's really interesting. You know, one of the things it was shift gears a little bit and talk about the industry in general. one of the things that, that I've, I've heard you say a couple of times, and, I I'd like to double click on it is that, that, you predict that the, the total. Market size of the edge.
[00:21:58] I believe it's market size. Well, maybe there's some other [00:22:00] measurement of size we'll seep out of cloud. And I'm just curious, like how, how you see that, what, what do you see driving that? but also what that means, what it
[00:22:08] means, what does it mean for the edge to be bigger than the cloud? and how does that
[00:22:11] Mark Thiele: Yeah, it's a good question. And luckily, I'm not asked to come up with specifics of how much percentage of that do I see in disk drives versus GPU versus, 5g deployments, et cetera, et cetera. Right. I'm allowed to get away with some, significant amount of nuance relative to what the market is made up, but generically speaking, when I refer to the size of the market, I'm talking about the infrastructure components, networks, the IOT devices and the software that would be used to run on them or make them run.
[00:22:42] Yes.
[00:22:44] Matt Trifiro: that housed them,
[00:22:45] Mark Thiele: not the, not the buildings. Well buildings. Yeah, certainly from a data center standpoint. Yeah. 5g IOT. I mean, even, even things like based face recognition, cameras, or cameras as IOT devices, et cetera, you know, all of that [00:23:00] fits in there. and the reason, you know, there, there are a number of reasons, man.
[00:23:03] And I, you know, I have to believe I'm singing to the choir here, but, the primary drivers. Is that I see from a value standpoint is that first is, is how I look at the market. Right. And I look at the edge market today as, roughly 1993, maybe 1994 internet. Right. And, what was happening in 1993, 94?
[00:23:23] Yeah, pretty browser priests, pretty standard. Yeah. And, and, and some of them, a very small percentage of us thought this is, this could turn into something big, you know, and, people started doing things and, and, but there weren't very many of us that were going to be Bezos or, you know, surrogate brand or anybody, but, you know, just people playing basically going from computer to computer over this new network access capability and.
[00:23:49]within four years in, you know, 93 to 97, basically if you were a mid sized company and you didn't have a website, you didn't exist. [00:24:00] Right. And so when I think about the edge and I think about, ITE history in general, it history is replete with, situations where the, the desire, the opportunity has always been there.
[00:24:16] But what, what it's waiting for is a barrier to entry to fall. Yeah. In order for it to be financially viable, in some cases it's a pure technical, other cases it's money. But the, the, the question isn't whether the opportunities are there, the question is just whether or not you can deploy technology and capability to make it financially viable.
[00:24:34] And we're getting to the point now where. Even without 5g, there are plenty of options to deploy capability at the edge for demand that is being built independent of many of us already demand in the form of smart cities, smart buildings, smart roadways, autonomous
[00:24:53] Matt Trifiro: Right. And fi fiber works fine for the well autonomous vehicles, maybe not, but fiber
[00:24:57] works fine for anything that's fixed.
[00:24:59] Mark Thiele: That's right. That's [00:25:00] right. And so you think about all that stuff and thinking, and so I did a little bit of math, right? on a small city, a hundred thousand people. And I, I picked a small percentage of the town having smart buildings and smart homes. I think it was like 25%. I'd have to go dig up my napkin that I wrote it all down on M and a, you know, a thousand hours in a day of, mostly autonomous vehicle data collection and sharing, a thousand hours a day of people doing VR or AR, and a number of other things that might be common in a more edge infused environment, but only for a small percentage of the city.
[00:25:35] And again, it's a city of a hundred people and. Turn into 35 extra petabytes of data a month. 35 extra petabytes over what they were already creating, doing whatever they were doing, playing games and, and running their videos around their homes or whatever else they were doing. 35 extra petabytes. Well, there aren't that many cities that are a hundred thousand people that even have the [00:26:00] network to traverse.
[00:26:01]35 petabytes on a, instant basis
[00:26:05] over the course of a month on one 10 gig link. That's 324 days.
