The first part of this Q and A with Vodafone’s Joanna Newman, senior director of technology partnerships, covers the operator’s dedicated vs. distributed MEC approach, the logic behind its various hyperscaler partnerships, and addresses how many MEC use cases involve common sets of building blocks. Taken from a session at the recent Telco Cloud & Edge Forum (available on demand here), the second part of the conversation looks at the intersection of private networks and MEC, the importance of ecosystem development, how operators assure the combo of connectivity and compute for business critical operations, and if there is a single killer MEC use case—or if asking that question is missing the point. The following has been lightly edited for clarity and length.
The private mobile network, MEC Venn diagram
Q: Regarding the relationship between mobile private networks and edge computing, can just maybe give us a bit of insight into if you’re an enterprise buyer and you see a potential role for edge computing, why would you also potentially pursue a private network to meet that business outcome?
A: That’s not always the case. I think that when you’re looking at bringing new use cases to life that have not been able to be provided before, that require ultra low-latency or even low-latency capabilities to deliver, you also need the connectivity to be able to cover the facility so that could be done. I think that there’s a story about a tractor manufacturer, tractors and agricultural machines are very big things, and I have a five-year-old who’d know far more about this topic than I ever particularly wanted to, but they’re very big things. And so of course, the manufacturing facility is a very big place as well. And actually, where they kept their just-in-time supplies, they weren’t able to have any coverage at all because of the size and the shape of the facility unless they had their own cell tower built.
So they used a telephone and a guy with clipboard, who was obviously very competent. But if you want to have just in time capabilities maybe delivered via autonomous robots to keep pumping your supply line, then they’re going to need connectivity that goes along with that. So one of the nice things about edge computing and mobile private networks is mobile private networks by themselves provide connectivity that unlocks some of those new use cases that haven’t been able to be done before. And you need that low-latency capability that’s best done. You can use the geographical function if you like, but you’ll get a better ultra low-latency and compute that’s dedicated to you if you have an edge compute with it. So it’s a with, not a must, if that makes sense from that perspective.
Now, there are companies that are exploring edge computing on their own and have all the connectivity that they need or might have connectivity from a different provider or however that happens to function, and want to explore edge computing for themselves. Because remember, these business cases can be quite substantial. Bringing new capabilities online is not free and sometimes there’s proof of concepts or trials, and of course the edge computing can be provided, just obviously via Vodafone but also direct as well, to explore some of those use cases that might work for a proof of concept or a further pilot or a trial. But ultimately, the mobile private network unlocks capabilities that are best delivered alongside edge computing and not 100% the other way around
Q: In assembling an ecosystem or ecosystems to address the sweeping opportunity of MEC, how do you maintain a degree of focus in your own ecosystem development while also maintaining the flexibility to opportunistically pursue projects as they come up?
A: It is quite a complex space of partnerships and has taken a long time to evolve to get here. And of course, there’s also the system integrators that go to site and set up these networks and set up these edge computing as well, so there’s another player as well. But to come back to your question, so Vodafone’s approach has really been twofold. So first, we’ve run an edge innovation program for more than two years now. Keep in mind that our MEC capabilities for distributed MEC in London, which was our first site, came online about six months after Verizon’s did, and that was only summer ’21, so this is a very recent technology that’s actually been available to customers.
And so what you have to focus on for the edge innovation program is that it’s been a front door, a joint front door to come to us to say, “Hey, I’ve got this use case. I would like to understand more. How could it work in my facility?” Or, “I’m an ISV that does this for a specific type of computer vision for a port, for example, like Dedrone. Dedrone is a great use case because they identified drones that are interfering with the port and run a geolocation and the detection, etc… And they came to us via the Edge Innovation Forum and said, “This is what we do. We think that that would be really appropriate for edge computing. How can we go about doing that?” And then we supported them to do so. Doesn’t necessarily mean that Vodafone has acquired these companies or there’s a partnership in place, but we have a mechanism to be able to solve the problems of that particular industry and that particular use case through who we’ve met through the Edge Innovation forum itself.
