Jeff Johnston: Hello there and welcome to the All Day Digital podcast where we talk to industry executives and thought leaders to get their perspective on a wide range of factors shaping the communications industry. This podcast is brought to you by CoBank’s Knowledge Exchange group and I am your host, Jeff Johnston.
On today’s episode we get to hear from Marc Dyman, Senior VP Wholesale and Public Private Partnerships for Kinetic, which is a division of Windstream, and Brian Porter, a telecom consultant with the Tensor Group, to get their thoughts on edge computing and the opportunities it presents for rural telecom operators.
The growth in low latency type applications such as augmented and virtual reality, or advanced manufacturing processes that utilize machine learning, are some of the driving forces behind edge computing. As data storage and processing moves into smaller, rural cities, new opportunities will emerge for telecom operators in these markets. Marc’s company Windstream, is actively involved in supporting this trend and Brian’s work as a consultant make them ideal guests to discuss what’s happening at the edge.
So, without any further ado, pitter patter let’s hear what Brian and Marc have to say.
Jeff Johnston: Brian and Marc, welcome to the podcast. It's great to have you guys here today.
Brian Porter: We look forward to joining you. Thanks for having us.
Marc Dyman: Yes. Thanks for having us on, Jeff.
Jeff: So guys, the term “edge computing” feels like a little bit of a nebulous term. It seems like it comes in many shapes and sizes and I get the sense that there's a decent amount of confusion out there in terms of what is edge computing? Brian, maybe I'll start with you, if you can just help us understand, how would you define edge computing?
Brian: Thanks a lot, Jeff. Maybe I can bring some clarity to that term because I agree with you it's used generically and just to clarify things right away, the edge has been around forever. It's not some new magical thing. It's existed where it needs to be. Right now there's a driver towards the placement of compute and storage hardware at or near the customer's location.
That's really the edge. You're pushing applications and workflow closer to the user in various forms and shapes. Usually, that's combined with connectivity. That's how I see edge computing and that's being driven by certain business drivers. I'll let, Marc, expand on that as well, but it's faster, more secure access to applications and workflow. It's lower operational costs for cloud, greater flexibility from cloud. Those are the business drivers pushing that architecture out.
Jeff: Okay. Let me ask you this, Brian, could a smartphone be considered an edge computing location or an edge computing device, or could like an on-premise data center? Does that fall within the definition of edge computing or is that something entirely different?
Brian: Marc, I'll let you take that.
Marc: I think it does, absolutely. I think if you think about the prior to the edge, there's a core data center network that was developed. This is the New Yorks and the Atlantas and the Santa Claras and all these places. As Brian was talking about, this is just that moving closer to the customer in areas where now in maybe the far outreaching areas or rural areas, that's now emerging in the same way. What started in the core, which was people had on-premise data center applications have now moved those into the cloud.
That's the same thing that's occurring at the edge. It's just a different economics and maybe not as much density that drives some of that. On the mobile phone, yes, it could be, right? I think about anybody that has a mobile phone these days, and I think about TikTok as an example, they're generating content and now that's getting propagated out there in the network. It absolutely can, anyone with a smartphone is now a content creator, so it's going the other way.
Jeff: Actually, that's an interesting thought the whole TikTok content creation for edge. I never even thought about that.
Brian: You're gathering and computing and so here's how I look at it, Jeff. You think about it this way, 5G. 5G is an application that will be hosted by two different architectures. One is on-premise, which is infrastructure as a service edge compute, and then you have near premise, which you have no infrastructure at the customer location, and that'll be an API to a hosted 5G core. Those are the architectures that are playing out. That's how I look at the edge and where that goes is dictated by performance, which is typically latency and cost. Do you need it on-premise? Do you need it near premise? Do you not need it in the cloud?
Jeff: Great. That's helpful. Actually, that's a nice segue into where I wanted to go next, which is around some of the main applications, or I don't know, maybe you call them use case scenarios that are driving this whole edge computing industry. I've seen some examples of revenue projections growing at like 40% compounded annual growth rate for the next 10 years to these really large numbers.
Maybe Marc over to you. It’s sort of a two-part question I guess. One would be, what do you think are some of the main applications or use case scenarios that are really driving this whole edge concept? Then I guess if we had to use a baseball analogy, where do you think we are in terms of inning here and as it relates to the adoption and rollout of these use cases and applications?
Marc: Yes. I would say we'd start just ahead of the applications to look at the investment and if you look at the two big cloud guys, they spent over $40 billion last year, investing in infrastructure. The reason they did that was to position for enterprise workload migration into the cloud from that premise-based solution to a more flexible, scalable, more in control application that has better reliability. By the end of last year, there was only 20% to 30% of those workloads actually have been migrated to the cloud so you just-- it's really to your question, I think it's early innings.
