This episode of Over the Edge features an interview between Matt Trifiro and Dalia Adib, Director, Consulting at STL Partners. In this episode, she explains what telecoms are doing to expand use of the edge, how it can help companies navigate the digital transformation, and what are some of the best future use cases for its implementation.
This episode of Over the Edge features an interview between Matt Trifiro and Dalia Adib, Director, Consulting at STL Partners. In her role, she manages and leads teams to help telcos and their technology partners evaluate new opportunities. Dalia founded the Edge Computing Practice at STL Partners in 2019, which included building their own proposition in the space and establishing the company as a leading adviser for those looking to grow business opportunities in the edge computing ecosystem.
In this episode, Dalia explains what telecoms are doing to expand service through use of the edge. She defines and outlines where the edge is, and what it is doing to help enable companies to navigate the digital transformation. Dalia also talks about how organizations are thinking about the edge, and what might be some of the best future use cases for its implementation.
---------
---------
“Where the edge becomes interesting is when it becomes really dynamic. In that, we'll have distributed environments where workloads and things can move across each other, and that will further emphasize what is true edge - what is just simply a, a static form of compute or some static compute with an application dedicated to it in an edge-like location.”
“I think generally how we distinguish traditional, let's say data center infrastructure or on premise compute with the edge is at the software layer. Probably more than anything in that for it to be true edge computing, it should be using cloud native software infrastructure, things like containers. Not everything is yet container based, but it should definitely at least not be about monolithic systems that are inflexible.”
“I think telecom operators are still figuring out their role. One thing that they are doing though, whether it's with edge or different services, they're trying to become more digital solution providers in general.”
“There just needs to be more edge infrastructure out there. Otherwise, we're going to be stuck with either the chicken or the egg, whichever way you look at it.”
---------
(02:15) Getting into Technology
(04:45) Journey to STL Partners
(07:15) All about STL Partners
(10:15) Starting the Edge Practice
(13:15) Distinguishing On-premise computing from Edge Computing
(16:45) How Companies are Thinking about Edge
(25:15) Telecommunications Ecosystems
(27:15) Defining Mobile and Multi-Access Edge Computing - enabling vs providing, mobile, etsi layer
(33:30) What is Local Breakout
(36:15) Use Cases for the Edge - video analytics
(43:15) Trends in Neutral Host and Shared Infrastructure
(51:45) Important Edge Technologies to Push
--------
Over the Edge is brought to you by Dell Technologies to unlock the potential of your infrastructure with edge solutions. From hardware and software to data and operations, across your entire multi-cloud environment, we’re here to help you simplify your edge so you can generate more value. Learn more by visiting DellTechnologies.com/SimplifyYourEdge for more information or click on the link in the show notes.
--------
Connect with Dalia on LinkedIn
[00:00:00] Narrator 1: Hello and welcome to Over the Edge.
This episode features an interview between Matt Trifiro and Dalia Adib, Director, Consulting at STL Partners. In her role, she manages and leads teams to help telcos and their technology partners evaluate new opportunities. Dalia founded the Edge Computing Practice at STL Partners in 2019, which included building their own proposition in the space and establishing the company as a leading adviser for those looking to grow business opportunities in the edge computing ecosystem.
In this episode, Dalia explains what telecoms are doing to expand service through use of the edge. She defines and outlines where the edge is, and what it is doing to help enable companies to navigate the digital transformation. Dalia also talks about how organizations are thinking about the edge, and what might be some of the best future use cases for its implementation.
But before we get into it, here’s a brief word from our sponsors…
[00:01:02] Narrator 2: Over the Edge is brought to you by Dell Technologies to unlock the potential of your infrastructure with edge solutions. From hardware and software to data and operations, across your entire multi-cloud environment, we’re here to help you simplify your edge so you can generate more value. Learn more by visiting Dell.com for more information or click on the link in the show notes.
[00:01:21] Matt Trifiro: Two years ago when I started the Over the Edge podcast, it was all about edge computing. That's all anybody could talk about. But since then I've realized the edge is part of a much larger red.
Pollution. [00:01:40] That's why I'm pretty proud to be one of the founding leaders of a non-profit organization called the Open Grid Alliance for oga. The OGA is all about incorporating the best of edge technologies across the entire spectrum of connectivity. From the centralized data center to the end use devices, the open grid will span the globe and it will prove performance and economics of new services like private, 5G and smart retail.
If you want to be part of the open grid movement, I suggest you start@opengridalliance.org where you can download the original open grid manifesto and learn about the organization's recent projects and activities, including the launch of its first innovation zone in Las Vegas, Nevada.
[00:02:06] Narrator 1: And now, please enjoy this interview between Matt Trifiro and Dalia Adib, Director, Consulting at STL Partners
[00:02:15] Matt Trifiro: Hi Dalia. I've really been looking forward to this interview. One of the things that interests me the most is, you know, if you think back to your, I don't know, your childhood or whatever it was that you became first interested in technology, like what stands out, what, like, how'd you first get into technology?
[00:02:27] Dalia Adib: That's an interesting one, and it always sounds like an interview question, which I've actually never had to answer, but I was trying to think about it. I think I got into it as a teenager and you know, back then, The beginnings of social media and there was, there was actually this platform which allowed you to create your own website.
I remember it was called Pixel, p i c, Zeto, and it was kind of a, you know, creative outlet, a predecessor Facebook place where you could, but you could essentially create your own website, use html, et cetera. So I think that was the early beginnings. I was one of those slightly more nerdy users who took it to its full form, wanted to use all the features, learned how to use Photoshop, all those types of things.
So that was probably my, my first inroad into technology. And back then I had no real idea of what I wanted to do, but I knew, I, I liked the idea of using technology for creative reasons and to [00:03:20] create, you know, to, I guess, Probably didn't think of it that way back then, but to create an experience to engage people in a different way.
So that was the early beginnings. What, you know, what I'm doing now is probably quite different. But even in my role as a consultant, I know we have a bad rep of being a bit of a PowerPoint slide machine, but we do, I think, you know, now consulting's changing and a lot of it is about how do you communicate your ideas in a new way using new platform.
So in some ways, like this is my first time thinking about it, but it is quite cool that early beginning of what I was doing as a teenager is still manifesting itself and to what I do today.
[00:03:53] Matt Trifiro: That's neat. And, and what was your, your schooling like, What did you, what did you study?
[00:03:58] Dalia Adib: So, uh, university, I didn't do anything related to technology.
I did politics and economics. I didn't either. Yeah. So I, I came out of school. I mean, in the UK I guess it's different to the US where you have to choose what you want to do probably quite early. So we have three years of university and you, you need to pick one or two subjects. Usually you don't get to do much more than that.
And I was sort of like, I was okay. I was fairly good at maths and I liked history and I thought politics and economics was somewhere there. And the reality was I just, I wanted to keep my options open. So yeah, I did that, didn't, didn't do what I probably set out to do originally, which was, had big ideas of going to international development and things like that, but came into technology in the consulting side and said,
[00:04:42] Matt Trifiro: yeah.
And, And what was your journey to STL Partners?
[00:04:46] Dalia Adib: The journey to stl. So I, I went to the London School of Economics and another place which can have a bad reputation of being a place where Goldman Sachs bankers are made essentially. So [00:05:00] going there, it was, yeah, interesting environment. There were, there were some really cool people, some great friends, but there were some serious, you know, basically people who wanted to get on with their career.
ASAP and make a lot of money, and that's great and that, that's a, that's a good ambition to have. No, no judgment there. But I, I was kind of exposed to honest to consulting as a, as an industry. It was something I'd never heard of before and how, how people described it was solve problems. You get exposure to different things.
