Home5G‘It’s the Android model’ – Google Cloud on how it will remake the 5G edge with carriers

‘It’s the Android model’ – Google Cloud on how it will remake the 5G edge with carriers

Note, this article is taken from a new Enterprise IoT Insights report, called The Role of Hyperscalers in 5G and Edge. To read the full article – including about Google Cloud’s go-to-market strategy with operators, the breakdown of use cases across the cloud-edge, and roles and responsibilities in the edge-5G ecosystem – grab hold of the report, which is out later this week (check back here for availability).

The enterprise 5G market has its head in the cloud, in case you haven’t noticed; or rather, the cloud cover has come so low, as hyperscalers have distributed their componentry into the edge, that enterprises hardly have to raise a hand to grab the requisite compute functions to hook-up their highest-fidelity 5G operations. These technologies – edge compute and edge (mostly ‘private’) 5G – are being combined as the foundation for powerhouse IoT and AI.

They are being set, as we write, as the springboard for wide-ranging industrial change. And as the market watches, and plays along, the two sides supplying these technologies are rubbing up against each other, sometimes in partnership and sometimes in competition. For the telco community, which has dragged a fifth generation of cellular through development and standardization, enterprise-geared 5G appears like a shot at redemption.

For the old dogs of telecoms, it presents a chance to reassert power and influence, among the big beasts of tech, by animating a zillion ‘things’ in the enterprise space – all the robots, machines, and processes in all the buildings, plants, ports, and cities in all the world. But the hyperscalers are plumbing their wares into these venues at the same, sometimes in local 5G network infrastructure, and exhibiting more than a passing interest in the 5G game as well.

Regulatory moves on cellular spectrum in most major economies, and ongoing developments with standardized network hardware and cloud-based network software, mean 5G is up-for-grabs as the ‘far’ edge – on enterprise premises, where the action is. This is where the perceived tension between telcos and hyperscalers, as the most awesome of their new 5G competitors, is highest.

Of the three ‘big’ hyperscalers, Microsoft and AWS appear more like rivals on 5G. Microsoft bought 5G core network provider Metaswitch 18 months ago to serve 5G to enterprises (at the same time as it acquired Affirmed Networks to supply operators with the same). AWS announced a new box-fresh mass-market private 5G offer last week with a number of unnamed partners, going under its own brand for now, to ride the gathering CBRS wave in the US.

In between, they have struck deals variously with tier-one operators to piggyback on their public multi-access edge compute (MEC) infrastructure in their 5G networks, and work through them to go to enterprises. Google Cloud, by contrast, appears quieter. It has the same kind of MEC deals, most prominently with AT&T, and is working inside carriers to support their own cloud migration and network modernization. 

But it maintains a clear line – in its conversation, and also in its activity – that it wants edge-5G ‘partnerships’ with operators that serve both parties. In conversation with Enterprise IoT Insights, it does not swerve from this line; this is a point of difference, it says. Its distributed edge strategy, explains Alicia Abella, managing director for the firm’s telecom, media, and entertainment (TME) industry solutions division, has developed over two years.

Its work with the carrier community at the so-called MEC edge has developed over the last 12 months. In October, it unveiled its Google Distributed Cloud product, to extend its cloud-compute infrastructure to the network and enterprise edges, and provide managed enterprise IoT support, alongside other customer applications. It is starting with rollout to 140 “global network edge locations” in mobile operators’ MEC footprints, plus to customer premises.

The key part of the Google Distributed Cloud product, available as a ‘hosted’ and an ‘edge’ service, is the new Anthos hybrid-cloud platform, which makes developers’ software inventions work the same way all across the edge-cloud continuum. It is the same Android model, just for enterprises, says Abella; it is the common thread that will make all the future digital-change wizardry scale, and the edge-based 5G-compute springboard bounce with business. 

And Google Cloud is in charge of it – which, one might argue, could reflect its comparative magnanimity about who holds the customer relationship on enterprise 5G. Here, Abella discusses all of these aspects at length; all the answers below are hers. The full interview is available in a new Enterprise IoT Insights editorial report, called The Role of Hyperscalers in 5G and Edge, which is out this week, and will be made available here.

Put Google Cloud’s work with mobile operators around enterprise 5G connectivity and edge compute / storage into some context. Why is Google Cloud putting such focus on the carrier 5G market, and why now?

