Menu
When it comes to edge computing and the telecom industry Jason Hoffman, President and CEO of MobiledgeX, does not mince his words. In this interview with 6GWorld, he addresses industry mindset, edge’s roles, and shares his vision on the market’s future.
It’s All About People
“The structure of your organisation, of your people, always mirrors the technical architecture that you’re working with,” Jason Hoffman explains chatting with 6GWorld, “What we often don’t think of is that all of our technical systems are done by systems of people. When we propose technical architectural changes, we often don’t actually change the systems of people.”
We are talking about the elusive ‘mindset’ that can support or hinder business agility over the coming decade, and how that pertains to the telecoms environment. Hoffman has a way of setting out a broad theory and then applying it to the situation under discussion.
“Amazon themselves had to undergo a cloud transformation, you know, back in 1999 to 2003. They had a hard-ass CIO that drove it. And he went ‘people that don’t do it get fired’ and things got restructured.
“The mobile network operators haven’t really undergone the same internal mind shift that a lot of other enterprises have had to undergo thanks to the cloud, because these guys still tend to do their own infrastructure and the like. And so when we look at that, yeah, the systems of people in the telecoms space haven’t really mirrored the new technical systems.”
The Third Application Environment
MobiledgeX has been in the news recently as it announced a swathe of new investors, new operator engagements and a new version of their Edge-Cloud platform. Does MobiledgeX still face questions about what exactly the ‘edge’ means, though? After all, it can be used to refer to a wide variety of different locations in the network. For a service that seems inherently tied to localisation, Hoffman’s definition is surprisingly unrelated to network specifics.
“We started thinking about there being an embedded application environment in a network, an environment that lets you go and do new things in there,” he begins.
“In the case of the device, there were clearly embedded application environments. In the case of smartphones, it’s basically down to iOS and Android. And there are a lot of singular frameworks that you go and develop in, and it spits out applications for both those two environments.
“Then for the internet we’ve also had the same thing happen, where there are at-scale embedded application environments in the internet. I say that is what the public clouds are. Amazon, Microsoft, Google are full-blown embedded application environments for internet.”
Hoffman points out that the network in between the cloud and the device has never been treated like that, never had an embedded application environment which people can develop for. Instead, the major vendors have historically developed their own stacks in-house.
“In the case of 5G, what we saw was that, even in the major RAN vendors, everybody at least standardised on Kubernetes has a base platform. So we see the beginnings of that,” Hoffman notes.
However, MobiledgeX, in Hoffman’s view, can and should act as a catalyst to go much further.
“How do we serve up our own embedded application environment in what we really have as a collection of operators now? Not what we think we have, but what we have – this very heterogeneous infrastructure footprint that’s in a bunch of locations? How do we bring that up into a common model? And then how do we take the device, plus that edge, plus the cloud all together in a sort of an end-to-end system that solves the hardest, biggest system problem for developers?”
This is an approach that MobiledgeX is uniquely able to take, thanks to its ownership by operators including Deutsche Telekom and SKT. “Literally the reason Deutsche Telekom had to start the company is because there’s not an existing vendor in the world that we motivated to ever make it,” Hoffman chuckles.
“I mean, what? Amazon is going to come along with an edge suite that lets you use Google really easily? Really, they’re going to do that? Or an Ericsson, Nokia, Huawei’s going to come along with something that makes it really easy to take their C-RAN and just run it anywhere? There has to be a globally available product in the space. It solves the real ecosystem problem of how to bring all these things together.”
That product, essentially, comes back to the concept of an embedded application environment that can be used to enable services.
“And I think when you sort of look at that base definition, that ‘clouds’ are the embedded application environment of the internet and ‘edge’ is the application environment embedded in a wireless network. Then you start to say ‘okay, in the case of wireless networks who owns them?’”
The Uncanny Enterprise Edge
Hoffman’s eyes light up as he starts to unravel some of the implications for enterprises.
“Wireless networks are typically owned by an operator or by an enterprise – they’re either in public use or private use. And then some people will call that a ‘telco edge’ and ‘enterprise edge’. But the characteristic that an edge is ‘an application environment that’s embedded in a wireless network’ is always true.
“The thing that makes it strangely full-circle for people is that ‘edge’ becomes the collective term for everything that’s left at the enterprise. Well, you start having some strange things happen, as that one strange thing is that your mainframe strategy is now categorised as ‘edge’, right?”
It’s clear that Hoffman relishes making absurd-sounding statements, even as he lays out the rationale for it.
“Even CIOs like to have three simple things that they describe stuff in – ‘Father, Son, Holy Ghost’, you know, easy to say. I mean, people tend to do that. Historically CIOs have had a SaaS strategy and a cloud strategy, but then they’ve had a whole list of other crap underneath.
“Now what you’re starting to hear from consulting companies and the like is, ‘Here’s your SaaS strategy. Here’s your cloud strategy. Here’s your edge strategy.’ And under ‘edge’ is everything that’s left out of those other two.
“Well then, even your mainframes are part of your edge strategy. When people hear that they’re like ‘That doesn’t make sense. That’s not progressive!’ Well yeah, everything that’s left on-premise is now ‘edge’. And that’s going to include some non-sexy things, some of which may even be a legacy system of record.”
What the Edge is Not For
Hoffman can’t restrain himself when 6GWorld mentions the idea, sometimes shared by operators, that edge services could be used to reclaim some revenue and influence from the cloud players.
“You’re correct. That’s a correct use of ‘reclaim’ because they typically will say ‘reclaim’, but it’s very funny when people want to go reclaim something they’ve never had and didn’t exist before,” he snaps.