You learn about the commercial world and you learn about different industries and often, I mean, not always with consulting, but often there's a technology side as well, so, I kind of thought that that sounded good and then came across STL partners in that search to find a consulting firm. And what I really liked about STL is at the time, and I mean still today, we, we position ourselves as wanting to encourage, we focus on the telecoms industry mainly.
I mean now it's more telecom slash tech with, you know, say the cloud and telecoms converging together. But it was always about how can you think about the business possibilities with new technology or how telecoms operators as technology based companies can innovate and can solve problems in other industries and other domains.
When I first joined, there was probably more of a industry trend around consumer digital services. So back then we saw the mobile operators launch, you know, their digital, What, what year was this? Roughly? How long? This is 2015? Yep. Okay. When I, when I first joined and yeah. Mainly it was like digital apps, I think back then.
Yeah, companies like Telephonica, some of the Verizon, I think, I can't [00:06:40] remember the name of it, but they had they've kind of video content app that they'd launched around that time. and they'd, I think either then or soon after they acquired Yahoo. So there was a lot around media and apps that, how to engage customers in that way.
Now, I mean, we'll probably talk about it more, there's more of a shift towards enterprise and new opportunities using technology, but to kind of deliver outcomes in, in the enterprise side. But anyway, long story short, so came, came across SDL partners. I like the idea that it was, we were focused on technology, but still in consulting, which is what I wanted to do and wanted to solve problems essentially.
Yeah.
[00:07:15] Matt Trifiro: SDL partners, when you think of like tech analysts, like Gartner and Forester and STL is, is quite a bit different. I it's pretty bespoke and, and niche. Can you help us understand like how big you are and what types of clients you, you
[00:07:27] Dalia Adib: tend to have? Yeah, yeah, definitely. So we do research and consulting and then the consulting is research based.
So there's, there's a bit of a gray area between the two. Anyway, we are now. Between 30 to 40 people. So we're, we are literally just in a That's nice size. Yeah. A bit of a fun size. Yeah. It's, it is a, it's an interesting side cause we're now growing quite quickly. Touchwood, we grew quite a lot as well during the pandemic and sort of in Yeah.
In that stage of scaling up quite quickly. Which is cool. Which is interesting in terms of, Yeah, so I mean, like you say, there are big, huge analyst firms like Gartner. We, we, we a specialize in an industry, mainly telecommunications and then some surrounding areas. We also focus on less around like core telco business models and core telco services, but really about what can communications do in other [00:08:20] industries or how is it changing, How is the technology changing?
What should you mean? Like how does it affect logistics,
[00:08:25] Matt Trifiro: manufacturing? Yeah, exactly.
[00:08:26] Dalia Adib: Healthcare. Okay. All that stuff. I mean it's not, and also it's not, I guess it's not just telecommunications. Like I said, it's kind of telecommunications and compute and cloud and all these things are adjacent to, to connectivity.
So yeah, that's us. And research, like our research business subscription model. So similar to the, to the likes like Gartner you mentioned. And on the consulting side we work with there, we do a strategy consulting. So we work with clients kind of hands on with helping them to develop a business strategy in a particular area.
And partly because of the industry we work in, it's often, oh, there's this new technology that we're, you know, hearing about what can we do about it? How do we monetize it? And we're part of our role is trying to shift that mindset to really think about, okay, well what's the customer problem you're trying to solve?
Say like you mentioned like manufacturing, like what, what are they trying to do? Then think about the technology pieces, and that's quite a big shift in the industry, particularly in mobile operators can get kind of obsessed with cellular technology and know 5g, but. Really, it's trying to encourage thinking about a broader solution set or portfolio of tools that telecoms operators and other technology companies should have.
And those need to be tailored to meet some certain customer problem.
[00:09:39] Matt Trifiro: So yeah, I'm, I'm fond to saying that enterprises don't buy edge computing. Yeah. Like it scares them, to be honest. They buy solutions to their problem. They wanna buy like robotics as a service, or the kinds of things can be delivered with edge computing, but they don't actually, they don't actually need edge computing necessarily.
Yeah. So you, speaking of edge computing, you actually founded the edge practice [00:10:00] at stl. Could you tell me a little bit about that and what the Edge practice does and how you, you stand in the industry? Yeah, it's,
[00:10:05] Dalia Adib: it's been a really cool, let's say, project, I guess within SDL Partners. So, coming from a telecoms background, I would say this was in like 2 20 17 or so, there was a lot of discussion around this idea of mobile edge computing, partly driven by.
Network virtualization, which is essentially moving the proprietary equipment that's used to run a net telecoms network to more server based, software based networking. That's essentially what it does. And what that means is you have to build the facilities and the infrastructure, which are more data center like to support those.
So the telecoms, what the mobile industry was saying is, listen, we've we're building these, they were essentially edge data center facilities. We were gonna be putting. Service for our network functions. We know that there are gonna be applications that are gonna need some form of edge computing. Why do we play a role in, in kind of seeding that market in, in, in being part of that market and essentially using some of those facilities and the infrastructure for other applications like robotics as a service or cloud gaming or whatever it might be.
So it was originated with mobile edge computing and what we were hearing from our clients. Like questions like what are the use cases for that? Why would enterprise or a, you know, even an app developer isv, a cloud gaming developer, whomever it might be, why would they use that infrastructure? What is really the benefit?
What's driving this? Like, are we trying to provide a solution to a problem? What is, what are the actual problems here? So a lot of it was around, you know, understanding use cases. And then the other thing was, you know, as a telecoms operator, how, what is our [00:11:40] business model here? I mean, are we just going to be, I mean, not just, but are we going to be a sort of an edge colocation data center provider?
Or can we do more? Can, can we really play a role and how does that change with 5g? That was the whole beginning of the edge practice and just to speed up the long story, maybe we, we basically decided to formalize it. We were quite early as a research consulting company to, to really put forward some thought leadership in that space.
Put, you know, we had this kind of five edge business models paper, which got a lot of traction and then slowly did a bit more work. Some of it mapping the ecosystem. And I remember talking to you guys probably three, four years ago now about where you found the ecosystem, that sort of thing. Then we created a research service around it called the Edge Insight Service.
So that now has a market forecast around edge revenues across the value chain, different use cases, different countries and regions. We also have a forecast around those telco, the edge data center. I was talking about how much capacity there'll be of that. So there's a few things we've kind of brought that all together in the practice.
Why I said it was inter an interesting or a cool project is that it's kind of, it's been a catalyst for us to also expand beyond telecoms quite a lot. So a lot of our clients now are. Data center operators or data center businesses, cloud providers. Yeah.
[00:12:50] Matt Trifiro: Everybody that's looking at that. Yeah. That's really interesting.
Yeah, so that's,
[00:12:53] Dalia Adib: that's been really cool.
[00:12:55] Matt Trifiro: As you mentioned, we've known each other for a few years and, and I started my Edge journey about five years ago too, and I latched onto that. In fact, helped establish it. I created the State of the Edge, which is now a Linux Foundation project. My podcast is named Over the Edge, but I've been a little disappointed by Edge computing as a term, not, not as a, not as a solution, not as a part of a solution.
And I'm interested in your, your thoughts on this. A lot of entrance when, when Edge Comput started to take off, a lot of companies [00:13:20] joined in and I, they started calling what I think of as on premises computing, as edge computing. And so I'm curious how, how do you distinguish on premises computing, which we'll be doing for decades out at the edge from edge
[00:13:33] Dalia Adib: computing?
Yeah, it's a good question. And just to be transparent, as a consulting firm, I feel like that is an even harder question when we have potential clients who. There's, there's a bit of a balance where you're like, I don't know if you are, you're telling us you're an edge company. I don't know if you are, but we also wanna do work with you.