I would look at the applications, at least from my business in two different areas, one there's the enterprise application, which are those workflows, the typical ones that we all work in as we work through business processes, databases, you name it. Those all need to move into the cloud. Then I think the second one is IoT is a big thing. You think about all those devices that are just in your house. If you ever just looked at your wireless network and you go, "Oh my God, I have 60 devices in my house." Well, imagine that in a commercial enterprise application and I think about mining, I think about oil and gas, I think about manufacturing and all those devices and sensors and collection of information that needs to get processed.
I think those are big drivers of the applications on the enterprise side. Then on the consumer side, I think two big ones. One is, and it's a basic one, it's the video content and when you're out in the tier A cities that I mentioned before, earlier in the podcast, it's easy because everything's cached and it's local there and you can get access to it. Now you push out in the far-reaching areas and it's like, "Hey, I want to stream something but that server is now in the middle and somewhere else, far away from me."
We sometimes laugh here where we say, "Well, we just gave you a gig connection to your house," and they go, "Well, that's terrible." Well, no, the gig connection is great, but the content and processing that's occurring is so far away that now that latency and everything makes it a terrible experience. Then the last one, and we've seen it thrown out here quite a bit lately is the metaverse and gaming one which I think is a real thing and I chuckle a little bit because I got one of those devices for my birthday this year.
Now I find myself jumping in there with my niece and nephew and my son, and we're in there walking around, looking at concerts and playing these games and you're like, "Wow," we're talking on there and that only as just like, I think about the devices that we have from a TV standpoint. You were at 720, you went to 1080 now you're at 4k and beyond 8k whatever, the same thing is going to incur in those games and in the metaverse, and that's just going to drive more application and those devices are flying off the shelf. It's a real thing, so those are the applications I see from my side.
Jeff: Yes. That certainly gets a lot of press right now the whole augmented reality, virtual reality, Facebook, now Meta is spending $10 billion a year every year with, I don't know, something like 10,000 employees just working on building the metaverse. It's fascinating how much capital is being thrown at this thing. so that's the sexy stuff or the fun stuff to talk about, but are there any other and more enterprise applications, like virtualizing network operations, whether it be in wireless or C-RAN that you think is also playing a role here?
Brian: There are, let me take that Marc. AI can actually tell if someone's actually inebriated. It gets very intelligent so it's a weird application, just videoing everything basically, and then analyzing it. It can tell you not to turn a machine on because there's a piece of dust in the saw blade, stuff like that and that's proving to be extremely useful in cost savings.
Jeff: Great, okay. No, that's helpful. There's an unprecedented amount of government money flowing into the communications industry right now. We've got the Infrastructure Act that included $65 billion for communication, grants and subsidies. Then you've got another-- at one time it was a total of $20 billion in the Rural Digital Opportunity Fund.
I'm just trying to figure out that and how that might relate to maybe it's not so much an urban issue but as we get into tier two and three cities, what's missing? What kind of infrastructure do you think we need in order to bring this whole edge computing architecture to life?
Marc: I'll jump in first. I think the first thing, and you hit it, it's like there's never been a time probably since the Dot-bomb days where there's been this much capital inflow into the market for infrastructures specifically. We've taken advantage of that by going after it from a public and private partnership standpoint. I think back to your question, it's all of those things. These are served or underserved areas that the business case didn't work before.
It couldn't pass the public or I'm sorry, the private go to market on that where we could make a business case to get a return. Now with those funds and that subsidy, you can combine the two and be able to make that work. That's both for just the physical fiber infrastructure, the broadband, the PON, all those things that we've been developing here as well as the data center that's required to host those enterprise applications that maybe they need dedicated space, they need racks, they need cage space because they want to control it or to introduce the cloud guys to come into that environment that's stable.
That's probably the biggest missing piece on the edge. After you get the fiber in place it's like where's the aggregation point like you have in those tier A cities. There isn't any. It's in somebody's closet and they probably got some fans on it and hopefully isn't going to burn up. You need better environmentals and a condition space that has connectivity that brings you back to the core. I think that funding would enable some of that to occur here lot quicker than just trying to find that anchor a success-based opportunity.
Jeff: That's helpful. Brian anything to add to that?
Brian: I do. Actually, Marc and I have been talking about this recently where you have this need and the need is growing because you have the demographic shifting. Not only do you have subsidies for this, you have demographic moving into these rural areas and they require bandwidth like oxygen. It's actually a great investment for rural companies right now to get ahead of this and put infrastructure in place because it's going to replicate as it replicated in the dense urban areas.
I actually lived through this because I had to go out and get rural co-location space about a year ago. I couldn't get the bandwidth I needed it. It wasn't the space I couldn't get. I couldn't get a 10-gig interface. I would've paid for the build. There wasn’t an infrastructure available at that time for me to do what I wanted to do. If I had done that, it would've made that data center more attractive and that thing would've gotten gravity like they get. I think now's the time for people like Marc who is actually doing this to create these aggregation points that are high quality which is what Marc was saying.
