Today’s episode features an interview between Matt Trifiro and Iain Gillott, founder and president of iGR market strategy consultancy. In this interview, Iain discusses the many challenges and intersections between wireless, mobile, and edge computing, his views on where the telco operators will fit into edge, the role CBRS will play, and much more.
Today’s episode features an interview between Matt Trifiro and Iain Gillott, founder and president of iGR market strategy consultancy.
Iain is a leading wireless and mobile industry analyst, and since 2000, has run iGR, a market strategy consultancy focused on the wireless and mobile communications industry.
In this interview, Iain discusses the many challenges and intersections between wireless, mobile, and edge computing, his views on where the telco operators will fit into edge, the role CBRS will play, and much more.
Key Quotes
"Problem number one is people don't realize how much edge is actually out there. A lot of people think, ‘to do edge computing, I’ve got to wait for 5G. There'll be no edge until there's 5G.’ But some of the very first edge compute was put in with ethernet wired connections…The second problem is that everybody has a different definition of the edge. My definition of the edge is if you take one more step, you fall off a cliff. That’s the easy one."
"The perceived value of edge compute in the mobile network is with the cloud vendor, not with the mobile operator, and I think that's a problem. It doesn't mean they're not needed. The mobile operators are needed, but from a mobile operator perspective, their customer is the cloud vendor. It is not the enterprise or the end-user…In this respect, I hate to say it, but they are being pushed into an edge bit-pipe relationship where they're going to rely on the cloud guys to deliver that business to them."
"When we've seen things take off in technology, they're one of two things: They're either huge volume, mass market, low cost. Or they're very specific, super high requirements, and expensive. It takes a while to build the middle ground."
"The operator will tell you it's the connection that's the value. The application developer says it's the application. The edge compute guy says it’s the servers. But the whole solution is the application, some compute, and connectivity. But the CIO looks at the entire solution. The answer is we need all three."
"I think the problem with edge compute is, it's very difficult to pick winners and losers right now. Some operators are going to be successful with their strategy and others will not even be involved. So picking winners and losers is difficult."
"I think people who've got a very rigid view of ‘the edge is here, and I'm going to do this and sell it to these people,’ that's a problem. You've got to be much more fluid and say, ‘to this customer, the edge is over there, for this customer, the edge is over here, and for this customer, it's actually completely different.’ You’ve got to be really fluid and flexible, probably like we haven't been before as an industry. And I think this is why the operators will have a problem because they're not flexible."
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 Ori Industries. Ori Industries is building the world’s largest edge cloud. Their products power the next generation of intelligent applications through unparalleled access to major communication networks worldwide. Ori is laying the foundations for application developers to seamlessly deploy to uncharted edge computing infrastructure across the globe. Learn more at ori.co
Links
[00:00:00] Matt: [00:00:00] Hi, this is Matt Trifiro, CMO of edge infrastructure company, vapor IO, and co-chair of the Linux foundation's state of the edge product. Today. I'm here with Iain Gillott, the founder and president of IGR. And we're going to talk to him about his background as a technology analyst and get his thoughts on the current state of the edge and the intersection between the edge and the mobile wireless industry.
[00:00:20] Hey Iain, how are you doing
[00:00:21] Iain: [00:00:21] today? Good. Thanks.
[00:00:24] Matt: [00:00:24] Yeah. Well, this is awesome. It's been a while since we've talked, what I like to do when we start up these interviews is like, just find out a little bit about you. And so I think w I mean, how'd you even get into technology?
[00:00:37]Iain: [00:00:37] well, I'm an engineer.
[00:00:39] I'm a secret closet engineer by training. Yeah. So my degree was actually, computer systems engineering. so I, I've got a degree and after college I became a programmer and, realized fairly quickly that there's not much money in programming. All the sales guys were having the fun. And so I actually became, the market, [00:01:00] sales support guy between the sales group and the engineering groups for, the company that did the first billing systems and clearing systems for the mobile operators back in 1992.
[00:01:15]so my job is basically to work with the operators, define needs and, requirements, then work with the engineers to say, you know, can the product do this? And then build the cost model and the proposal and all these things and hand it off to the sales guy and say, Go close. and, the, my clients at the time were, Southwestern bell, mobile Ameritech because bell Atlantic, mobile GTE, I said, ITI, yeah.
[00:01:44]right. and you know, they've been through three rounds of mergers, I think, to get where we are today, four rounds. So, so, so I do have a technical background, which is actually interesting being a research analyst because, you know, wireless is very [00:02:00] technical as, you know, and we, you know, you can get lost in the bits and the bites and the acronyms pretty quickly.
[00:02:07]and so, but I do have that engineering background, which does help, when we start talking about packets and megahertz and, throughputs and you know, all those good things.
[00:02:16] Matt: [00:02:16] Yeah, there's certainly, a lot to be gained by having a working knowledge of the engineering, especially now when the entire mobile industry is transitioning to 5g, there's just a lot of acronyms and new terms.
[00:02:28] What, tell us a little bit about what a research analyst does.
[00:02:32] Iain: [00:02:32] Yeah, a good question. I've been doing it for awhile. no, so what we do basically is, you know, I use this description with my in-laws and my parents, is, you know, cause they were always asking, what do you do? and so, Hey, you go into the grocery store and you go to buy cereal. Right. You walked down the cereal aisle with your mask on and, the cereal, the cornflakes, or the Frosties, or whatever, catch your eye.
[00:02:59] And you [00:03:00] think, you know what? I'm not going to buy the Cheerios because I'm going to buy the Frosties and you take the Frosties and maybe you take a bigger box and you were planning because there's a deal on for the bigger box. Right. And you walk out the store and what you don't realize is somebody actually planned that aisle.
[00:03:16] They planned where that box went on the shelf, how high it is on the shelf. They planned that the promotional box it's a little bit bigger and better value stands at the front, et cetera. and we do the same thing. So somebody, we did market research to say, Hey, you know, people who like Cheerio's occasionally they like Frosties and they cross shopping and people like this and believe it or not, I'm being serious now.
[00:03:42] They actually do height things. So, sure. Women tend to be a little bit shorter. We'll buy certain products. That'll be lower on the shelf. my wife's five foot three, she can't reach the top of the grocery store shelf. So she doesn't buy products from that. somebody worked all this out. We do the [00:04:00] same thing for wireless and mobile.
[00:04:01]so somebody has to build the networks. Somebody has to design those networks. and that is based on usage and consumption and behavior. So we look at. How people use the networks, how you use those devices, how much data they use, what they do with them, where they go. we used to do a lot of traffic analysis, looking at the traffic usage throughout the day.