So yeah, it's definitely a delicate balance. I think generally how we distinguish traditional, let's say data center infrastructure on or on premise compute with the edge is at the software layer, probably more than anything in that, for it to be true edge computing, it should be using cloud native software infrastructure, things like containers.
Not everything is yet container based, but it's definitely, it should definitely at least not be about monolithic systems that are inflexible. So I, I think that, and it's, I know it's not the most punchy definition, but I think that's the key differentiator between some of those traditional environments and what is true edge environment.
And although I think as a industry, and we're not there yet, but the real benefit or the, where edge becomes interesting is when it becomes really dynamic in that, you know, we'll have. Distributed environments where workloads and things can move across each other. And, and again, that, that will further emphasize what is true edge and what is just simply a, a static form of compute or some, you know, Yeah.
Basically some static compute with a application dedicated to it in [00:15:00] a, an edge like, uh, location.
[00:15:02] Matt Trifiro: Looking at its continuum is actually a pretty good answer. One of, one of my colleagues actually said that the difference is in, in, in his mind was the control plane extends back to the cloud. Mm-hmm. , which I, I think is, is a partial good answer.
But, you know, the one that, that I've actually stumbled upon, which comes back to your original point, which is like what enterprises want, and that is the, the edge is the ability to deliver performance and security of on premise workloads with cloud economics and as a service with flexibility. Yep. And cuz that, cuz that doesn't, it gets away from the technology and really gets away, like, doesn't matter where the workload runs as much as it runs in a way that the enterprise.
Can think about it differently. Mm-hmm. , and I'm becoming fond of that. Yeah. That's, that's at least where I've gotten in my evolution. I think
[00:15:45] Dalia Adib: that is that, that's sort of what I was trying to say, but you've put it much in a much better way, because I think that's right. You know, when I was talking about with cloud native, Okay.
Those are the tools that will enable flex. I think it's the techno, you can move containers around and That's right. Yeah. Technological flexibility, but commercial flexibility, it's all, I mean, what we always hear from enterprises is they don't wanna be locked in. They don't want, you know, vendor lock in or even technological lock in, and it's all these sort of business drivers that are lending themselves to edge, essentially.
Yeah.
[00:16:14] Matt Trifiro: So you mentioned that in the early days when you first started the practice, the hill coasts, were trying to figure out, are we gonna build edge data centers? Are we gonna be data pipes? Are we gonna be a combination of both these, Are we gonna be cloud companies? What are we gonna be, what, what are you seeing?
Like I'm, I'm seeing companies really establish themselves by planning polls in this market. M my knowledge is mostly US based. Mm-hmm. . And I know you're based in uk and so I imagine a lot of your business is, is more global than, than my knowledge. What, what are you seeing? [00:16:40] Like where, where are the big like flag planting, where's that happening?
Yeah.
[00:16:44] Dalia Adib: With, with telecoms operators specifically, are you saying with Edge?
[00:16:47] Matt Trifiro: Yeah. Telecoms offers specifically?
[00:16:48] Dalia Adib: Yeah, I think so. One thing that's definitely materialized is there's, there's, I mean they, they think about Edge two domains. There's kind of, we call it network edge, but it's the tell these data centers that they own.
And then there's the kind of the on-prem edge, which is quite fragmented because you, you can go from like enterprise CP boxes, which are running like network services, firewalls, SD one, et cetera, adding some more compute to them and they become a, a form of edge compute. Or you've got what's happening with mobile private networks and Edge, again, like using essentially like a few servers to run a private mobile network for.
Enterprise and then you, you can run some applications there too. And then you've got the on-prem edge kind of data centers and, and small configurations as well. So I think telecoms operators are still figuring out their role. One thing that they are doing though, whether it's with Edge or different services, they, they, they want, they're trying to become more digital solution providers in general.
I think, you know, what we've seen with connectivity is that in a lot of markets their, their revenues are in decline. Like where as consumers we're getting bigger and bigger, like, I dunno about you, but I've got a a hundred gig data bundle per month and I pay the same amount as I did on my mobile subscription.
I mean that as I did like a year ago. So, you know, it's difficult for them. They're investing their network and they're not, they're not able to monetize it in that way. So they're looking at, okay, what kind of services do we bring? And I think they've, and I mentioned at the beginning, they've tried different things.
They've tried sort of all sorts of things. Media, gaming, manufacturing, [00:18:20] a lot of them are trying to double down on a few areas. Mainly based on like what their markets are. So, I mean, the US huge market for a lot of things, but it's, you know, definitely a huge market for safe for manufacturing.
[00:18:30] Matt Trifiro: Can you, can you talk about some specific examples?
Yeah. Where, where you're Yeah, like,
[00:18:34] Dalia Adib: Yeah, sure. So what I mean, one closer to home is Vodafone they're doing, what they've tried to do is bring all their bits and pieces, let's say in the enterprise business together. So they've got some iot they've been doing for a number of, They've got some, they've got edge computing, they've got this kind of private 5G network, they call it a mobile private network offering and, you know, general connectivity stuff.
And then what they realize is, okay, there's a, there's an opportunity in some of our, in Europe around manufacturing and they've got, they're trying to develop a sort of like digital manufacturing offering. And it will, what's, what's, so, what's really different is that they're including applications. So they, you know, it'll be kind of like video analytics, computer vision applications, totally different to their core business.
And in most cases they, they're working, they're partnering with software developers who are specialists to do that. But the idea is that they wanna be able to bring that end to end solution to their enterprise customers. And really, so the robot is
[00:19:28] Matt Trifiro: becoming. They're almost becoming system integrators
[00:19:30] Dalia Adib: in a way.
Yeah. And this is to us, like, this is a debate cuz they're, they're sort of like, we, we want to be the solution provider. We don't have the skills inhouse to be the, the, like the end to end systems integrator. So they're also partnering with the systems integrators to actually deliver it. So they think, I mean, this, this is one of the debates, but it's like sort of what, who leads that?
Who leads the interaction with the enterprise? And in some cases it's, it's whatever reason, more appropriate for an SI if they've, they've got the relationship and then the telco's role is more [00:20:00] bringing that, bringing the solution. In other cases it might be the telco leads the role and they then they will need to have a consultative.
they have a sort of consultative role to play, but they still need an SI because, you know, like just bring it to life. I mean, not speaking for Vodafone, but if you take a Vodafone like digital manufacturing solution, so say something like video analytics, it might be running on at an edge in a factory and it's being used to spot defects on an assembly line, for example.
You'll need, when, when that happens, you probably need an alarm that links in with the production line to stop it, for example. So there's like, there's some systems integration that needs to happen there or needs to integrate into the, the SCADA system or whatever it might be. So, and that's not like, and not to perver on the spot, but it's not telecoms, operators bread and butter.
So it's all, it's all around partnerships. So they're trying to do more, they. Build solutions, develop solutions with partners, play a different role. And I think that's with traditional enterprise, it's also with, you know, with things like gaming as well, Slightly different model, but it's how do we work with gaming developers to create something different in the us?
I think a interesting example is Verizon with events as well. So they've set up an. Kind of business now that's like targeting events and stadiums to be delivering edge compute 5G solutions to specific verticals. So do
[00:21:18] Matt Trifiro: you mean fix in a venue or like temporary, set it up live for a motorcycle race that's through the streets of, of Finland or something?
[00:21:25] Dalia Adib: Yeah. I mean I think now it's probably more fixed just cuz of, you know, like as a starting, but definitely that's the beauty of edge computing. Right. And I mean this isn't Verizon specifically, but I've heard of a No, I know of a case study where they're working with, with, with Formula One. Cause that's, those are some of the [00:21:40] interesting use cases.
Right. Or scenarios where you have say, a sporting event that moves around different countries or at least different areas within the same place. And you could stick it all on a track. You really could. Yeah, exactly. So I mean, and like broadcasting news that, that's really cool as well to be able to quickly spin up the, the, all the setup that you need.