[00:26:11] Matt Trifiro: Yeah, for very large data sets, sometimes putting it on a, on a disc drive on a train is the fastest way to get data
[00:26:18] a couple thousand
[00:26:19] miles.
[00:26:19] Yeah.
[00:26:20] Mark Thiele: right.
[00:26:21] So when I think about that, and I think about how much data is likely to be created, and then you, and, and, and, and the fact that people are gonna want to make use of that data in something approaching real time. I'm, you know, I, I start to think, well, what is that going to do? What kind of backlog is that going to create locally?
[00:26:36] And how do we, how do we look at that backlog? internationally. And I start thinking, then I, then I, you know, I read, I wrote my first blog about this, like in 2017. And then in 2018, Gardner came out and said 75% of data created in 2025 will be created outside of the, on premises data center or the cloud data center, 75% of data.
[00:26:57] That doesn't mean that. [00:27:00] If we had four zettabytes today that the three of them will be created at the edge. That means however big we grow between now and 2025, 75% of that new number will be created at the edge. And that's. And, and when you think about what that means from a mind boggling number of, of, of storage devices and gigabytes that have to be transferred, et cetera, and you start to realize what kinds of business models are likely to be built on top of that infrastructure.
[00:27:30]the more data we collect, the more we drive down prices for IOT devices, the more networking we make available, the more people find ways to put it, to use. And so for every first layer opportunity that people come up with, like, finding the best path out of a building when there's a fire or running the air conditioning more effectively, somebody else two years later is going to say, they're going to do air conditioning modeling across 20 different towers and determined, [00:28:00] efficient ways to, to balance power for the city or something.
[00:28:03] Right. I'm making sh** up here, but you get the idea. So the first layer opportunity is one that most of us can think about, but as we get. Compute and capability becomes more available and more permeated in society. at what we would consider something like the near edge, what will be amazing I believe is, is the new business models that will be discovered that will be built on top of that new data collection.
[00:28:28] That's exactly right. And so, when I started adding those numbers up, I realized. you know, cloud is going to continue to grow. we're going to continue to migrate apps, build net new apps in cloud where cloud centralization makes a lot of sense, but, the edge eventually will be like, a combination of, enterprise workspace and, of the iPhone or Android store for applications.
[00:28:53]and the potential permutations of application and application types that could exist in, in that [00:29:00] kind of future are just they're
[00:29:02] way beyond, you know, the million plus apps that you would find in one of the phone stores.
[00:29:07] Matt Trifiro: Yeah, that's, that's a, that's really interesting. So there's two things that I'd like to branch off of off of that. one is, my CEO Cole Crawford is, is fond of, of offering up a heuristic for edge computing that I think is pretty interesting, especially in light of what you just described, which is the edge computing is most interesting at the intersection of three Venn diagrams, data gravity, data velocity.
[00:29:30] And data sovereignty. So David gravity is like where the data is created and how, how expensive it is to move it around, right.
[00:29:38] Mark Thiele: right. What are, what are
[00:29:42] some of the, what are some of the tags and metadata associated with it that make it valuable where it is versus somewhere else?
[00:29:47] Matt Trifiro: That's right. Or, and then it's just maybe time and cost prohibitive to ship it anywhere else. And then the question of velocity is like, well, what's the half-life value of that data, you know? I mean, do, do milliseconds matter to do hours matter? [00:30:00] and then finally, like, is it even allowed to leave? Right.
[00:30:02] Is it, you know, is it HIPAA protected or is it, is there a, you know, a County boundary or a country boundary that we need to respect? And if you start looking at all of those demands, it, it makes a lot of sense that there's going to be these, these tremendous amount of demand for localized compute, storage and networking.
[00:30:19] And I think that speaks directly to your thesis.
[00:30:21] Mark Thiele: Yeah, no, I, I, every, every time I do the math, I'm astounded by what the potential is. And, you know, frankly, to something you said earlier, which is, you know, a big part of why I even started at Giovana. and I, I'm certain a big reason why, you guys are in business is that there is no way that the infrastructure we currently have at the edge, will be, enough to accommodate the kind of growth.
[00:30:47] We'll see, three to four years from now.