[00:04:23]et cetera, Jason. So we do market research. We do surveys and studies of people in how they use. Tech mobile technologies package that up and turn it into reports and research and sell it to the industry basically. so give you an example, the, you know, the networks we build today, you build and your customers build, they don't just put in capacity.
[00:04:48] Everywhere. Right. They put in capacity where it's needed. Right. And with COVID-19, one of the things that's happened is that people are not driving to the office in the morning [00:05:00] and they're not driving back in the afternoon. They're staying at home. So the traffic usage within the Metro area has changed.
[00:05:07] Right. And that's impacted the mobile network. So understanding that is important to the engineering side, then the planning and all those things. And it, obviously it ripples down to the cost of networks and things like that. In essence, that's what we did. We do just like people study cereals and baked beans and people.
[00:05:27] What they prefer in their date, deans, we looked at it from a mobile perspective. So,
[00:05:32] Matt: [00:05:32] so you're a research analyst in wireless grocery stores with a specialty in Cheerios.
[00:05:40] So are your customers, are the system operators, the telcos.
[00:05:44] Iain: [00:05:44] Yeah, not primarily. so, obviously within wireless and mobile, we've got a very long tail on the ecosystem. Right. and as I said, at the beginning, we used to have a lot of operators at the top. And now in the U S we've [00:06:00] essentially got three major ones and then a lots of small regional guys, and we do deal with them as well, but we do deal with the big operators, but then below them, the infrastructure folks who are building that, there's the construction guys, actually physically building the network that, there are the tower companies who.
[00:06:19] People are putting the network on. there are the fiber guys who obviously connect all the networks together. there are the handset folks. There's the applications. There's security. There's the packet core. there's the data center. what else? and then there's a lot of. operational companies in between.
[00:06:38] So a good example is roaming. You know, we all roam across the U S or used to roam across the U S and didn't think about it. you're old enough to remember back in the day where you used to have to put your credit card in Rome, right. And then it became automatic, and now it's across the world and we take that completely for granted, but there's an amazing amounts of systems in the [00:07:00] background that make that happen.
[00:07:02] For authentication and security and handoff and billing and reconciliation. So there's vendors in that space as well. So wireless has a very long tail on the ecosystem just to support, you know, relatively few operators, compared to what we used to have.
[00:07:19]Matt: [00:07:19] , so iGR, your company is focused primarily on the wireless industry, but you and I came to know each other through edge.
[00:07:26] So w how did you first get interested in edge computing?
[00:07:29]Iain: [00:07:29] so yeah, so when I started the company is actually 20 years ago, 2000, The reason we started, it was wireless. I was working for another company, had various different groups and I said, Hey, we should focus all the wireless together. The wireless is going to be big.
[00:07:43] And at the time wireless was around 15, 20% of the telecom industry. If you like. So I mentioned all those players, you know, the big players were not. Southwestern bell mobile. It was Southwestern bell, right? Bell South was the big guy. [00:08:00] Yeah. I think
[00:08:00] Matt: [00:08:00] PAC bell was my original cellular provider.
[00:08:02]Iain: [00:08:02] yeah.
[00:08:02] Right, exactly. And of course now fast forward, 20 years, everything's wireless, right? Everything's mobile. the edge compute came up probably four or five years ago. There was the original, the, the MEC multi-access edge compute working groups. And they had a conference in, Munich, which w they asked us to chair.
[00:08:25] So I ended up chairs and it was in the middle of October Fest. It's the best time for a conference. And, so that's how we got involved in edge. And at the time it was very much, Hey, we think we can do this. we need some standards. The telcos has got to play. We've got to get into the networks.
[00:08:42] We've got to kind of, you know, work out where everything fits and what we're going to do with it. And the big question at that conference, and I think it was four or five years ago was, our applications going to move between different edge nodes. Right. And how that was the first question. do we even need, that was the [00:09:00] first question.
[00:09:01]and so, and there were some demos at the time. I remember as different people doing video demos and they're showing video acceleration, and, things like this. This is all of course pre 5G, right. We were still in an LTE world. and so we weren't really talking about. Very high bandwidth networks or ultra low bandwidth networks.
[00:09:20] We were talking about offloading workloads for local processing, like video analytics and video processing. and that's how we first started. Started got involved. And the, there were several sessions, it was kind of interesting. nobody had really size the market, looked at the, you know, kind of laid it all out, who the players are.
[00:09:40] We did that. and so, yeah, it was, it's kind of the, the engineers would just. The finished kind of playing with it, the kind of working out, okay. Now what are we going to do with it? Right. You know, that process as it goes through. but, then fast forward now, of course we've got, you know, a variety of different companies involved doing different things.
[00:09:57] So,
[00:09:57] Matt: [00:09:57] so, so what's, and I know this is, this could [00:10:00] take an hour, but give me the thumbnail sketch. I mean, what, how do you view edge computing today?
[00:10:06]Iain: [00:10:06] I think a couple of things with that, number one is people don't realize actually how much edge is actually out there. and I think a lot of people think, Oh, well to do edge computing, I have to have 5g and I've got to wait for 5g.
[00:10:20] There'll be no fight. There'll be no edge until there's 5g. Some of the very first edge compute was put in with Ethan net. Well had connections a bit more about that.
[00:10:32]Matt: [00:10:32] you know, part of the challenge is there's 150 definitions to edge computing, which is one of the reasons I started co-founded the state of the edge was trying to get through some of that.
[00:10:40] But, so what do you mean by edge computing and what are these things that were hooked up with ethernet? So,
[00:10:47]Iain: [00:10:47] so the second problem I say is that everybody has a different definition to the edge. My definition of the edge is if you take one more step, you fall off a cliff, right. has the easy one.
[00:10:58]but basically, [00:11:00] everybody, if you look at them, let's just take a mobile operator. A mobile operator looks at the edge and says, It's the tower, it's the radio, right? Actually a device vendor looks at the edge and said, well, actually, yeah. It's the device itself, it's the closest to the end user.
[00:11:17] Right. if you look at the cloud provider, they looked for the back in the network. if you look at a data center company there, well, the edge is the wall of our data center. Right. But some of the very first, edge compute stuff, we still was actually where, I'll get one of examples, the manufacturing plant.
[00:11:36] And they had senses and machines all over this big plant and they were all connected for monitoring and control. They all have different connections and they were coming into different systems. Right. And they wanted one view or the biggest problem they had with getting one view was that they have all these different feeds.
[00:11:58] So they took all [00:12:00] those feeds, put them into a one server, one edge compute box it's on the edge. It's right next door. From the machine to the box, consolidated all the traffic, normalized it and put it into their, into the tools and monitoring systems. So in that definition, the edge was the tool, right.