All the, you know, the video broadcasting, not just the equipment, but all the software you need as well. Have it close to where the broadcaster is also, I think. One of the things these edge compute applications does is means you don't need a team of specialists who have to set things up. It's really about trying to make it software driven.
So those are the types of things that yeah, Telcos and others are trying to do in, in events and stadiums. I've also, I dunno about you and I to us, I dunno the specificities, but one of the key, let's say like verticals that's getting a lot of traction with Edge seems to be casinos as well in the US So I mean they've got a lot of video.
[00:22:32] Matt Trifiro: I mean we have a huge installation in Las Vegas and yes, the hospitality and entertainment in in there is a lot of cameras in use. Exactly. There's a lot of opportunities for video inferencing and they don't wanna have data centers on their hotel property. Yeah,
[00:22:43] Dalia Adib: yeah. And probably like the fact that it's like Vegas is, is a good example cuz it's so you knows like so congested, so many things you want to be used using shared infrastructure as much, much as possible doesn't make sense for each, if they need to have its own thing.
So yeah, some interesting stuff, but back telco, so I think that that's something interesting that they're doing. Plus they are exploring sort of infrastructure, the service models. So more like typical cloud like things you'd expect from aws, Microsoft, they're doing it in partnerships. So just taking Verizon of Vodafone examples, again, they've partnered with AWS to deliver this, you know, infrastructure service solution called or service called wavelengths.
We've got at and [00:23:20] t partnering with Microsoft. So there's a lot of these sort of like partnership hyperscaler partnership models happening as well where both parties are taking it to market together. Yeah, I think,
[00:23:29] Matt Trifiro: you know, it's interesting like, you know, when we, we, when I started my journey, I think there was this, this, this huge amount of fear in, in the telecom world about what the cloud fires are gonna do.
And I'm sure that still exists, especially with like Amazon announcing, you know, their private 5G service. But if you think about it, every time the telcos have tried to go into a broad cloud business, they've gotten their clocks cleaned. And that's because the developers are already working with the three big cloud providers.
And if you're want the richest. Pool of applications, that's where you're gonna find them. You're gonna find the developers that are writing for the existing cloud. And so what I'm starting to see now is, well, two things. One is service providers. I mean the developers of the application, right? Not the underlying infrastructure, are building these services.
on a cloud or on multi clouds, and they want distribution. Mm-hmm. and the clouds can't give them as much distribution as the wireless providers can. And the wireless providers, as you say, are interested in, well, how do I monetize this investment I made in my network? And so what I'm starting to see is companies and sometimes even third parties that aren't a telco or a cloud, are assembling these like end to end, these end to end systems that can be white labeled, gray labeled by a service provider.
And I think that's a, that's starting to get some traction in the US because then as a Verizon or a Dish or a T-Mobile, I can offer my customers an entire portfolio of services. Some of which maybe I've driven to create differentiation and others like I just have to provide because it's table stakes, right?
Like everybody's gonna want [00:25:00] video inferencing from XYZ because they're the best. Yeah. And so rather than trying to compete against xyz, why don't I just like bring them onto my platform? Are you, are you seeing that in
[00:25:07] Dalia Adib: other places in the world? Yeah. I think that's a really good point. And I think that's also been a shift in the telecoms industry where.
Let's say, you know, five to 10 years ago, the telcos, when they wanted to capitalize or capture this, these new opportunity areas, they always want us to do everything themselves. And it's not, I mean, it's not easy to, You have to totally shift your business model. You're going into a space you don't understand.
So I think that's been a real shift, which is we need to work as partners, and I think most of them do see their role as this like central point of providing digital solutions. They talk a lot about we need to create ecosystems. And I'm sure you talk about this in your podcast, but like competition always comes up as a topic.
You talk about the hyperscalers. I think there's still some fear, uh, when it comes to I not, well, rightly so. Yeah. I mean, yeah. Yeah. Like everyone was always like, Well, Amazon, you know, the, um, Yeah, the AWSs buying
[00:25:58] Matt Trifiro: up the healthcare companies, they may buy Dish. You know, I would be, There's there's some existential threats there.
Yeah. But at
[00:26:04] Dalia Adib: the same time, it's like, well, they also, they, the telcos hit from their customers, Listen, you know, we're a Microsoft house or we're a AWS house. And they also we're multi-cloud every enterprise. Exactly. And I think the multi-cloud thing actually is, is key because that's where telcos, they wanna be a neutral service provider.
And it's, it's like, and it, I think it's. It's quite complicated as a model, you have to be really slick at being able to provide a service to your customers and manage all these partners, right? Cause you're managing first there's a multi-cloud thing, which is at the infrastructure level. But then, you know what I'm talking about is they've also got this application ecosystem now they're having to manage and they don't [00:26:40] have the slick marketplaces that like AWS have yet.
That's where they, that's where it should get to. But
[00:26:44] Matt Trifiro: complicated. Yeah. I mean we, we put together a system in Las Vegas that was an end video inferencing system and there were no fewer than 21 partners involved. And everybody needs to get a little piece of money. There has to be a charging system that works.
It's, it's very, very complicated. Which means there's a lot of opportunity, right? A lot of opportunity for people to solve these, these really interesting problems. Back, back in sort of the core edge computing thing, I'm wondering if you can speak to, So one of the things that, that I find a lot of confusion around, are there.
Mobile edge computing, lowercase, mobile edge computing. And then there's mobile edge computing, uppercase et c standard. Can you explain the difference?
[00:27:21] Dalia Adib: Oh God. , uh, you know what, like two years ago we just, we were like, we just cannot use the term, I think it's mainly the access edge
[00:27:35] Matt Trifiro: right now. It's multi access.
[00:27:36] Dalia Adib: Excuse me. Yeah. Yeah. So it was the mobile industry came on. They were like, Oh, listen. Forgot there were loads of access technologies. Let's not, let's make mec multi access edge tech computing. But the, so the, the, the thing that is really complicated is that I've noticed each telecoms operator and different parts of the industry mean it to describe different things.
Yeah. So I think in general it does still mean mo it is still referring to mobile edge computing, aka edge computing. And these mobile, oh, sorry. In these edge data centers designed to run the mobile network just because of 5G being kind of a key catalyst in this. And some there are exceptions. We've got like Cox Communications in the US building, not building, but they, [00:28:20] they, they're putting an edge.
They've got Cox Edge, which is a platform which runs from their. Data centers and they don't have a mobile network. So it's not, you know, it's not mobile only, but this concept of mec, I think is still quite mobile driven. And yeah. And then so then the, Yeah, the challenges that telcos mean different things.
So sometimes they actually refer to MEC as the kind of the, this is getting a bit technical, but there's this 5G breakout mechanism, you know, which allows you to sort of, rather than go through the, the whole, let's say 5G network, you can break out at a standpoint, So it could be at like a customer premises, like an enterprise site manufacturing plant, whatever, retail shop.
Or it could be at the base station, for example. So sometimes it actually is not referring to compute, but, but this kind of breakout mechanism. And you've got, like, for example, at and t's MEC proposition is that is mainly that at the moment. So they, in a lot of cases, they're not, say providing compute, they're enabling edge computing, but they're not providing it.
Other telcos are talking about MEC as, as referring to them providing the edge compute. Platform underneath. So very confusing. Then you've got the atcl, then you've got the ATC layer, which is a standard, which is, which drove a lot of the, let's say, innovation in the industry four, five years ago. And to be honest, and this happens with sta, especially telecoms industry standards all the time, technology moves faster than the standards do.
And what we've seen is a lot of vendors and part technology partners have started to develop their mobile edge compute solutions, which aren't necessarily in line with the standards. The idea is that they want to make it as compliant as possible and kind of future proof so that when those standards are more advanced, they, they [00:30:00] do work to them.