[00:30:48]
[00:30:48] but
[00:30:49] Matt Trifiro: I'm, I'm fine to saying the internet will break,
[00:30:52] Mark Thiele: but, No.
[00:30:56] Matt Trifiro: will, will, you know, the seams will burst.
[00:30:58] Mark Thiele: Yes. Yes. And the [00:31:00] opportunity associated with companies like vapor io and edgevana. although you could argue that, in fact, I would argue that they probably, I no edge mana, could likely be, partners in some way in this particular model. They're certainly not competitors. is that, the opportunity that we represent in many cases is how do we actually get more value out of preexisting, footprint,
[00:31:23] Matt Trifiro: Yeah.
[00:31:25] Mark Thiele: beyond just the capacity of the existing white space floor.
[00:31:29] Right.
[00:31:30] Matt Trifiro: What does it, what does that mean, Mark? I
[00:31:32] Mark Thiele: to me, what that means is that just as a, as a simple example, right? I'll mention a couple of things, but if you're, if you're Amazon or Microsoft, doesn't matter, And you're going to go out and identify a new campus for your next a hundred megawatt data center. You're probably going to spend somewhere between two and 5 million on due diligence, right.
[00:31:51] People traveling, you know, looking at the land, determining what you're going to need relative to permitting brings.
[00:32:01] [00:32:00] That's the point. That's exactly right. So if, if we can mint existing installs, where there is already power, where there is already network, where there is already land permitting and water, et cetera, et cetera. We can, we can, again, to my earlier point, lower the barrier to entry for, expanded access and, and longevity of existing infrastructure, rather than forcing any one of a thousand different people to try to figure out what the best way is to find and build a one to two megawatts in, in any town USA or Europe
[00:32:38] might be right.
[00:32:39] Matt Trifiro: Yeah, there aren't that many full city blocks left in San Francisco. For example. and so even from a, just a, a, you know, just, just a real estate standpoint, it's a lot easier to assemble, you know, seven parcels of 200 kilowatts each, than it is to find one large parcel of dropped down a megawatt.
[00:32:56]so yeah, that's a, that's an interesting dynamic that, that isn't [00:33:00] often thought of, which is just how difficult is it to add capacity into the existing urban infrastructure. that's, that's really interesting. The other question, well, not really question, but, one of my hypotheses that I'd love for you to react to, is that this whole discussion, that kind of it compares centralized cloud edge, but in my.
[00:33:22] Pointed in my view will seem sort of silly in five to 10 years. And the reason for that is like, it all just becomes part of the internet. I mean the cloud, right? We think of the cloud. Okay. Well, that's the centralized data center. We'll know. Every cloud providers pushing capacity out to the edge is signing leases and edge data centers is going to incorporate, edge compute into their traditional offerings.
[00:33:44] You know, if today I can only provision easy two instance in U S Western U S East. There's some point in the future where Chicago, West and Chicago, East and Atlanta at West Atlanta that used to become options for me, or they're going to build it into their IOT tool chain or add it to their, their security offering or whatnot.
[00:33:59] So, [00:34:00] you know, the, the, the. That distinction between edge and core, in my view will largely disappear. It'll just become part of the internet. And I'm curious if you, if you share that view or if you have kind of a nuanced, alternative to that.
[00:34:15] Mark Thiele: You know, I, it might be nuanced, but I'm not sure it really matters to the big discussion I've been in several of these debates just recently about, well, you know, the cloud will still lead and you know, you should design from the cloud out and the cloud will own the control plane and all that kind of stuff.
[00:34:31] And those are all arguments that. I can understand where they come from. but in many cases, those arguments are made from the perspective of, you have to lead with the cloud because you never want to build edge first. Cause they're just too expensive. and I think that statement, is wrong.
[00:34:45]on two counts. One is, you know, edge is only too expensive if. The business model you have, doesn't have an ROI to, edge being too expensive is based on the assumption that we're going to build out edge the same way we've built out traditional enterprise data centers, [00:35:00] even traditional cloud data centers.