[00:12:21] They've taken multiple feeds and different things, pulled them together to give one view. just like you could say that you have multiple cell sites with traffic coming off, then come to one. Edge processor in a mobile network, do some things and pull it out. and so that's, that was one of the very first ones.
[00:12:40]there's been some wifi things. one of the other ones we dealt with fairly early on, was, share business model for a large retailer. And they had a very interesting problem. And that this was to me was kind of indicative of where we are is. they had a thing come up with it. I think it was a hundred megabit per second [00:13:00] connections into their stores and they were overloading them because people would come into the store and need to upgrade the software on the device.
[00:13:08] So if you go into a lot of these technical stores, not to tell you who it is, but you probably have to work it out. Hello. Hey.
[00:13:18] Yeah, but you used to go in and say, Hey, I've got a problem here. And the first thing to do is, Oh, your software needs updating. Right. Well, every time he did that, it's one and a half, two gigs of data. and so there were pulling that down each time they had all these devices, all these products, they had a wall of screens showing video that was being streamed.
[00:13:37] Then they had, and every time this is interesting. Every time the salespeople came to work. They have videos to watch, to train them on the new devices, new functions. And so there was streaming all this stuff down from the main data center, the cloud, and they were having to upgrade the connections into those stores from a hundred bags to one gig.
[00:13:58] And it was costing them [00:14:00] X thousand dollars a month to do that. So the project was pretty simple. Take an edge server, some software on it, put it in the store, overnight cash, all the training, videos, cash, all the software updates, cash, all the content. Right? Keep it updated. Keep it in sync. Keep everything in the still local.
[00:14:19] You can keep the 100 megabit per second. back home to do that. And the economies we worked on, the cost models for this and the price, everything this justified itself extremely quickly by not upgrading the backhaul. Not going to a one gig connection. And it's the most boring business case. There is, you know, you want to say, Oh no, you need, augmented reality.
[00:14:47] And actually the company was planning to do more demos and more stuff, and they couldn't do it on a hundred Meg connection. So they actually put in the architecture at the edge that was upgradable that would [00:15:00] allow them to grow in the future and do those things. And it. it's quite amazing. didn't need 5G back then.
[00:15:07] This is three years ago. so it's interesting.
[00:15:10] Matt: [00:15:10] Yeah, that's really interesting. So our CEO at vapor Cole Crawford is fine to fond of saying that the killer app for edge, is cost avoidance
[00:15:19] Iain: [00:15:19] and
[00:15:21]Matt: [00:15:21] cost avoidance, and the easy button. Yeah. And in this case, it makes sense, right? If this customer could have pushed a button and saved a bunch of back haul.
[00:15:29] So the first example you used with the factory and all these different streams, how is that any different than on premises computing?
[00:15:42] Okay. Fair answer.
[00:15:43] Iain: [00:15:43] Right. Yeah, I don't think it really is except that, if you've got multiple manufacturing plants doing that, right. Yeah. let's say you've got multiple warehouses. actually I'll give you another example. I can remember. there was a company, was, [00:16:00] taking refrigerated trucks.
[00:16:02] And, this refrigerator truck, they had all these sensors on it to know if the temperature got too high and your broccoli started to rots and all this stuff. And, all that data was going straight up to the cloud. So Amazon is using Amazon web services. So there's a data connection that's painful and there's pay for the cloud.
[00:16:20] And it was the sensors going, Hey. Yeah, I'm okay. I'm okay. I'm okay. Okay. And there's like 20 sensors on these things. and that was just on the trailer, not the truck itself. And they literally took a raspberry PI platform, built a little unit that all those senses went into the raspberry. And the raspberry looked for access a stored all the data and be in look for exceptions.
[00:16:44] So it didn't send up to the cloud. I'm okay. I'm okay. It waited until I'm not okay. And here's the problem. Right every hour. So it did a data dump or something like that, but the cost, the ROI on that was [00:17:00] actually avoiding the mobile charges of all that data going backwards and forwards and the cloud charges again.
[00:17:09] So is that
[00:17:11] Matt: [00:17:11] I think I have an answer. I think I haven't answered all the questions I was thinking about it just now. So. I think the concept at edge only makes sense in the context of centralized computing, because that's what you're doing. You're either moving that compute to the edge, right.
[00:17:28] To a closer location or your. You know, doing things like data reduction and storing it in centralized compute, but it almost doesn't make sense. It's just on premises or embedded systems prior to having a relationship to centralized compute. Do you agree?
[00:17:46] Iain: [00:17:46] Yeah, I mean, I kind of looked at it as you're moving the workload.
[00:17:49] Right. So if I've got a workload that I can move. So in the case of the truck, Those sensors weren't doing anything. There was no workload on the truck. It was [00:18:00] all being sent up to the cloud.
[00:18:01] Matt: [00:18:01] That's a great example. So we're literally pushing it out to the truck on a raspberry PI.
[00:18:05] Iain: [00:18:05] So what are we going to do here is we're moving the workload out to the cloud.
[00:18:09] We're moving it down to the truck. So in that case, the edge is the truck we could make. Maybe we could've moved that workload to the local cell tower, maybe to the local operator, maybe. Yeah. Somewhere in there. Right. We're still moving the workload to somebody's edge the edge of something. So I kind of looked at it and it gets difficult.
[00:18:31] And you see, we're raising this issue here of where is the edge, what is edge compute? Because, and now it gets more confusing with cloud. We have cloud edge, right? We've done all those types of things. And we given these different architectures, but in a sense to me, it's. There's a workload, something is happening here.
[00:18:51] I'm going to move where that happens. And as a result of that, I have to put a processor out there. It could be as simple as raspberry PI. It [00:19:00] could be, some of the, the installations you guys do are a little bit bigger than that. but you know, it could be, multiple megawatts sitting at the cell tower eventually, right?
[00:19:10] Matt: [00:19:10] Yeah. Yeah. that makes sense to me. ,
[00:19:12] so, what are you seeing out in the market in terms of how the telco operators are looking at edge computing and investing in it?
[00:19:23] Iain: [00:19:23] Yeah, it's a good question. And, some people may not like the answer. but this is the reality, I think. And I'll talk about failings first. So one of the problems with the telco industries in general is we take too long to do stuff. Right. We always think about, Oh, right. So, so the edge can be the MEC bodies.
[00:19:45] I mentioned we were in Munich, then we're in Berlin years ago, we sat there for years talking about stuff to get standards approved and deployed. And so everybody had the same thing and the [00:20:00] hyperscalers just go. Standards. Oh, no. You just want to do this straight down the back. Yes. Yes. So what we've ended up with is the operators and the operators have done this before.
[00:20:14]they'll do it again. I could point to some of the things, with messaging and, some of the, the rich communication suites and all those types of things that tried to get video conferencing on. The mobile operator networks with standards and they develop that stuff for years. And then, well, we had FaceTimes
[00:20:36] Matt: [00:20:36] Skype.