But there's a bit of a, there's a kind of a timeline delay. I think even, Yeah, capital sort of, capital MEC of Yeah, definitely means different things. I think, and it, this is a challenge and I'm sure you find it being really deep in the edge industry, is it's not, it's not always easy. Like when I, whenever I speak to someone, I have to really say like, what do you, what type of edge computing are you talking about?
Where is, uh, how big is the, It's kind of annoying. We have to like, go through these precursor questions. Like, are we talking about a server? Are we talking about a data center here? Like really different environment. Like are we talking about in, in, in the network or is this a manufacturing site? Is this on the floor?
Is this like in the back room? Like what's happening? And I don't think, to be honest, we've. Quite like we, I think there are some times we're using more on-Prem Edge. Pretty standard. Yeah. But even we call it network edge. That thing I described about putting edge in the mobile operators, data centers, sometimes it's called Telco Edge, sometimes called distributed edge a lot of the times.
So yeah, we, we, we definitely need to standardize these terms as an industry. Maybe Matt, you and I, we
[00:31:03] Matt Trifiro: tried, I tried with the open blast against computing. I tried very hard
[00:31:08] Dalia Adib: and I think, I think yeah, it is, it is used and it, I think it's also, there's different ecosystems, right? Cuz yeah. Like you, the state of the edge stuff, is it with the Linux Foundation Edge?
Yeah. So, and that obviously has a community around it, but it's a different community to say Etsy and the European telecommunications style. That's right. That's right. It's like guy, everyone needs to talk to one another, but it's, Yeah. Maybe wishful thinking. Well
[00:31:29] Matt Trifiro: my, my personal opinion is it's all just gonna become the internet , you know?
Cuz you think about like, all these technologies just collapse into this. Like, it just, I just get this. When I need it. Right. And I [00:31:40] think eventually we're gonna get there. Well, a couple things on, on, on me. One of the things that seems absolutely essential to me is the ability to have an API to the ran.
And the reason for that is if developers are gonna build applications that have autonomic functionality, they can self-heal, They can automatically distribute themselves. They can like restart themselves in a low carbon envi, you know, say I wanna be in a low carbon routing situation, so I'm gonna move myself to these other data centers and things.
In order to do that and be able to deliver SLA backed services, you have to know the status of the rant. And that's always been a closed system. And I think one of the things that Etsy is doing is opening up the, the status to the ran. So for the benefit of my listeners who don't know what local breakout is, let me take a stab at describing it and you can tell me if I'm right or wrong.
All right, So today's mobile networks, let's say lt, the common mobile networks, and even the five, the new 5G networks for the most. The wireless signal goes to the radio, comes down and and goes. Even if it, if it wants to terminate locally, like you and I are sitting in the same room and we're using WhatsApp to talk to each other, that signal may go all the way back a thousand miles to some central office or some facility where then the data is broken out, becomes ip, ratable goes onto the internet, hits Facebook's WhatsApp servers, and then comes back and it has to go the whole same trip all along.
Local breakout. My understanding. Is that, as you say, you can, you can break that data out closer to its origin. Mm-hmm. , which means that then you can route it much more efficiently, especially in a local market. Like if you and I are sitting next to together, like intuitively you think, Oh, it goes to the radio and comes back to you and no, that, that's not what happens today.
And for, you know, humans interacting with WhatsApp, that's plenty fast. It's, it's a hundred [00:33:20] milliseconds, who cares? Right. But for a machine that's trying, trying to like send an alarm to stop the factory, a hundred milliseconds may mean damage. Yeah. Yeah. And so I think, How's that as a description of local breakout?
[00:33:33] Dalia Adib: I haven't described it in this way before, but that is exactly it. It's kind of, I mean, The network is, Yeah, we talk about traffic as in like signals and packets, but it is, it does work in a similar way to car based traffic, vehicle traffic. And it, I mean, it's kind of like, you know, when you turn the wrong or you take the wrong exit or, or you miss the exit even and you're stuck on a road and you have to go all the way to the next junction and come round.
Like you can think about the junction is almost that central office. So you talked about, you
[00:34:01] Matt Trifiro: know, Oh God, I love that analogy. Yeah. And you, Cause when you, that is so frustrating, you know, you like five more miles and turn around to meet. Yes.
[00:34:07] Dalia Adib: That's exactly what it's, And it's kind of like on those less constructed roads, you can just take a u-turn, you know, there's, you don't have the barriers in between the lanes and you take, so it's kind of the, the break, how is doing that is doing that u-turn right.
Where it makes sense too, because you just wanna turn back quickly. And I think in a, you know, you're talking about those applications that need it, it maybe the analogy extends where the, the longer the road, the more likely congestion can make, you know, there could be things that go wrong. Like you could have a, a tree fall down, I dunno, the middle of the road.
Or you can have. Too much traffic and it slows you down. And I think that is why you want to shorten the route. And that's why edge computing plays a part. Cause a lot about, a lot of edge computing is about shortening the route, reducing latency, aka so shortening the route from the packet or the signal, uh, the packet going from the device of origin back to where it, you know, needs to go.
And it's the same thing, you know, And [00:35:00] what you're trying to do is avoid having to go through, say the whole of the 5G network or the whole of the fixed network, whatever it might be, or even the
[00:35:06] Matt Trifiro: internet with its best ever routing, right? Yeah. Like, just to be able to control, you know, not just the, the latency, the delay, but the predictability.
Like how, you know, can you control a laser LA without knowing exactly the timing between those. And I, I think that that's the industry sort of coming around to realizing that. And that's actually what I, why I like the idea of saying, What edge computing is really becoming, to me, from a business solution problem is all the power and of on-premises, all the performance of on-premises, but having to worry about the on-premises stuff.
Mm-hmm. , because you can run a laser late from a server that's sitting on premises. People know how to do that, they just don't know how to do it in a way that's like from the cloud today because those, that technology doesn't exist. So back in the, the heavy early wishful thinking days of edge computing, everybody was spouting like autonomous driving, like edge computing is gonna be autonomous driving.
Now we know where that's gone. Right. Like that's not a use case for edge computing today. It may eventually be, but it's not today. Mm-hmm. , we're starting to see, you mentioned casinos and video inferencing and those sorts of things. Where where are you seeing these, this kind of convergence around verticals in use cases that are, are relatively near term, meaning today to, to 18
[00:36:14] Dalia Adib: to 24 months?
Yeah. Um, bit, bit of a plug to my recent article, which was I've, ie this article research is, you know, uh, where is money being made in the edge? And that was because. We, uh, you know, wanted to like focus the mind to answer that question. And, you know, um, I know you were saying, you know, you're not disappointed about edges in industry, but maybe the term, but there are people who are disappointed of edge computing at the moment and that there was a lot of hype and it's maybe not delivered on the hype.
And I [00:36:40] think that's part of it is because, say the suppliers or the service providers of the industry were, were kind of navigating which use cases, like you were saying, there's a lot of emphasis on autonomous driving. Well, there's a lot of things that need to come together for autonomous driving to exist.
So I think the first thing that's materialized is you as a technology provider, can't expect to get an industry like manufacturing. I know keeping bringing up or retail to completely change. Like it's basic business processes. So an example of that is like, let's say robotics in the manufacturing industry.
That was a, an interesting domain for edge computing and it still is. But if you're gonna take the, the smarts or the controls of a robot out of the robot itself and run it in a nearby edge compute and no allow it to be autonomous. There are a lot of things that have to, like the, the indu, the manufacturing industry has really high standards when it comes to health and safety.