[00:35:01] And I would argue that there are many opportunities based on new application designs that will proliferate over time at the edge. That the infrastructure itself will actually become cheaper over time. and so the, the, the per cost per, you know, node, whatever, is likely to not be that much higher from a premium standpoint as central cloud would be.
[00:35:23] And so, again, once you bring those barriers to entry down, what does that do to expansion of workloads? That being said, I'm just like, I don't argue that middle data centers will go away or that edge data centers will be, you know, a little half rack boxes or, or, you know, one or the other. I think all of them will play.
[00:35:44] And I think what matters in the end is. How do we best deploy applications to solve customer needs while maintaining a, you know, not only a combination of SLA and availability, SLA on performance and availability metrics, et cetera, et cetera, but [00:36:00] also on concern around compliance and security and, and et cetera.
[00:36:04] And I think public cloud will fit into more and more of the scenario, over time. But right now, average access to public cloud is far in excess of a hundred milliseconds, even inside the United States.
[00:36:18] Matt Trifiro: Yeah,
[00:36:21] well, and it's not, and it's not just that Mark it's it's you know, you can't, again, when you're dealing with old latency workloads, you can't just rely on the average latency.
[00:36:30] Mark Thiele: No, no,
[00:36:32] Matt Trifiro: latency. Like what is the worst case? Because gender is actually a bigger problem than latency.
[00:36:36] So you, in a lot of
[00:36:36] cases,
[00:36:37] Mark Thiele: especially in some of the applications that we're talking about. Right. So, yeah, so, you know, I, I think, cloud will play a bigger and bigger part going forward, but it's not because it'll take away from edge. and I think if you think about another analogy that I like to share with folks, when they ask me, well, Mark, the cell phone is getting, you know, the, the smartphone is getting so smart.
[00:36:59]you know, why would we need [00:37:00] all that edge compute? Why wouldn't we just do it on the phone? And I say,
[00:37:03] well, your phone was super smart in 1997
[00:37:06] Matt Trifiro: Right. And where, where in the last two decades, have we seen any example where the devices get more powerful and they steal workloads from the cloud? Yeah. It creates more, it creates more in both places. Yeah. So I, I definitely definitely agree. There's a little bit of a paradox going on with, device capacity in the cloud.
[00:37:27] well it's so, so that suggests to me that, there may be some opportunities for, for. Either new startups or companies that, that aren't in the, in what we think of the cloud today. So leasing of infrastructure to emerge and, you know, some are obvious, you know, some of these, these, agile bare metal companies, like a packet or Zen layer, are creating really interesting opportunities by specializing in edge.
[00:37:52] But also if you look at like the CDNs, you know, if you really look at the history of edge, you know, my first, reference to edge computing [00:38:00] in the sense that you and I are talking about it, comes from the original research paper that the guys at MIT wrote before they founded Akamai. and what you're seeing now is a lot of the CDN companies, are starting to look like cloud companies where they'll run workloads, you know, I think in many, you know, you look at like, even like a company like stack path, which like they'll run containers, they'll run VMs or on serverless workloads.
[00:38:21] And so do you see. Do you see an an there being some disruption to the existing cloud players,
[00:38:30] Mark Thiele: Well, I'll what I'll say, Matt, is that I hope there is.
[00:38:34]I mean, I don't have anything against, the offerings from, any of the large cloud providers. They all solve real problems in there, and they're a boon to business by and large. and so I don't have anything against them, but I don't necessarily, want to assume that the future of edge belongs to five global companies.
[00:38:55] Right. I would, I would much prefer a more vibrant, [00:39:00] environment and, and frankly, I think that companies that find the right way to leverage and gain access to markets. With smaller initial footprints, on shared infrastructure, whether that's shared infrastructure is network or data center or compute or all of the above.
[00:39:16]in many cases that already exist, I E building out an , installing, just adding some additional capacity and capability for, for cloud offerings or at CloudFlare, et cetera, et cetera. I think those companies and companies like them, standard, real opportunity to build. the next new set of large, call them cloud players if you want.
[00:39:40] But, I would call them, you know, edge infrastructure or,
[00:39:43]edge application delivery companies.