[00:20:37] Exactly.
[00:20:37] Iain: [00:20:37] Yes. Set standards for me. You just want to use this browser thing. It's
[00:20:41] Matt: [00:20:41] called TCP IP.
[00:20:44]Iain: [00:20:44] and that's a great example. And I think the operators have done that again. The telcos have done that again. they, as I said, that was pretty early in the edge space. but what we've ended up now is, partnerships basically, and that they seem to be announced every [00:21:00] week is between a telco and a cloud provider where the cloud wants to take their capability and.
[00:21:08] Push it down into the mobile network, which is great. That's where we want it. We want it as close to the user as we can get. We may, we don't have the traffic to do it all yet, but, eventually we'll talk about apps and services and things. but, you know, we want to do that. Now. The problem I see is the operators I've spoken to.
[00:21:28] Hey, I've just done a partnership with cloud vendor. A I've done a partnership with cloud bend debate and their stuff is in our network and it's right down there and it's at the local data center. It's next to the EPC. It's we're in the ecosystem like, congratulations, you are, you go to the cloud vendor.
[00:21:50] Can you say, show me your marketing materials. So your edge. Compute capability. So there you go. And I've done this and it [00:22:00] makes no mention whatsoever of the telco or the mobile network. It says, Hey, cloud customer, you want to collect your edge devices, your device, your mobile devices. We can move that processing closer to them.
[00:22:14] We'll make it happen to you. So the perceived value of edge compute in the mobile network is with the cloud vendor. Not with the mobile operator. And I think that's the problem. It doesn't mean they're not needed. The mobile operators are needed, but from a mobile operator perspective, their customer is the cloud vendor.
[00:22:32] It is not the enterprise or the end user on the other end. and so I hate the term bit pipe in some respects because it's the problem with the term bit pipe is people think it's bad. Actually being a bit, pipe is good if you're very good at being a bit pipe and don't carry all the overhead of stores and retail and customer support and all the things mobile operators have.
[00:22:56]but in this respect, they've, I hate to say it, but they're really a [00:23:00] bit, they are being pushed into a, an edge bit pipe relationship where they're going to rely on the cloud guys to deliver that business to them.
[00:23:09] Matt: [00:23:09] Yeah. I mean, it always seemed a little nonsensical to me. and this was years ago, not so much happening more when the telcos were posturing around having their own cloud.
[00:23:18] It's like, well, what developer is going to develop? For the telco cloud. Like I just don't see that happening. Right. Developers want to build for the major cloud providers and employers want to hire those developers and that's how the ecosystem works. But to some extent, you know, there's a couple things that I think will change with 5g and I'm interested to get your reaction.
[00:23:37]so one of them is network slices, and that is in some senses, like really special bit pipe. and the reason I think networks are important is. Even if I can get a workload right next to the Rand. Right. It doesn't do me any good unless I have a solid SLA down to the receiving device. And it seems to me that the best way to get an SLA is to buy a network slice.
[00:24:00] [00:23:59] And the only way to do that is to get it from the telco. Now, the telcos may, sell it wholesale to the cloud providers. So when I provision an edge instant on Amazon, I get a Verizon. You know, network slice. So, so there's network slicing. And then, and maybe this is really where mec will find some success.
[00:24:17]you know, Mec was going to be essentially a cloud was going to run workloads and they thankfully got away from that. but now it's an API to some of the things on the Ran that you would never otherwise have access to, you know, like network congestion and things like that. So do you, are there.
[00:24:31] Do you agree with those two hypotheses and then are there other things that the telcos in this more restrained view of their role, bring that are of, you know, of really legitimate, exceptional value in this new world?
[00:24:47] Iain: [00:24:47] Yeah. I do agree with your first statement. If we're talking about mobile wide area coverage.
[00:24:55] So if I'm, a logistics transportation company and I've got [00:25:00] trucks and things going all across the U S or Europe or wherever, and I want communications. Right. And, I want, and I want edge compute. Maybe I'm doing some video stuff and things like this has been discussed. one of the applications I heard was that, you know, you've got a.
[00:25:17]logistics, transportation truck, and, you know, guy pulls up in front of your house and the hops out and delivers your package to him. Right. those drivers are on their own and there's actually been discussion of having a button in the truck with, if there's a problem, they hit the button and the doors lock and the video camera in the truck goes full HD video.
[00:25:41] Instantly and starts recording in the cloud or locally. and in the, your plan is that they can actually take control of the truck and drive it out the way. But if a driver gets incapacitated or something like this, right. And [00:26:00] think about that now you're thinking, okay, I really want edge compute to do that.
[00:26:04] Right. I really need low latency and I need an uneven network slice because I'm going to want bandwidth right now at that place dedicated to me. I don't want some teenager playing fortnight, taking the bandwidth on that cell site. Right. I want it for me and I'll pay for it. Right. So
[00:26:23] Matt: [00:26:23] you may want to buy a competing slice, a little lag.
[00:26:29] Iain: [00:26:29] Yeah. So in that example, I'm going to be on the public mobile networks on the tower, and I'm going to want the slice. And I think your scenario in that one is very good. there's values as that we can realize that requires edge compute the other side to this as well. I can build a mobile network today using CBRS in the U S
[00:26:49] Matt: [00:26:49] private mobile network.
[00:26:50] Yeah.
[00:26:51] Iain: [00:26:51] Even a private mobile network without the need of an operator. So, I could be a warehousing distribution center operator and I've got distribution [00:27:00] centers all across the U S I put private networks, each of them to do stuff. I am the network slice. now this, now I've defined the slice that says, well, we're going to have these machines running this, and then we're going to have security cameras on this slice and dah, still going to have edge compute because I'm gonna do a lot of processing locally.
[00:27:20] Right. And, yep. so. th one of the interesting things with slice network slicing came up as a 5g discussion that the operators would take slices of this network and sell it to people and we'd have dynamics and they do it on a weekend, or they do it on a Monday and all this stuff actually slicing that actually come from the enterprise with a private network.
[00:27:44] Before it shows up with an operator. that's a very real discussion because there's a lot of big enterprises, who have got the facilities and the means to do that. And it could be a very simple example. Like I've got security [00:28:00] cameras and I've got, Manufacturing machinery. I don't really want them on the same network, but I'm going to have, I'm going to have one network too.
[00:28:08] Right.
[00:28:09] Matt: [00:28:09] But I may, I might create a virtual network over two different dedicated slices. yeah. Yeah. Well, and I actually, that, that reflects kind of what I'm seeing in the market. I'm seeing a very rapid cost reduction and, well back to the easy button and cost reduction, right?