They can't do that overnight. They need to make sure that a, a robot running from an edge can roam around, feel, won't, won't bump into another robot, won't bump into a human that's working there. It also be about changing, you know, the production line and how it works. Cuz today a lot of it, not all in all, you know, manufacturing scenarios, but there is gonna be people there, there are people doing tasks.
You can't autonomize everything overnight. So the thing, the first thing is, you know, what are some of the low hanging fruit when it comes to edge? And I think you were talking about, you know, How do you differentiate between Edge and maybe some traditional on-prem compute? A lot of it is that, you know, flexibility, scalability, and there are some applications that, um, enterprises have wanted to do, but it wasn't, you know, maybe it wasn't easy to deploy or there was a large overhead in trying to, to test it out because you'd have to maybe get [00:38:20] specific equipment or the commercial model attached to it was, you know, let's say CapEx based.
So you couldn't try it out and it wasn't as in service. You couldn't try it out and then scale it back. You really had to make a commitment up front. And when you have to make a commitment, then you have to make a big business case. So what we're seeing as some of those applications is, is stuff like video analytics.
So talked about like using video analytics for kind of quality assurance mechanisms and you know, in some cases it's completely automating that. But even as at the short term, it's about complimenting what say people are doing on the assembly line when their quality checking with a camera to do that.
And it's really about, rather than an in, like an individual human can probably check, I dunno, one. One item at a like per five seconds or something along those lines. Whereas, you know, if you use video and, you know, real time processing, you can do dozens of products that are going through an assembly line at a time.
So,
[00:39:10] Matt Trifiro: and honestly, that's not a very attractive job for a human. Yeah. Yeah. That, I mean now we, we we're, hopefully you need to create other jobs for those people and retrain them. But like, sitting there watching widgets go by on the, or eggs go by in the chicken factory is just, I, it just must be mind numbing.
Yep. And a really bad use of human talent that we could exploit otherwise. So, So one of my favorite use cases that hasn't really emerged yet, but I feel is, is ripe, is automated checkout at retail. And the, the reason I say that is retail was getting really damaged by e-commerce. Mm-hmm. and when Covid came around, At least in the us that was a big stimulus for a lot of the big box retailers to change the way they think about their product in their stores.
Like the stores became inventory hubs. Like, why can't Best Buy delivery me an MP3 [00:40:00] storage device as fast as Amazon or Faster? The Best Buys right here, Amazon's Warehouse is at the airport. Yeah. And I think they've, they've kind of completely revolutionized it and, and because I've gotten so used of not going into stores with Covid now that I go into stores, I just wanna pull my hair out.
Like even when they have the self checkout, like it's so long line at the self checkout cuz it's a little faster than the checker. Mm-hmm. . But it's still like I have to, It's a lot of work and it takes a lot of time. It's really my knowing at the Dallas Fort Worth airport, which is near me. There is an automated checkout store and I, I don't know who does it and why they're doing it, but one of the things that was most remarkable to me, well, the fact that it works, that was pretty cool, but there's a camera, like every two square feet on the ceiling, and there's a scale on every single item.
Every item has, has weight. So they know a combination of who you are by falling on store, and they know what you're taking off the shelf and they can correlate all these things and then charge you correctly as you walk out the store. Of course it's all stored because if you say you didn't do that, they've got it on camera.
They say, Oh yeah, well you're right. You didn't, you didn't buy that $10 item. We owe you $7. And I think about like the cost of the infrastructure to, to run that thing in a store and the headache of having to manage. All the, all your stores that have like little servers in them, like little data centers.
I mean, this isn't like a, a Windows server underneath the counter. This is like a rack of Nvidia GPUs. Yeah. That have to be cooled and conditioned and, or, you know, you might have and have, you know, a hundred thousand dollars of equipment there. And so one of the things that, that I see being really powerful and, and I imagine coming around, but I I haven't actually seen it yet, is I just pick a retailer in the us.
I don't know if they have it in the UK seven 11. So these little, like these little stores or maybe a thousand square feet [00:41:40] and they're everywhere. So in a given market, a large market like Dallas or or Atlanta or something, there must be a hundred of them at least. And I can't imagine, I mean, I can imagine being sound 11, I said, Gosh, look how more efficient we could make our, our experience if we did automated checkout.
But look how much it's gonna cost and what a headache it's gonna be. And yet if you can pull all that out into the infrastructure and deliver it as a cloud service, you get. There's all kinds of, of what makes it easier for someone, like someone potentially to buy is like, Okay, just hang the cameras and somebody else.
I pay two bucks per day per camera, or $2 per month per camera. Whatever the economics end up being, somebody else is responsible for that. And it, and it scales and it, it can scale across multiple stores. You get have one server that serves multiple stores. So I see that as a really interesting use case.
And what I wanted to take this though and get your opinion is what's really powerful about that in addition to just make it easier for a customer to buy. The, the fact that you're now starting to look at sharing infrastructure, so in the case of multiple seven elevens, you can just share across all the seven elevens.
But why just the Seven Elevens, right? Like why can't you look at automated checkout as a service delivered to not just the seven elevens, but the AMPMs that are part of the Arco gas stations and all these like little stores, maybe even big stores too. And then there's like this, so that's, that's like one company offering all these services and that's shared infrastructure, multi-tenant infrastructure, and you get all kinds of economies from that.
But also there's shared infrastructure in the kind of the neutral host sense, meaning lots of, lots of. Providers, whether they're telco providers or automated checkup providers can come and share the same infrastructure. How do you think about neutral host and shared infrastructure and the economics and what the trends are?
It's, it's a really fascinating topic to me and I, [00:43:20] I haven't seen it converge yet. Yeah.
[00:43:22] Dalia Adib: Yeah. I think it's same here. Like it's, it's quite, I think where I've seen, you know, where neutral host models are being picked up is really about on the, on the telecommunication side, mainly I've seen it in that, you know, if you are, like, say if you're a big hotel and you wanna get good coverage, they have these like distributed antenna systems.
Now sometimes they're looking at private 5G networks, and so you and you as a hotel, you've got a bunch of different users and you've got also your hotel guests. You need to have multiple carriers supported, and that's where a neutral host infrastructure comes into play. But to us, I think that that's been around for some.
It does. I think, I haven't heard specific use cases, but there's definitely been like exploration activity of extending that. So it's sort of, if you have a neutral host infrastructure per hotel, why not put, like, a lot of them have like keyless, like software based keyless authentication stuff for the hotel rooms.
You could run that application on the same neutral host infrastructure. Some of the, the maybe secure video security type applications run on there. I haven't, yeah, haven't yet seen it, to be honest, come to life. But that's, that's not because it's not happening, but just because you haven't come across it.
But that's where it is on the shared, let's say on the like edge data center shared infrastructure side. I mean, I guess this is, you know, where you guys play a big
[00:44:39] Matt Trifiro: role. I mean, it's part of our core model. Yeah. So I think about it a lot. But I'm interested in, in, in what you're saying.
[00:44:45] Dalia Adib: Yeah. And I think, at least from what I'm seeing, I feel like, and this is, this is probably a, maybe a European or UK centric view of the world, but there's more, there's more investment on the supply side than there is on the demand side.
So, you know, even the other [00:45:00] day we have, there's a lot of these infrastructure providing What do, what do you mean by that? So like, SupplySide demand side. So you've like, we've had companies who are even like real, real estate sort of businesses who, you know, are saying, Oh, we've got space for edge compute or for like some, you know, shared infrastructure, who's our customers?
So you it's that way around or it's, you know, it's like, oh, we've got, like you say, you know, I'm a, there are some businesses we've come across like I guess some edge compute, edge data center type businesses and they're like, We've got some, we're gonna build a, I think there was one that was building a edge compute infrastructure around the main motorway around London, the M 25.