[00:39:46] Matt Trifiro: Yeah, it's, it's really interesting in this, this, innovator's dilemma, scenarios that, that creates, you know, I had a conversation with one of the major wireless carriers, which I will name. and, and again, even though these large companies, you talk to [00:40:00] different people in different divisions, they have different points of view, but this, this gentleman's point of view was, We build infrastructure and we own infrastructure.
[00:40:06] So we put up our own towers, we trencher it on cable. We build our own data centers, and that's how we've always done it. And that's going to be our competitive advantage. and you know, you kind of think, well as a CFO, agree with that. and does that create a career more disadvantages and advantages?
[00:40:22] Certainly advantages of owning a lot of your own infrastructure, but there's also complexities and you look at a company like dish. That's like they have a complete Greenfield and. You know, from what I can tell, they're like, well, we'd like to see a world where we don't know anything. Yeah. And I think the, one of the things that's largely overlooked, and maybe it's just cause it's kind of boring to talk about, but just the economics of shared infrastructure are so compelling.
[00:40:45] If I have to build a data center. Everywhere. I want to put a rack versus leasing a rack from someone that's already built the data center, or I have to trench cable versus leasing a wavelength, but the economics [00:41:00] are so different. And then the ability to ship, ship shift CapEx to OPEX. I mean, if I'm going to virtualize my network, why would you even want them on servers?
[00:41:09] Why want to just lease them from, from one of the cloud providers or one of
[00:41:12] the bare metal providers?
[00:41:13] Mark Thiele: Right. Well, I mean, we sort of were touching on that earlier. and so easy answer is that I totally agree with that. the idea that shared infrastructure is, the opportunity going forward. And, you know, from a, from a specific example, you know, Ivana is find to help enable that.
[00:41:29]make better use of what's already there, but also it's that, shared infrastructure tends to help you focus on what's actually important to, your customers rather than, what. Is comfortable, right. from my perspective. Right. And, when you look at, you know, any one of the big wire or wireless companies in North America, for many years I've looked at those companies and said, wow, these guys, I mean, even [00:42:00] before the term cloud became really big, I mean, we're talking about, you know, 2006, 2007, I was talking about this and saying, These companies stand to make them make mint in this circumstance right.
[00:42:11] Of cloud, because they own the network and the network is the most important part of this. Right. but I looked at it, the network, most important part, not as a way to specifically make money on network, but as a way to enable greater use of other environments. Right. And so to your point about shared infrastructure, you think about, the super highways in North America or even Europe.
[00:42:34]you know, they were, they were built in the forties and fifties and sixties by and large, and the United States still makes a buck 10 on every dollar. They spent building those in additional tax revenues. And they don't charge tolls on any of them. So how does that happen? Well, it happens because by making commerce easier, so trucks have shorter distances go to get to, deliver their loads.
[00:42:56] And, there are more reasons to build industry [00:43:00] or retail on the sides of freeways, which create more tax opportunity, more gas, sales, more Slurpee sales and so on and so on. And so on. that the, the, the highway seems like something you would forget. And yet it's foundational to all that enterprise that occurs above it.
[00:43:18] And so when companies get too tied to owning their infrastructure, they really run the risk of saying, well, I need to figure out how to make money on that infrastructure. And
[00:43:29] instead of looking at it as how is this an enabler, for a bigger picture down the road.
[00:43:35] Matt Trifiro: That's really just the, the, the, the highway analogy. well, the, the, the economics effect of highway analogy or the railroads for that matter was same thing as a really interesting way of looking at it. And it makes me think, do you, do you think that government has
[00:43:57] Mark Thiele: Well, I do think that there's potential opportunity [00:44:00] in, in certain places like, you know, making more wavelengths available, helping standardize, you know, 5g, if possible, potentially creating more right of way opportunities. But realistically, if I had to pick one thing that it would be that.
[00:44:13]I would love for us to do something the equivalent of the, you know, the Tennessee or the original, you know, phone in every house, regardless of cost type of initiative from the federal government, because, you know, frankly, I just think that, the world turns on internet access these days, whether it's for your education or for shopping or for your utilities or whatever the case may be.