[00:28:22] With private networks, especially with, unlicensed spectrum, it is becoming so cost-effective and easy to deploy and operate or pay somebody to deploy and operate a private network on your behalf. That's a fascinating insight.
[00:28:38] Iain: [00:28:38] The good thing with LTE and subsequently 5g is we are really good at policy control.
[00:28:43] We're really good at authentication. We're good. At security, we can do handoff between cells really well. Right. Right. And that you just took the words right out of my mouth. Wi-Fi doesn't do those things particularly. Well, it does other things. Well, [00:29:00] but, but when I've been talking to enterprises, looking at building private networks and we do a lot of work on this it's they're like, no, we want to consolidate our workloads onto one network that can manage all this and it's secured.
[00:29:12]and by the way, it's got to pass the 5G. Cause my boss says. Yeah, he's gotta be in 5G, two years, things like that without knowing what 5G is,
[00:29:22] Matt: [00:29:22] five is better than four, right?
[00:29:23] Iain: [00:29:23] Yeah, that's right.
[00:29:24]Matt: [00:29:24] yeah, exactly. so let's talk about CBRS and unlicensed spectrum. so first of all, I suspect there's a lot of people in my audience, who are deeply into edge computing that may know very little.
[00:29:38] And anybody thinks CBRS is CB radio. So can you give us a simple definition of CBRS and explain why it's so important?
[00:29:45] Iain: [00:29:45] Yeah. So what happened is the FCC, found the chunks of spectrum at three and a half gigahertz right in the mid band and, do some, astute lobbying and, persuasion.
[00:29:56]they will persuade the suede to not just go sell it to the mobile [00:30:00] operators as they normally do. Right. But Rob, let's try something different and say, Hey look, there's a lot of people out there with like a little bit of spectrum. They'd like to do some cool things, but they don't need 40 megahertz.
[00:30:12] They don't need a hundred megahertz and they don't need the United States. They need that little chunk. For where they are covering their warehouse. So the idea was they came up with CBRS and, it's 150 megahertz spectrum. and basically, you and I, anybody can buy a radio, and you buy this radio and you plug it in and it runs LTE.
[00:30:34]and you, if you've got my phone 11. it has LT. It has CBRS support actually in it already, and more devices are coming. There's lots of IOT devices out there. Use it, the security cameras, as I mentioned. So you can actually build yourself a little private networks now, and we get a little bit more technical.
[00:30:53] An LTE network is not just the radio. We need a packet core. You need the back end of it as well. And [00:31:00] within the CBRS community, there's plenty of people have said, Oh, we'll try that in the cloud. So you can, we'll the sell you a bit of packet core. You can sign up for a service. So literally you and I, you can do it in your house.
[00:31:12] You, it, in a small office, you could do it in the warehouse. You can go buy a couple of radios, throw them up there and you are the mobile operator. Now we haven't gone to ATT Verizon or T-Mobile here. we've done all the work ourselves. And, you go in and configure your EPC and you say, yes, I want to turn my iPhone 11 on.
[00:31:31] And I'm the only user on this, that, or whatever it is. Right. but you are out running the network. so that's at its very simplest point. Now there's an interesting thing here. It is actually lightly licensed. It's not unlicensed.
[00:31:44] Matt: [00:31:44] I've just been reading a lot of news stories about CBRS auctions. So like if I could use it, why do I have to buy it?
[00:31:50] How does that work?
[00:31:52] Iain: [00:31:52] Right. So let's take an example. So it is wifi that is unlicensed, right. And everybody knows wifi gets [00:32:00] congested. Right. There's too much traffic. since my kids moved home for COVID, they tell me daily how bad our wifi is at home, but, you know, it's just my wife and I here. We don't get, so now what.
[00:32:13] LTE is. I said, he's good at lots of things, policy authentication, et cetera. It's good at traffic management and that packet core, it's very good at traffic management. So what the, FCC did, and this is very clever is, they created something called the SAS, and I never remember what the acronym means, but basically when you buy that CBRS radio and plug it in, it actually registers with a SAS.
[00:32:38] There's five of them. And Google has one Comscope, federated wireless, Amdocs, and Sony, right. And those SASs keep track of where your radio is. So, and it says, Oh, Matt, you've got a CBRS radio though, in your office, you can use channel one, [00:33:00] right? Your neighbor gets one and it says, Oh, you know, Matt's, neighbor's got one.
[00:33:04] You can have channel two. Okay. So we, it manages the interference between the devices. If you buy a third device and put it into your house, it'll go, okay. Now we'll give you channel three. And there's 15 channels. It's 150 megahertz, 10 megahertz each, which is great, right? Because now you get into this management of, and what it can do is, let's say there's a problem.
[00:33:28] Let's say your neighbors. Radio goes on the Fritz or something. It'll come to you again. Hey, Matt CBRS radio changed channel 13, please. And it'll do it. Right?
[00:33:39] Matt: [00:33:39] So these SAS products, their cloud services, it sounds like, and not to be confused with S a a S products, which
[00:33:46] Iain: [00:33:46] they probably are sold
[00:33:48] Matt: [00:33:48] is S a a S that was a good acronym.
[00:33:52] They picked. I like that. Yeah. yeah. so they basically does collision avoidance.
[00:33:56] Iain: [00:33:56] Yes exactly. Got it. And you buy your [00:34:00] CBRS radio. And actually some of the ways they've been sold is that when you buy that packet call you subscribed to a practical, it comes with a SAS subscription as well.
[00:34:09] So you radio, I mean, we're talking about a couple of bucks a year type thing. we're not talking thousands of dollars here, so, so everything's good. Well then they said, hang on a minute, you know, we've got 150 megahertz, right. And there's going to be some people who want. Priority access. They'd like to own the spectrum.
[00:34:26] And the FCC said, no, we're not owning anything here. This is all shared. And we're all, we're having a big luck here, come by out. Right. But what we'll do is we'll have, some of that spectrum, 70 megahertz of it. And we'll divide it into 10 megahertz slots, so seven channels and we'll sell priority access.
[00:34:46] So, and that's the auctions that they just had. And to give you a perspective on this, the. The pal priority access license, right? Seven of them per County in the [00:35:00] United States. And there are 2000 counties just over in the U S so a hundred. We've got the number come on. Was that 28? Yeah. 14,000 licenses is what it is.
[00:35:12]anyway, this, all of them, the one in LA, a PAL in LA went for 50 million bucks. Right. cook County in Chicago went for
[00:35:24]Matt: [00:35:24] all that COVID right
[00:35:30] Iain: [00:35:30] now you get to West Texas. actually one of the last licenses to be sold, was up in Napa, near you. Right. And, it went for 30, $40,000. I think something like that, some of the licenses in West Texas were literally hundreds of dollars thousand, a thousand bucks. Right. Because nobody lives there. yeah.