And they were like, We're gonna have dozens of edge compute data centers, but not really sure why. So that's what I mean by, by this, like Yeah. Right, right. You might say more than the demand side. And I don't think that's, That's chicken and egg problem. Yeah. Yeah. And that's like, you know, where we're at with edge in general.
And I think, yeah, I think it's just, just a case of the demand side. AK the customers aren't aware of this as a possibility, but I like, at least my feeling is that it's, it's pretty early. The, the other thing is that the telecoms operators, our friends are telecoms operators. Are they So on the one Stubborn, Yeah.
you said it. But I think it's more that they're too much, they're trying to figure out what they do with their data center infrastructure. Cause on the one hand, they want to play a role and they wanna be, you know, they, they feel that and they do. Existing data center infrastructure, why not be the provider of services?
Sure. But they also have a, a balancing act to play, which is, you know, they want to remain a core service provider for their customers. And, you know, maybe don't wanna share that, share some of that with their competitor with [00:46:40] others. And so in general want to, you know, keep their own data centers within the, under their roof.
I think there's some, there are are some interesting models though, right? Like with Dish in the US as a mobile network operator, and I think a lot of these, let's say Greenfield, so the new mobile operators aren't necessarily building their own data centers anymore and they're using third party infrastructure providers.
And that, depending on how you don't get a mobile operator, you know, like every week a new one. But, but you've got, like, say you've got others in other parts, like even Rakuten and Japan, it's a new operator. They have, I mean they have their own edge data centers, but you've also got, in Malaysia actually, they're building their 5G network.
Kind of totally different way. The government has decided, and the regulators decided that they will have a shared radio access network infrastructure, which will be government owned, and they will lease that to the mobile operators. And so they, but they've not built their own, at least from my understanding, they've not built their own sort of edge data center infrastructure.
They're using existing compute infrastructure to do that, or existing data centers to do that. So I guess what I'm saying is that from the telecoms industry perspective, there are, there are, we're already starting to see demand for shared infrastructure because you know, the ones I mentioned, Dish, um, dnb, the one in Malaysia, et cetera, they're using shared infrastructure already today.
It's just a case of, you know, how quickly will some of those incumbents or will those incumbent. Be open to using shared infrastructure. They may, as they may well do, as they start to distribute their 5G network. So they'll need more, basically more edge data center infrastructure to run their network.
And they'll basically, they'll have to make a decision of, is it more cost effective for us to build our [00:48:20] own network of data centers or should we use others? And that I think will drive some of the shed the demand for shared infrastructure. Yeah, it's, I
[00:48:27] Matt Trifiro: mean, and I bet a business on this, so I believe it's inevitable.
And you're right, I think it is more of a timing thing, but the economics are so compelling, especially you look at like, Densifying to millimeter wave in tier two, tier three and tier four cities. Right? Like, like the math just doesn't work. To have every carrier do that. Like, it just, it just, it's broken.
Like you, they don't have enough money. There just isn't enough money to do it. Maybe in some other countries where it's government run or monopoly or something, you, you can do that, but I, I don't see it being done without, without shared infrastructure. The tier one cities, I think the all carriers wanna build it out cuz they want to run that network and, and all that, um, by her Interesting story.
Yeah, an interesting story. So one of our investors is Crown Castle, which in the US is, you know, one of the largest owners. Of towers that, that shared. And one of our board members is, is an old Crown Castle used was that back in the early days. And he said, Look, in the early days, all the carriers wanted to get these towers off their balance sheet.
Yeah. So we bought 'em all up. But when we took a Verizon tower into an at t meeting and said, Hey, why don't you put your radio up here? They'd be like, Over my dead body, you might have put an at t radio on a Verizon tower. And the guys at Verizon say the same thing about the at t towers. And it took a, took a little bit of time, and then the CFOs said, , you start using those shared towers because it, we can get to market 30% faster.
And it's a third the cost, like, why would we do it? So I, I, I sort of suspect, I mean, we already have shared, we have shared towers, we have shared cloud computers, we have shared fiber. So I, I, I do think it's, it's inevitable and I think it's gonna be economics to drive it. I'm interested
[00:49:56] Dalia Adib: if you Yeah, no, definitely.
And I think, I think you're like, you're right. There's a [00:50:00] timing dynamic where today, I mean, like if you, you know we talked about Yeah, the central officers or like the telecom's, core data centers. I mean, you've got some big countries like the US and China, but the vast majority of countries, they'll have a handful of those data centers that the large data centers running, that the brains of the network, and now they're building edia data centers.
But it'll be, you know, it's kind of in the, in the range of like maybe definitely under 50, maybe between 10 and 20. And what we're seeing is of those that are being used for this, you know, the mobile edge compute. Concept we were talking about, it is really a handful. It's like, you know, between four and 10 in MO in a lot of markets.
So it's more of a case. They've not got to a stage where they've, they need more capacity. But there's things like this move towards virtualized run or open run, which is again, moving the radio access network, which is really distributed. It's basically like the base stations, the towers that you talk about, and needing some server, like some compute infrastructure at those.
And I think that once that really kicks off, and most operators are kind of in exploration mode right now. Some are doing proof of concepts, the technology's being proved up. But once that. Gets going, that will be a, you know, a real driver for that. For the need, probably for, for shared infrastructure. Yeah.
[00:51:17] Matt Trifiro: I, I, I think I agree with that. I think I agree with that. So let's looking out to the future, right? One of the ways I, I like to provoke an answer, which is interesting, is, okay, imagine like your godlike, right? And you can look out at all the things that are blocking or slowing down this evolution towards like all these edge compute case that we've talked about and imagined and know would be useful, right?
I mean, medical [00:51:40] imaging is a service like at the edge makes a lot of sense, but it's not happened yet. If you could like, nudge one or two of the blockers, what, what, which ones would you, would you push?
[00:51:51] Dalia Adib: Whoa, that's a , That's a hard question. Okay. Well give, gimme an example. What do you mean by a block?
[00:51:58] Matt Trifiro: Oh, okay.
Like a telco who's stubborn, the old way of doing things. Okay. Or there is no large investment in shared infrastructure and until there is nobody's gonna use it. Or standards or, I mean, I don't open source. I I I'm just interested, like if there's any, if there's any like something that you think is slowing the industry down to their own detriment.
Yep. Because of just the, they're not talking to each other or they're not. Is there anything you see, and it's okay to say you don't see anything? I mean, this is like
[00:52:26] Dalia Adib: futurology stuff. No, I mean, like, I do see things, but they're probably a bit boring, but I'm gonna say them anyway. So I think the first thing is we, there aren't that many, this is telco view of the world, but there's, the telcos sort of come to us and they say, Oh, we don't, where's the money in edge?
What's happening? And what I always say is, Well, where are the edge data centers? Like why would a developer go out of its way to create an application that's distributed when they're, I mean, even a year ago there were like, if you, if you look at the telcos who are building edges in the, in the us I think.
And I don't quote me on this, but like Verizon probably had around 10 nodes. Cox had a few, I think Lumen had a few in other countries. It was, you know, it's probably a handful that are generally available. So you, you sort of have to think like, what, but what on what official, how am I as a developer gonna scale my application when it's only gonna work in a few cities?
So I think that's one thing which is, [00:53:20] And it is, it's taking a risk, right? Because, and it's this whole chicken and egg thing, but there just needs to be more edge infrastructure out there. Otherwise we're gonna be stuck with either the chicken or the egg, whichever way you look at it. So they, That's venture capitalists.
Pay attention, . Yeah. Just like pop some more money guys. And I think the second thing, and I actually wanted to pick your brains about this, not to become, not for the interviewee on the interviewer, but you know, you said at the beginning about enterprises don't buy edge computing. And I'm not sure what the nudge is, but we, I think, you know, as like service providers, you've gotta figure out what is.