Jeff: Is it like a build it and they will come sort of thing here in more edge or far edge locations. I'm just trying to get some perspective on what advice would you give some of these companies? Do you build it ahead of time or would you say, "No, listen, you need to really focus on more of a partnership approach on a case-by-case basis." Just looking for what kind of advice would you give somebody in rural America or a similar type of market that lacks this type of infrastructure?
Marc: I think Brian was saying it there, which is I think you do have to have some type of anchor to at least make part of the business case. You don't have to make the full business case like you did before because you have some of the subsidy and money to be able to make that work quicker and give you a little bit time to operate. I think that's the value in all these things that are coming out. Everybody's looking in a silo right now when I look at the cloud guys are going, "Hey we've got a premise thing and here we're going to ship you this device and throw it in your closet."
Then the enterprise goes, "Hey, that's not my core competency. I need some management services to deploy that. I don't have a data center here to put that in," and all those type of things come to light. I think you have to have the partnerships, I guess, to be able to drive this forward. I think the other advice I would give these guys is maybe you look at a partnership with a similar entity. I think about the mobile and I think the hyper-scale business, there's larger entities that have relationships and some operators have infrastructure in niche areas.
I think sometimes it's hard to get the mind share of those cloud or those mobile guys to be able to spend time, to be able to get a contract done and realize the benefits of a return on investment, maybe by partnering with a larger entity and coming at it together and going, "Hey, we're solving a bigger problem together and we're representing much bigger ROI." That's definitely a way that we've seen and we've done these type of partnerships with some smaller operators. I think that's a great way to go at it.
Jeff: Got you. Would you say the two driving forces for edge computing, just in general terms are latency and cost? Is that what's really driving us towards adopting edge computing? Cost in the sense that whether it's virtualizing network operations or on-prem or close to on-prem, so you're not having a lot of transport fees, that kind of stuff. If you've got applications like AR/VR that require sub-10 milliseconds, so you don't get vertigo, you need better latency. Is that really what we're talking about here, you guys? Do you think that's true?
Marc: I think that's true. I think the third one too, is the ability to scale. When you're on-premise, you're always waiting for that, "Hey, I need that equipment. I've got to do it myself. I've got to store it in the cloud." It's like you click a button and you can scale things up. I think that's the third piece. Yes, I agree with you on those two points, Jeff.
Jeff: Well, we've covered a lot here guys, this has been great. I want to make sure that we cover all the salient points while I have you. With that, I'll just open it up. Any closing thoughts, comments, anything that you wanted to share or cover that we haven't already done so?
Brian: Sure. I'll give my closing thoughts, —Jeff. I see the edge movement as a transition to just a more efficient architecture, where you have decentralized and disaggregated elements of the network getting closer to the user. Okay? It's hard to do that. It's hard to get the economies of scale that you need and the cost structure in place to do that in today's market right now because the guys who are active are self-performing so you’d have to take a lot of risk.
The guys who are going to be successful here are going to be the same trusted guys, and it might take a tribe to do it, are going to build the infrastructure in places where these guys can't get to in time. There's going to be a lot of opportunity for people because this trend is going to continue for a very long time. I've watched Marc, I've watched what he's done in my job as a consultant, and he's just doing exactly this.
He's bringing in smaller guys, he's aggregating their traffic, he's bringing the technology that makes it cost-effective. He's out in front of it, probably what Marc? Two years. That's the kind of behavior you need to get this done.
Marc: To build on what you were saying there, this isn't a light switch thing. There's some time, there's some looking at different things that are going to scale that can make all those business cases work. I know we've said it a few times here, but the partnership side to me is critical. No one person can do all this. I think together you become a collective better force of executing against the needs.
The last thing I would just say is, listen to the customer. That's the age-old thing. Listen to what they're saying, what their desires and needs and what they are because that ultimately drives what's going to happen here in my mind.
Jeff: Great. Well, let's leave it on that optimistic note. I think that's great. I love opportunity and I love when people are ahead of opportunities. Kudos to you Marc for being in that pole position. Thanks so much, Brian and Marc, it was a pleasure having you on the podcast today.
Brian: Thanks, Jeff.
Marc: Thanks, Jeff.
Jeff: A special thanks goes out to Brian and Marc for being on the podcast today. So despite the fact that defining edge computing is easier said than done, my two biggest takeaways are: one, in order to be successful in the edge computing market, partnerships must play a critical role. These partnerships come in all different shapes and sizes. For example, they could be public/private partnerships, partnerships with hyperscalers, enterprises, fiber providers, and so on. And secondly, we’re still in the early innings of this whole edge computing market, but the opportunities are big. And those who start to align their business with edge computing now should be well positioned to reap the benefits in the future.
Hey, thanks for joining us today and watch out for our next episode of the All Day Digital podcast.