[00:35:52] Matt: [00:35:52] We're going to attach sensors to the,
[00:35:55] Iain: [00:35:55] yeah. So what we've so what happens is now I own a pal, [00:36:00] right? And, now I've got my CBRS radio. My access point and I say, yeah, get on my pal. When I'm using my radio. I get priority access. Okay. I get guaranteed 10 megahertz service. That doesn't mean that if you're on my, let's say we're both on channel.
[00:36:20] You come along on channel one that you get no service. the SAS moves you out of its channel. 13 moves you out of the way. Right? So, but I get priority access. If I'm not using my radio, then it goes into the general pool. Anybody can use it, but as soon as I
[00:36:38] Matt: [00:36:38] licensed all spectrum, that way,
[00:36:40] Iain: [00:36:40] it's funny.
[00:36:41] You should
[00:36:42] Matt: [00:36:42] say underutilized spectrum.
[00:36:44] Iain: [00:36:44] Yes, this, and this is what the FCC wanted to avoid was we didn't want people sitting on spectrum for years, speculating it. Yeah. To sell it on later or say, Oh, well, I'm going to deploy that in three years time when I build the cell tower or whatever, right. This [00:37:00] way the 150 megahertz is available, but then the pal users can get it now.
[00:37:06]I said those seven PALs for every County, the most you're allowed to buy was four. Okay. Okay. So the operators, Verizon and dish spent a lot of money, AT and T did not participate and the T-Mobile did. And, so Verizon's bought four licenses in a lot of markets and there's actually going to use it to supplement the mobile network, but other people who bought it were, Oxy, the, chemical oil, chemical company, Chevron, John Deere, couple of universities.
[00:37:38]the problem and we'll get, I'll get, I'll tell you one more thing. Problem is like, yep. So let's take, Berkeley where you are, right. That County is pretty expensive and, you know, it's pretty big, right? It went for a lot of money. So, Cal. [00:38:00] University of California up there. There was no way they were going to bid on the PAL because it was too expensive for them.
[00:38:06] Right. The universities that bought pals are in the middle of nowhere, frankly, or in small towns, they could afford it. but there's, what's called a secondary market and the FCC is encouraging this. So if you own the pal, you'll being encouraged to invite people to come and rent it from you, lease it from you.
[00:38:25] So Cal could have, you may have a football game one day, right? People will come into the university and want services, et cetera. So for the weekend of the football game, the university could lease the PAL. One or two pals from the local owners, right. security cameras. Maybe they want to use them between midnight and 6:00 AM.
[00:38:50] Right. Great. Let's go use a pal, things like that. So there's a secondary market developing of people who say, yes, I only need this geography. I don't need a [00:39:00] whole County. I just need my university campus. I need my warehouse. I need my hospital. so they've sub you can subdivide the County down and lease.
[00:39:10] Access to it. which is fascinating. We've never done that before. and there's a lot of people around the world looking at this model saying just as you did spectrum, doesn't lie fallow.
[00:39:22] Matt: [00:39:22] It goes to zero the next second. Right. If I didn't use that spectrum, it's not going to go. I lost it.
[00:39:29] Iain: [00:39:29] Yep. Yep.
[00:39:31] Matt: [00:39:31] Maximize utilization. so, so. all of this, you know, frequency, hopping and sa controlling and stuff like that. is that an example of a real time workload that someone might need to run at the edge because you've got so many devices you need to move around in so many frequencies that might be colliding and so much predictive work, or is that just something that's gonna run in the plane on the centralized cloud?
[00:39:53] Iain: [00:39:53] No, it has to run in the cloud because remember I said there was four or five sasses. Oh,
[00:39:57] Matt: [00:39:57] but in an edge environment, in the cloud, you [00:40:00] know, like you said, you know, Amazon is offering edge nodes. Yeah.
[00:40:03] Iain: [00:40:03] No, they could do that. Yeah. Yeah. we could. The other thing that happens is you start say, okay, well, I'm going to have a private network.
[00:40:09] I'm going to use CBRS. Right. I've got a warehouse and I want to do all this. I want to, I've got my cloud, but I want to do some local processing immediately. off of that traffic off of that radio. Right? So what we've seen is a lot of the, CBRS vendors saying, Oh yeah, by the way, you know, you'll need that edge.
[00:40:30] It's almost like private network and edge becoming the same.
[00:40:34] Matt: [00:40:34] Yeah. I'm seeing that we've actually invented a term that seems to be catching on. That's called near prem. And this is the idea that like, if you're within 50 kilometers of a co-location data center, you know, not only could you put all the intelligence to run your private network radios, but you also could put.
[00:40:52] Servers next to it to run your compute workloads. And in fact, Amazon could put servers next to it, to run your compute workloads. [00:41:00] If you wanted to do that. And that removes the, you know, the headache and operational complexity and time to market of, you know, dropping a data center on the factory floor and managing it and all that.
[00:41:10] So it's a really interesting. You know, when I first started getting an edge computing, let's say four years ago, five years ago, this isn't really what people were talking.
[00:41:19] Iain: [00:41:19] and I think
[00:41:19] Matt: [00:41:19] all those other applications will happen. I mean, you know, some of them are super fanciful, but a lot of them, you certainly will see because the technology makes sense.
[00:41:27] But like these, like, like you said earlier, sometimes these really prosaic use cases, save money or more convenient or more reliable. are the first uses of edge compute?
[00:41:39] Iain: [00:41:39] Yeah, I mean, if we go into the mobile, go back to the mobile operators, the way they configure their networks and they're architected, you know, we talk about virtualized, ran radio access.
[00:41:50] That works right, and well, virtualization means separate software and hardware to run them on. I mean servers to run them on. So [00:42:00] now I'm going to put some servers out at the base station to run the ran or the radio or compute.
[00:42:07] Matt: [00:42:07] So that's really interesting. Yeah. I mean, you know, people put it in these clever phrases, like, you know, you don't need 5g to edge computing, be new edge community 5g.
[00:42:15] And like 2% of the audience gets it, but that's exactly what it is. Right. it's like, You have this very complex network running very complex software on, you know, pretty beefy servers. And those things need to be located in something that looks like a data center. And that data center needs to be within 15 to 20 kilometers of the Radiohead.
[00:42:35] Yeah. And that's edge computing
[00:42:37] Iain: [00:42:37] compute. We've moved the workload, right?
[00:42:41] Matt: [00:42:41] Yeah. So, so yeah, so it's really fun that you brought up these, you know, like you said, these really prosaic examples. and then these far off. You know, AR VR, drones, all this stuff. what are some of the exciting use cases you're seeing emerging or you predict will emerge in the [00:43:00] relatively near term, say the next, you know, within the next 18 months?