Now, as we talk about ecosystem development, Vodafone won an IBC Champions award last year, and IBC is the International Broadcasting Convention that happens in Amsterdam every year. This is made up of broadcasters globally. And we displayed an end-to-end use case, but we were part of the group of people that demonstrated an end-to-end use case for location-based XR, so AR and VR. It’s a game called Hado, which is basically you’re throwing flaming balls of fire at one another and building walls all through VR headsets, et cetera, and it’s location-based and it also happens to be the top growing e-sport in the world, which is why it was selected. And to build and demonstrate that end-to-end use case required eight members of the Champions Committee to be able to do it. Now, to be fair, two of them were Vodafone, one in the Netherlands, which is a joint venture and one is per Vodafone, that was Vodafone group, but that still leaves six.
So that ecosystem of all of those components for a simple a use case as a location-based XR has a lot of pieces associated with it, and that ecosystem needs to mature as well. And one of the things that we’re able to do through the edge innovation program is to understand what’s happening with the little fish and also how they can be part of those use cases. Or a customer will come to us saying, “We have a long history of working with this developer who runs our manufacturing line that generates 80% of the P side of our P&L. Therefore, we really don’t want to give this to a new, unproven ISV. We want to work with them. How do we do that?” So there’s a consulting side of it as well to help them bring their application to light.
Assuring the edge
Q: So when you’ve got that type of network, that type of customer, and you’re running a latency-sensitive application, how do you extend the NOC capabilities we think of residing in a big centralized facility, and have them be geographically in an appropriate place to where you can actually do something if you see that there is a degradation of a specific KPI?
A: It’s something that come up to us with these global use cases that are coming and global RFPs that are coming. And Vodafone was selected for some parts of Rio Tinto because of our capabilities in this space. So if I think about how you monitor a facility, obviously the level of monitoring and the level of redundancy that you want is actually up to the customer. We’ll make a recommendation of course, but it’s up to the customer to choose. We also need to monitor the health of the private network, we need to monitor the health of the applications if we’re accountable for that, or we might monitor the health of the applications in the resource stack, either as a direct part of the sale or as an indirect part of the sale, etc… And we have a global NOC capability anyway, as we operate across, as I said, all of the Vodafone and Vodacom footprint and all of the JVs of which there are myriad, and then across partner markets.
We actually have a lot of NOCs that are in location that are able to query the resources. I think the other side of that, especially regarding edge computing resources, is that there are resources that are dedicated to you or distributed in a network facility, but if you have dedicated edge, then somebody needs to be accountable to ensure you have enough resources for what it is that you’re doing and ensure that that scales as well. And the AWS shared responsibility model is really what we’re led by here. Absolutely, Vodafone or any operator can take accountability for ensuring that there are adequate resources provided, so through monitoring, etc… But also, if the customer wants to monitor themselves, that they need to be sensible with what they’re running on those resources as well and ensure that there’s enough to support, especially when these microservices spike, et cetera, that there are enough resources to do what needs to be done.
Is there a killer MEC use case?
Q: Based on previous conversation with you, there’s maybe not a single killer global MEC use case. And then, not to incorrectly paraphrase you, but I think you took it a step further and said if that’s what you’re looking for, you’re maybe missing the point a little bit. So can you just expand on that line of thinking?Â
A: Sometimes technology is talked about requiring the killer use case that’s going to drive adoption across a geography, across a continent, across developed nations, etc… And sometimes edge computing is described in that construct as well. And of course you can see the CAGR of edge computing over the next six or seven years and it’s really steep, which is great, which indicates that there are some use cases that will be widely adopted, driving that compounded growth. You see a similar CAGR in private networks…which aligns to our view of the world of them being together. And then distributed requires more use cases, which will come online that happen to do a session management east-west, the open gateway project from GSMA is talking about that.
But when it comes to the killer use case for edge, I really think that people are potentially limiting themselves to thinking about just one, because edge computing is a platform capability that’s available to all sorts of industries, including some we haven’t talked about today. Retail is going to be using edge computing in an enormous way. We talked about logistics, we talked about heavy industry, we talked about far edge and remote locations. We’ve talked about manufacturing. And to think that there’s one use case that satisfies all of those industries and all of those enterprises I think can be a bit limiting. The way that we’re focused on it is to look at what those capability building blocks are between those use cases that are emerging across those industries, and to not close the door on some specialization if it turns out that manufacturing companies require computer vision and visual inspection with a little bit of an extra special sauce associated with it, you’re not limiting your market by focusing on that. There are a lot of manufacturing companies in the world to that can do that.
So I think focusing on a single killer use case for edge is a bit limiting. And in actuality, there’ll be 50, 75, 100, 500 depending on where we are on the CAGR scale, but we’ll probably find some common capabilities amongst some of them and those building blocks.