“Amazon went and disrupted. Fundamentally. If you look at where their money came from, it came from the system integration business at IBM,” Hoffman explains, “Literally if you look at, ‘where did the first $20 billion come from into the public cloud space?’ it shifted from system integrators. Then everything that came on top of that is net new. That didn’t exist.”
Hoffman seems like an exasperated parent.
“Some operators think ‘Oh, we’ll do edge and that’ll somehow shift money from the public cloud guys to us’. No! That was never your money in the first place.
“And it’s not even how it would really work. There’s no example of an application where you say, ‘You know, what we should do is we should stop working with one global company in seven countries, we should distribute the app across a hundred countries and have a hundred procurement relationships.’ Who’s going to do that, practically speaking?”
What the Edge Really is For
Hoffman’s enthusiasm spikes as we talk more about where the opportunities really lie.
“If you really look at the edge space, it is about emerging and net new devices and emerging use cases,” he explains, because of the role that the edge can play.
“The edge connects up a pipeline. And a lot, a lot of applications now are changing from ‘Here’s a backend that sits there and here’s this device app’ to much more ‘We design a pipeline that has a certain flow to it’. That’s because more and more applications are looking at the flow of data and building an AI pipeline around that.
“Edge is a required thing for that pipeline to be complete. If you look at something from the mobile operator space, all the use cases we then support are basically around completing and making that pipeline possible.”
This pipeline concept could create customers of even the largest players in the digital industries, such as Facebook or YouTube.
“Imagine if we sat down and said, ‘When we hand a video off to you, it’s actually metadata-tagged with a lot of things around, say, who created that, where it was created, all the initial filtering of whether there’s a head without a body, blood, nudity, anything else like that. By the time it gets handed off to you, it can undergo a much cheaper curation process than the 30,000 people you’ve hired from Accenture looking at all your YouTube videos today.’”
Hoffman emphasises the need for creativity and, once again, the need to break out of traditional ways of thinking and acting. Talking of the ‘pipeline’ model of device, edge and cloud might bring on telco terrors of being a dumb pipe, after all, and failing to monetise.
“What exactly are you doing to make your pipe not dumb? I’m sorry to say, but if you’re just running around buying hardware from people and sticking it on towers, and then opening up stores and selling cell phones, you know, your shit’s going to be dumb.”
Edge Use Cases Today and in Future
MobiledgeX has been working a good deal recently with leaders of smart cities; in transportation, logistics, and in industrial use cases among others. Hoffman identifies four broad categories of application that are good candidates for integrating an edge capability.
“The first one is the location-specific need for computing and data. The second somebody needs a highly precise and accurate GPS location for a given work, then okay. All edge cases have that.
“The second one is when you look at consumer-facing use cases – in particular where there’s somebody looking at it with a human eye. They’re multimodal, they’re multi-device, they’re personalised, they’re interactive, and there tends to be some sort of social or collaborative element to that.”
This is a very useful description of the qualities of a service, but what does that mean in practice? Hoffman explains. “You can call it AR or MR. You can call it whatever you want, but it is a bunch of people standing around having a common experience on their collection of phones in a specific place.
“There have to at least be metadata services that make that possible. Because we can say, ‘These three people are on a family plan. They’re now on their own Wi-Fi network, but they’re on three generations of these devices, and there’s a 120-millisecond latency difference between the cameras on those devices. It’s not a limitation in the network.’
“When you provide even that metadata back to an application developer they’re able to say, ‘Okay, the experience we’re delivering is based on the slowest device that’s present, and everybody else gets normalised to that experience,’ which now makes it possible to be social. I mean, what family wants to go play an AR game together where the dad or the mom with the fastest phone constantly beats the kids at the game? That’s not fun, you know? And so these little things that we can provide turn out to be of value.”
The third area of interest has to do with large amounts of video, for example from security cameras in a smart city or building. “And then on that video, everything that happens – pattern recognition, computer vision, AI pipelines, how that hybrid pipeline sits in the edge and the public cloud – all those use cases for us become this hybrid pipeline,” Hoffman enthuses.
The final element addresses items such as robots, drones or cars which are capable of independent movement. “The second you start having devices that do their own mobility, a natural use case inside of the networks is to do all the traffic management.”
The future of edge services seems bright to Hoffman, because of the range of services and the scope for customers.
“A smart city, smart building or smart factory is actually a combination of all four of these functional areas,” he explains. “There’ll be a combination of some location-specific computing and data needs. And we have some human beings as end-users inside of that. And those human beings, guess what they’re trying to do? Multimodal, multi-device collaborative, social, personalised, interactive content, at the end of the day.”
Low Latency… Giving MobiledgeX the Jitters?
Ultimately, the conversation comes back to mindset and the need to think differently about how we focus our attention, and – by extension – how we organise.
“As an industry, are we doing things in our networks that actually make it possible to do location-specific computing and data? Are we doing things that make it possible to support people who are doing multimodal, multi-device, personalised, interactive, collaborative, social content? Are we doing things in our network that makes it possible for us to deal with all this video that’s coming in and to really connect this together well for people? There’s this whole emerging space of autonomous mobile devices with control loops and structures. Are we doing things in our network to support those?
“No. Instead we sit around and we talk about 10 millisecond latencies, but that misses the point.”
[/et_pb_text][/et_pb_column][/et_pb_row][/et_pb_section]
Legal Notices. I Privacy Policy I © COPYRIGHT 2024 6GWORLD. ALL RIGHTS RESERVED.
Legal Notices. I Privacy Policy I © COPYRIGHT 2024 6GWORLD. ALL RIGHTS RESERVED.
Legal Notices. I Privacy Policy I © COPYRIGHT 2024 6GWORLD. ALL RIGHTS RESERVED.