[00:43:03] Iain: [00:43:03] Well, the one I'll answer the question and then I'll contradict myself. I always felt the big driver for this would be gaming. that I would put some edge compute into a mobile network. The operator would say, Hey, you're a serious gamer. because loads of people play games on mobile devices. and there's no reason not to have my son is upstairs right now using 5g LTE in our house for his primary connection.
[00:43:31] Cause he says the wifi is too slow. so why not play games on that? It's been, people have Apple TV and stream. Yeah, it was things like this as well. So, I always thought the gaming one would be really good that you would have, and you do want low latency and Google stadia, is a good example of an architecture there.
[00:43:49] They announced a year ago now. 18 months.
[00:43:52] Matt: [00:43:52] Yeah. Yeah. Where you're rendering it. Server side and pushing
[00:43:55] Iain: [00:43:55] pixels,
[00:43:57] Matt: [00:43:57] reliable, fast, low latency pipe,
[00:44:00] [00:43:59] Iain: [00:43:59] right. Yes. And when you look at what they, what their requirements were, it looks a lot like LTE with edge compute or 5g with edge compute within the mobile network.
[00:44:11]I kind of imagine the next version of Fortnite would have. AR VR goggles and all these teenagers running around doing this stuff. And maybe that's not that far off given we'll have to separate out, you know, the business issues they're having right now. but I thought that would be a good one.
[00:44:29]So, whether it is or not, we'll find out, maybe there's others, the other one that's coming up and I keep hearing about, and it makes a lot of sense to me is emergency services. the, and this, you do need a mobile network for it and you need edge compute, but, you know, there's an accident.
[00:44:46] There's car accident. you're in the back of the ambulance getting triaged by the doctor at the emergency room. For that connection to that ambulance, you want high definition, high bandwidth, low [00:45:00] latency, guaranteed service. I don't want any problems. Right. As you're driving through the app, the ambulance is going to there.
[00:45:09] That connection could be $500 for a gig. Okay. You'll probably pay it. And frankly, you don't want some teenager playing fortnight. Bandwidth
[00:45:22] Matt: [00:45:22] will, hopefully my insurance company will pay it, but yeah,
[00:45:25] Iain: [00:45:25] insurance company's gonna pay it. Somebody is gonna pay that. It's a high value connection. So when I look at things that when we've seen things take off in technology, they're either one of two things.
[00:45:35] They're either huge volume mass market, low cost. Right, or they're very specific. Super high requirements expensive, right. It takes a while to build the middle ground. Right. and I think the healthcare one is really good because it's monetized today. There's a whole. [00:46:00] Everything's in place to do it.
[00:46:01] And some people have started doing these things. It also plays in, there was an article the other day, the FCC is actually looking to include, to improve bandwidth and tele-health to rural communities because hospitals are closing in rural communities as the population declines. And frankly, with COVID the irony is we've lost a lot of hospitals through COVID.
[00:46:25] Because they can't do those routine surgeries that pay the bills. Right? So now you've got communities with no local hospital. Well, now you're going to have a remote type situation. And everybody says, well, you know, real robot surgery. No, it's pure tele-health, but again, there's going to have to be capability and server at the edge.
[00:46:50] To guarantee those services. Right. it's not just a case of give me connectivity. I've got to have guarantees and be able to do traffic flow and all those [00:47:00] things you talked about. Right?
[00:47:01] Matt: [00:47:01] Yeah. That's really interesting because, you know, you look at like your emergency service example. That is another example where the edge computing is kind of meta in the equation.
[00:47:10]because really you just need a network connection between the. The device that's in the ambulance and the hospital. but when you virtualize the network and the traffic management, Yeah, you need a lot of edge compute to run that on. Yeah. That's, it's really interesting, you know, that it actually brings up a, another yeah.
[00:47:31] Interesting, trend that I've noticed, recently, and that is the recognition that one of the biggest challenges of edge computing, at least on the infrastructure side now, not so much on the other side, but the service operator side of the network, is that it's not just about. The compute. In fact, the network actually is more important, you know, even in my business, you know, we talk about our kinetic edge and we've always described it as at data centers with a [00:48:00] network.
[00:48:00] And now I talk about it. Most of the time as a network with data centers attached, you know, you gotta put the networking equipment in something. And so I mentioned if you've got any opinion on
[00:48:10] Iain: [00:48:10] that. Well, as somebody used to say to me, It depends on who's doing the selling, right? If you're the mobile operator, of course the network is the best thing and the applicant is actually, yeah.
[00:48:24] I'll give you a really good example of this. and kind of answer your question is enterprises enterprise look at private networking. They don't look and say, gosh, I need a private network. What they do is they say, no enterprise has ever said, I need edge compute. What they've said is, Hey, I've got to do this.
[00:48:41] I've got to fix this problem. Right. And that problem means, Oh, well, we've got an application that can do that, but we need to run it locally. Oh. And we need a connection. We need a reliable connection. So the whole solution becomes the application, some compute somewhere or multiple places [00:49:00] and connectivity, right.
[00:49:02] That comprises the solution that the CIO signs off on. Right. So now the operator will tell you it's the connection. That's the value. The application developer says, it's the application. The edge compute guy says no. I've got the servers. It's that that's the value, but the CIO looks at the entire solution.
[00:49:21] So I think it's the answer is yes, we need all three. Who integrates that. Right. Who brings that all together is an interesting question. And it's the, answer's going to be, some operators are doing this, Deutsche. Telekom's got, T systems, right? They do a lot of things like this one
[00:49:41] Matt: [00:49:41] mobile edge X is a big investment of yours.
[00:49:43] Iain: [00:49:43] Yep. Yep. and it could be different people. It could be Amazon web services. It could be IBM. it could be, you know, actually I think one of the interesting ones is talent companies. I know that real estate guys, but they sit, they connect to a lot of infrastructure. [00:50:00] so it's a very interesting place, right?
[00:50:02] Matt: [00:50:02] I mean, you know, crown castle owns something like 75,000 miles of. Metro fiber
[00:50:11] network company in addition to a tower company.
[00:50:14]Iain: [00:50:14] yeah. But the answer is yes. All of the above, is the network critical show is edge compute, critical show. Is the application critical?
[00:50:21] Matt: [00:50:21] Yes. Yeah. So I realize we're at the top of the hour,
[00:50:24] yeah. So one of the things that I noticed you're working on is the 2030 project.
[00:50:28] And I don't know what it is, but it certainly sounds intriguing. Can you tell us what the 2030 project is?