What, you know, what is that? Oh,
[00:53:55] Matt Trifiro: I know what the nudge is. I, I, yeah. So this is this, and this has taken me a while to get here, but just identify what customers are doing on premises today. So many, many enterprises have started their cloud journey and they've moved as many workloads as they can to the cloud, but there's a set of workloads that are locked on the premises because it can't do it from the cloud today.
And so if a service provider or a cloud provider or something can unlock those workloads and get them out of the enterprise and turn them into managed cloud services, that, that, that is very attractive. The enterprise two reason, one, it reduces their CapEx spend, it gives 'em a ton of flexibility and scalability, and it becomes a consumption based model potentially, which is all very interesting to the enterprise.
Look what customers are already paying for today, like what they're already buying, what they already need, and what they, and then what they wanna get rid of and not have, but they feel they have to have it on premises for performance or security or other things. And those workloads are everywhere. Like that is a, that is a really, really attractive model.
And the enterprise doesn't know how to buy it. Like they, they know how to buy video management for their security system. Yeah. They just don't, Can't [00:55:00] buy it the way they want to. Yeah. Which is as a service, and I'm just picking that
[00:55:02] Dalia Adib: as one example. Yeah, yeah. No, it's good. I mean, the reason I was interesting to hear from you is because like, it's almost, I feel like the biggest challenges in marketing, like as a marketeer and you're trying to, you know, You're trying to get interest from all these industries.
How do you do that in, in a scalable way? And I think you're right. There are some things that'll comment. It's like, how do you, Maybe it is about how do you move those applications that are stuck on premises of how can you move to the cloud? It's a lot about like, How could you ease the way you manage your, your infrastructure, your applications, your workloads?
How do you, Yeah, how do you kind of like open up new branches or new sites quickly? So I guess there are some, some common themes, but, but at the same time, it can get quite specific, right? Like you talk like video management systems or you know, whether it's like robotics in manage, but those,
[00:55:49] Matt Trifiro: those application providers exist.
I mean, so, so, so the, a thought experiment is Amazon and Microsoft, which have gazillions of dollars and do a ton of market research, both know that there's demand to have their cloud, their APIs, their control plane on the premises, because that's why they have Azure Stack and outposts. Now, enterprises don't want Azure Stack or Outpost.
That's a, that's a pain cuz then they're still in the data center business. They wanna get out of that. So if you could deliver the capabilities of Azure Stack and Outpost, especially in a multi. Way enterprises just eat it up. It's just that nobody's, nobody's done that yet. Nobody's figured out how to, how to actually deliver that.
And those applications will come. I mean, once that's opened up, I mean, I, I, I just think it's, it's gonna be a, a waterfall.
[00:56:34] Dalia Adib: Yeah. It is interest. Cause you like, I think like what you're saying is that it is about bringing, [00:56:40] kind of bringing the cloud to the premises or closer, but at the same time you can imagine, like if you are an enterprise, You're like, Oh, these applications just can't go to the cloud.
Almost hearing something as cloud-like may put you off and maybe I'm just, you know, kind of overthinking. Yeah.
[00:56:56] Matt Trifiro: So that is the marketing challenge. That's exactly the, that is the marketing challenge. That's the, But the, but the cloud had that in the early days. Remember, I can't move anything off to the cloud.
I don't trust it. Well, that's, that's long gone. So I just think this is just a shade of cloud and you're right. I do think it's gonna take the Amazons and the Microsofts and the Verizons and the, you know, the Vapor iOS and the STLs to explain to the world that you actually can extend your cloud journey to include some, all of your currently on premises workloads and create more business opportunity or cost savings or flexibility than you had before in the same way that you did with your workloads five years ago when you moved them to the cloud.
Mm-hmm. , it just, you're right, it's an educational and a timing and,
[00:57:36] Dalia Adib: and that sort of thing. Yeah. And I feel, I mean, sometimes I feel guilty of it cuz I. Your, your podcast is, you know, around edge. We've got, you know, we've got some things around Edge and sort of like, are we, are we not helping ourselves?
Cause we're, you know, we're promoting the Edge band, but there's a reason to, I'm moving away from it. Okay. Okay. Interesting. How
[00:57:52] Matt Trifiro: are you doing that? I'm trying to move the industry away from it. And, and yeah, my, my podcast is named over the Edge, but I'm, I'm moving away from, from Edge cuz I do think, I think it's confusing and again, my customer's, customer doesn't care.
They just want automated checkout for their retail stores at a cost they can afford and a flexibility that doesn't make them feel locked in. And once you kind of like unlock that, like that's where the clarity is. That's where you say like, that's when the enterprises go ding, ding, ding, ding, ding. Oh yeah.
I definitely [00:58:20] would buy that if it existed today. Yeah. And so I'm, I'm really enthused it's taken us, I mean it's taken me five years to get to that point where it seems so intuitive once you realize it. Like, oh, just go ask the customer, look at what they're spending my money today and figure out how to make that more convenient, faster, cheaper, you know,
[00:58:37] Dalia Adib: more flexible.
And I think we were saying though, like, You've almost got a playbook there. You know, you're saying like, what you need to do is find out what are those applications that are stuck on Preem that are stuck with and can't move to a kind of environment, And they, There you go. That's that's all you need to do.
That's all it is. That's it.
[00:58:54] Matt Trifiro: Well, it's there and actually it is, It is. That is not that hard. I mean, assembling the supply chain to deliver that to the premises without the on-premises or with minimal is, is complicated. Mm-hmm. . But identifying what the opportunity is is easy. I mean, medical. Like why, why is medical imaging not delivered in a different way?
Why do you have to stick a giant computer inside every single medical imaging device? I don't know. I mean, I don't think it's technically reason. I just think it's, it's security and, and flexibility and you know, the inertia that the people that make these giant machines have and the commissions that the salespeople get when they sell 'em, and the hospitals that have bought 'em that way.
So I just think it's, it's, I do think it's an education thing. Mm-hmm. Yeah. Yeah. But economics is what's gonna drive it. Cause if you can go to a hospital and say, Look, I can save you a third of the cost on your radiology machines. They'll be like, Show me all day long. And I think you can go through every industry like that.
You can go through factories, you can go through retail stores, you can go through medical, you can go through casinos, you can go through public safety and all that. That addresses. So let's wrap this up so [01:00:00] people know. How can they get ahold of you, Dahlia, and how can they find out more about STL partners?
[01:00:05] Dalia Adib: So I think if you type in s stl partners, edge computing, that should take you to, uh, we've done, we've done a lot of, you know, SEO maximization, so it should take you there. And, but yeah, I mean to honest, kinda like you Matt, we're always looking to chat to people and just like love engaging the industry.
So we've got a ton of stuff we published, like articles and you know, we've got the research stuff that we talked about, webinar, probably webinar.
[01:00:32] Matt Trifiro: We've got a lot of really great stuff that is for free on your, on your, on
[01:00:35] Dalia Adib: your site, good stuff. And always doing these kind of research programs. So, I mean, I think kind of like you who are really keen to speak to is the end customers, especially in the ISVs who are really gonna drive demand and understanding what their needs are and figuring out.
Key. Those applications are the stock, OnPrem, et cetera. So yeah, Partner's Edge would continue to develop things, always looking for ideas and love, love speaking to people. Dahlia, thank
[01:00:59] Matt Trifiro: you very much for, for joining me on the podcast. We appreciate
[01:01:01] Dalia Adib: it. Thank you so much, Matt. Thanks. That does it
[01:01:05] Narrator 2: for this episode of Over the Edge.
If you're enjoying the show, please leave a rating in a review and tell a friend Over the Edge is made possible through the genesis sponsorship of our partners at Dell Technologies. Simplify your edge so you can generate more value. Learn more by visiting dell.com.