[00:44:35] And, I think. We stand in a better position, with a stronger economy and a stronger populace in a, in, in better chances to grow when we give that, what I would consider a right to everyone. And so if I could pick one thing for the federal government to do it would, would be to pull together the necessary resources and.
[00:44:54] And hire the necessary a hundred thousand people, you know, to work on a project, to make sure that everybody had access [00:45:00] to internet bandwidth
[00:45:03] Matt Trifiro: Yeah, that's, that's really interesting. And I, you know, it made me think of, you know, the, the, you know, again, the highways a really interesting example, right? So, so in many ways the highways became a platform upon which all these other. Businesses and economic opportunities were built and, that's one of the power of platforms.
[00:45:24] And, you know, we're entering a world where. That edge computing is going to open up that I don't think we've ever seen before. So you think about like precision agriculture, right? And it's, and it's clear, you know, we need new ways to feed our planet or populations growing. You know, we've got, you know, climate issues and, and all kinds of things.
[00:45:43] And so the ability to grow food more effectively is a very, very important. Art of what it means to survive in the future world. And what's different is so we can imagine lots of companies building technology that can apply to a farm, but once the farm [00:46:00] is realistically connected to the internet, I don't mean, you know, somebody PlayStation cable box at the farmer's house.
[00:46:06] I mean, the farmer's got sensors and, you know, GPS driven tractors and things like that, that, that you've suddenly turned the. Farm into a programmable platform. So that, so that then the next wave of startups can innovate without actually owning farms without actually owning tractors without actually owning sensors in the same way that, you know, Uber doesn't own any cars doesn't, doesn't sell or build cell phones.
[00:46:30] Doesn't put satellites in the air for GPS, but has built this amazing service with software on top of a platform that other people have provided.
[00:46:39] And that you can, you can easily see how that's going to contribute trillions of dollars to the economy.
[00:46:45] Mark Thiele: Oh, just amazing. I mean, I mean, really it's funny cause I have a sustainability bent and you know, the first thing I think of with farming and then I can think of a hundred application opportunities on top of it. Like some of the stuff you were talking about, but you know, I think about the farms [00:47:00] in California, where I'm originally from, or at least, you know, lived there for the better part of my life.
[00:47:06] And I think about the water woes that California struggles with all the time. And, you know, generally speaking, I realize I'm generalizing, but generally speaking, you know, the, the governor will yell at the citizenry to, reduce their water usage. Well, the citizenry only uses like 15. Right. And so when we could cut, you know, 15 or 30% of our use, and that's not even 1% of total use,
[00:47:30] Matt Trifiro: yeah.
[00:47:32] Mark Thiele: hand, if farms could cut 5%,
[00:47:37] Matt Trifiro: Which can be done. I mean, you look at like the drip, drip, drip irrigation in Israel. Right. Yeah. Yeah. That's really, that's really amazing. That's really amazing. Yeah. So I have a couple of last questions. So the first one is this let's look out into the future. Let's look out, you know, 18 months and, and two questions.
[00:47:55] So one is how, how does it, how does the world of edge look different to you? [00:48:00] and then if there was one domino you could nudge and topple now,
[00:48:06] To help make that a reality or accelerated what would it be? So how's it. How's it, how's it look in 18 months?
[00:48:12] Which domino would you nudge?
[00:48:15] Mark Thiele: Yeah, that's a great question. I think in, in 18 months, we're not gonna see too much. That's extraordinarily different. but I think we're going to see a considerable amount of development and, much broader utilization in several categories of edge or verticals of edge opportunity. and I think one of the, the biggest areas that we're going to see, I mean, we're going to see it in gaming.
[00:48:38] We're going to see it in some of the fundamental ones that people always talk about. Gaming, entertainment, retail. I think we're going to see a lot in those areas. Maybe even some in telemedicine. and potentially even in remote medicine, but maybe not in 18 minutes a month, but I think one of the biggest areas of opportunity, which may sound stupid, cause you'd think we'd been doing it forever, is working [00:49:00] from home, which I now call working from anywhere.