[00:50:33] Iain: [00:50:33] Yeah. So, okay. So many over the many years I've been asked many times we chair conferences of analysts and we've been on panels. You and I have been on panels, all these things, and we've been asked, well, why don't you do your own conference?
[00:50:46] Why don't you do your own trade show? Right. And, I look at those things and we can do the content. We've helped people put content together, but you know, when you do those things, you got to deal with the hotel and food and the badges, and right.
[00:51:01] [00:51:00] Matt: [00:51:01] It sounds like you just said, you know, why don't you put a knitting needle in your eye?
[00:51:05] Iain: [00:51:05] Yes. Exactly. Yes. Just give yourself a root canal. Actually, I had a co I had a COVID test the other day. That's the same thing. This thing is way up, right.
[00:51:19]anyway. So, friend of mine, Tim downs, he organizes conferences, where the content guys he's the organization guys. So we started the 2030 project. We actually did a trial run on an event last year. And the goal is really to look at starting events and conferences. We actually got one off the ground in March 12th in DC.
[00:51:39] On the CBRS pal auctions, actually. And we did it in March because it was right before the quiet period started. And we'd discussed all these issues. You and I just talked about and the secondary market and things, three days later in the DC shut down, we got in right before things took a dive. Right. but [00:52:00] we were kind of looking at it saying, look, There are lots of different things to be discussed.
[00:52:05]enterprise 5g is one that we've done private networking. we're looking at a variety of different things. Digital infrastructure is actually interesting to us and look at building communities and shows around it. Yes, we started that. The 2030 project is defined as a, as defined as a tradeshow.
[00:52:25] You know, conference company. So yes, we did start a trade show company just before the pandemic hit. And no, we can't have any trade shows. So we've been doing webinars and things and, because we decided that 2021 was too soon, 20 twenty-five seems like its only 5 years, 2030. We're looking ahead, you know? so one of the things actually we are looking at, and I'll tell you a little bit more is, digital infrastructure, right?
[00:52:48] That's what you and I have been talking about today. It's not just the tower. It's not just the radio. It's not just the data center is not just the fiber. It's all of it. And integrating all of [00:53:00] that to give that solution and support everything we're talking about, is, you know, is investible today.
[00:53:06] There are investors doing these things right now, putting billions into bringing those people. You mentioned crown castle, having towers and fiber. I think they've also got a data center company, right? So. Yeah. So things like that we're not just talking fiber Italia, you've still got the whole integrated solution.
[00:53:25]and so those types of things we want to get to, which we try not to do things that anybody's done, with trying not to be boring. it's been fun. It's been nothing like we expected because of, the pandemic. but yeah, that's what we're doing.
[00:53:38] Matt: [00:53:38] Yeah. So, before we wrap, I wanted to throw out a question to you.
[00:53:42] That is one of the favorites that you like to ask when you were, I'm gonna ask you a little, I'm gonna ask it a little differently. speaking to people in this industry and the edge computing that are, you know, banking their career on this or their company or their investors [00:54:00] money or their own money, or some combination of those,
[00:54:03]Iain: [00:54:03] what should
[00:54:03] Matt: [00:54:03] be keeping me up at night?
[00:54:07] Yeah. what should I, w what could mess this up and put us 10 years out, or like, or that I should worry about? Like, what should I be worried about?
[00:54:15] Iain: [00:54:15] I think you should be worried about. and I'm not saying that you've specifically, I know what you guys do, but I think you've got to really look at who your customer is.
[00:54:24] I hear a lot of people say, well, we're. We're going to help support the mobile operator cause they're going to be in this. I think the problem with edge computers, it's very difficult to pick winners and losers right now. It's very difficult to say. And actually, I wouldn't even say mobile operators as a whole.
[00:54:40] I could say that some operators are going to be successful with their strategy and others will not be in the involved. So picking winners and losers is difficult. the hyperscalers have. Obviously huge scale, massive resources reach, et cetera. and you know, as we've talked about [00:55:00] standards don't mean that much to them sometimes.
[00:55:02]so I think it's the, it's almost like people we've kind of started out this. We said, well, what is edge compute? Where's the edge. I think people, who've got a very rigid view of the edges here. And I'm going to do this and sell it to these people. That's a problem. I think you've gotta be a much more fluid with that and say, well, you know what?
[00:55:24] This customer, the edge is over there for this customer. The edge is over here. And for this customer, it's actually completely different. and. You gotta be really fluid and flexible. probably like we haven't been before as an industry. And I think this is why the operators will have a problem with this because they're not flexible.
[00:55:42] Telcos are not flexible. Right. commercial real estate, not flexible. but somebody is going to have to do something weird and you'll say, well, there's that edge? Well, actually to them, it is, and it makes sense. So, that's what I guard against is the, trying to put it all in a bucket and trying to define what [00:56:00] it is.
[00:56:01] There's plenty of it. You know, these people, there's plenty of executives in the edge, compute industry, stand up every presentation and say, this is the way it should be. Right. This is what it is. And I. At that point, you're like, okay, I'm not sure you're right. yeah, you gotta be a little bit more varied than that because I mean, just look at the conversation we've had what we thought this was going to be when you started doing this three, four years ago to what it's turned out.
[00:56:27] It's not the same thing
[00:56:29] Matt: [00:56:29] pretty much. So we're definitely in the early days of it. Interesting. I mean, I, you know, and I went through the birth of cloud. and it's a very different thing than, I mean, there's a continuous story. And when you look back, you can trace all the threads and it makes sense, but you know, if Jeff Bezos had never launched Amazon web services, we'd be in a very different
[00:56:47] Iain: [00:56:47] place.
[00:56:48] Yeah.
[00:56:48] Matt: [00:56:48] So I totally get that. so yeah. Thank you so much for spending time with us today and, offering some of your insight. If people in the audience want to find you online, what's what are the [00:57:00] places they could go to?
[00:57:01] Iain: [00:57:01] Oh, well, you can Google me and they'll probably spell my name wrong. So I A I N I have two eyes in my name.
[00:57:10] I'm a real Iain, Gillott, G I L O T T. so two L's, two T's in Gillott, and everybody gets, you're going to get the first one wrong
[00:57:19] Matt: [00:57:19] and I'm sure somebody will spell corrected in their search in the search engine. Yeah.
[00:57:22] Iain: [00:57:22] Yeah. but then we're online as well. It's easy. It's IGR dash inc com.
[00:57:28] That's it.
[00:57:30] Matt: [00:57:30] Are you active on social media at
[00:57:31] Iain: [00:57:31] all? LinkedIn? Sorry.
[00:57:34] Matt: [00:57:34] Okay, awesome. Well, Iain, thank you so much for joining us today. And, I hope we get to do this again sometime. Yes.
[00:57:40] Iain: [00:57:40] Great. Thanks.