[00:49:02] Right. And, I don't know what the estimate, the current estimate is for people not going back into the office, but, I feel like a conservative estimate is, is 20%. likely not go back to offices ever. Right. yeah.
[00:49:19] Matt Trifiro: personally. I never need to go back into an office.
[00:49:22] Mark Thiele: Yeah, same, same for me. And I mean, I specifically, I mean, you know, some of the companies that I would be thinking of, founders, who have told me, I'm not going to renew my lease, I'm not going to have a building start up anymore.
[00:49:33] We're not going to do it. And so. That's right. That's right. And so, you know, 20% is probably ultra conservative, but imagine if 20% of, companies, are buildings no longer existed. And so working from home took on a different meaning because when, historically when you were, I have said, we're working from home, we're working somewhere that's 30 minutes or 40 minutes from an [00:50:00] office.
[00:50:01] Right. Well, what happens when there is no office? Now working from anywhere is much more appropriate. Right. And so what does working from anywhere mean? Well, it may not mean a whole lot of difference for the average traveling salesman or consultant, but what does it mean for people that work on, high touch, interactive applications all day long for their company.
[00:50:22] Can can they afford to have the kind of interruptions and the poor performance that the average household suffers from today, what would that do to longevity for employees? What would that do for productivity for companies? So our ability. Over time to recreate life in all of its new forms, not just the remote office, but, but, distributed entertainment in the vehicle, on the road, in the park, at home, whatever the case may be to recreate that in a way that, that doesn't just do it as good as we've been doing it.
[00:50:55] But, but beats it because once people taste better, there is no [00:51:00] going back. And, I think that that's likely the area where we'll see the most growth and the most change,
[00:51:06]over the next 18 months.
[00:51:09] Matt Trifiro: That's a really exciting vision. It's a really exciting vision Mark. And so, and so w like what, what, what, which of the Domino's would you, would you,
[00:51:19] to make all that happen?
[00:51:24] Mark Thiele: I think, I think potentially the biggest domino is, is better, security. In general, right. I realized that's really boring and it's probably not what anybody on the phone or computer wanted to hear, but I think security is one of the, one of the biggest areas of opportunity because, and security combined with compliance is, The more, automagic we can make people being able to work securely and protect data sovereignty or data compliance issues for their companies.
[00:51:53] The more effective, these kinds of plans can be and more natural they can be for the user. So. Those are [00:52:00] the things, you know, over and above what people like Vapor IO and egdevana are trying to do for the market as a whole, those are the things i see as real opportunistic for that segment of the opportunity.
[00:52:07] Matt Trifiro: Yeah, and certainly security is, is I agree with you on a major, area of, of concern and potential innovation, that needs to happen. That's great. So, Mark, thank you so much as usual. having a conversation with you is, is a delight. We probably could have gone on another two hours, and maybe we'll have you back on and, and do part two, but,
[00:52:36] Mark Thiele: Yeah, I I'm a pretty active on Twitter, and, almost as active on LinkedIn. So feel free to go find me under Mark Teeley on LinkedIn and on Twitter. I am M T H I E L E one zero. and, I, I love interaction. In fact, I'm commonly, testing my assumptions on both platforms. So I love that interaction and learning every time I'm out there.
[00:52:59] Matt Trifiro: who has [00:53:00] MTLE one through nine. That's what I want to know.
[00:53:02]Mark Thiele: I, I don't know, but it wasn't old. It wasn't old. log-in that I used for an account somewhere. Mark Teeley at the time, I just did 10.
[00:53:12] Matt Trifiro: That's awesome. That's awesome. it, we'll put, we'll put all these links in the show notes too. So Mark,
[00:53:16] Mark Thiele: Awesome.
[00:53:16] Matt Trifiro: Really appreciate the conversation today
[00:53:18] Mark Thiele: It was fun.
[00:53:20] Thank you. You guys take care.
[00:53:23] All right, bye.
[00:53:25] Thank you.
[00:53:26] Matt Trifiro: Thanks Mark.