“It is a perfect storm [for them]. The CSPs are hurting; they’ve seen a stagnation in revenue – which is not expected to return to pre-pandemic levels until about 2025. [More than this] the investments they are making into infrastructure are enormous. Eighty percent of their investments will be focused on 5G in the next few years. They need ways to make their money back. And they are going to be challenged, as well, with this data explosion on their networks. 

Abella – all the stars, all the grains of sand

“Because the prediction is there will be 13 zettabytes of traffic on the networks by 2025. And you know, a zettabyte is a big number – one with 21 zeros, in fact. The same as all the stars in the observable galaxy, or all the grains of sand on all the beaches in all the world. That is a zetabyte, to put it in perspective. So the CSPs have a challenge on their hands, in how they cope with all of that data, and how they reduce the costs of operating those networks. 

“That is a lot of pressure – to make sure the network is humming, and there are no outages. And then there is this other trend, that CSPs are not doing too well when it comes to net promoter scores. They haven’t been able to break 50, recently – in the US and EMEA – which is a sign of the loyalty customers feel for them. They are investing 15-20 percent of revenues to improve the customer experience – and yet they are still not reaching any higher.”

And just explain your work with them; how is Google Cloud helping weather this ‘perfect storm’ inside this telco micro-climate?

“So, yes; these are the trends: big cap-ex investments, large volumes of data on networks, low customer brand loyalty, and then just the fact they are looking for ways to increase their revenue. What we did was to take these trends and focus on four key areas – in order to be able to address those trends CXOs are coming to us with. The first focus area is on monetising the edge. How do CSPs take advantage of the infrastructure? How do they partner with hyperscalers, and with their ecosystems of partners, to deploy use cases on the edge that will be able to unleash new revenue generating solutions for customers? That’s one way to help generate revenue at the edge.”

“The second is network modernisation – how do CSPs move to software defined networking solutions? Some have been on that journey for several years; some are just starting down the road. Some are bypassing virtual functions altogether, and going to containerized network functions – because that is the more modern diversion. And they are also embracing ML and AI to be able to build automation into that network – so they can create these closed-loop automation capabilities so the network starts to identify problems, analyse performance, and fix and heal automatically. That will also help with managing the network, especially in the context of so much data.

“[The third focus is on the customer] I worked on contact centre technology when I started in telecoms 20 years ago, and the calls that come into the telco contact centres are still the same. But the technology is better now, around customer analytics and call centre automation. Google Cloud is engaged with CSPs on solutions for contact centres – to produce automatic conversations with customers, get at their problems, and solve the simplest ones and pass the tougher ones to the reps. The same with customer analytics in general – looking at usage data around products and services, and how to improve those.

“The last one is IT modernization – how to improve and modernize the IT systems within CSPs, to get the benefits of moving to the cloud, and reducing that TCO.” 

And against this backdrop, can you first define the ‘edge’ for us, going from the cloud to the enterprise premise? Where is it, and how should we rationalize it? And then explain the new Google Distributed Cloud product, and how it serves those edge venues?

“The edge is everywhere; it is a continuum – on site with a customer, in the mobile network edge, or in a public or private cloud. How it is used, and where you put workloads, depends on the performance requirements for each application. If you need really low latency, you probably go to the [far] edge. But it will vary, even within types of industrial applications. Autonomous vehicles represent a good example.

“Because you have safety related applications that will never leave that car, because they operate in real time, and you have other kinds of use cases, as well, such as navigation and mapping, which can probably be served from the [mobile network] edge, and updated in near real-time. And then you will need software updates, as well, where latency is less important, which can be pushed from a regional data centre.

“Google Distributed Cloud is a fully managed family of hardware and software products to take Google’s cloud closer to the edge. There are two products: GDC Hosted and GDC Edge, where [the first serves use cases from Google Cloud’s regional points of presence and] the second pushes use cases and applications to the edge with CSPs and their enterprise customers.”

And explain, as well, the role of the Anthos system within the Google Distributed Cloud product, and perhaps how the Android philosophy informs what you are doing now in the enterprise 5G space.

“So Anthos is inherent in Google Distributed Cloud, as well as in our data centres, to provide a consistent developer experience, from data centre to edge. By having that consistency across all those locations, it becomes easier for the developer – to develop once and deploy anywhere. That’s the Android model, which we want to recreate. Because the ecosystem doesn’t just include us, and the CSPs, and NEPs, and big SIs. It also includes the developers. 

“The developers have to be part of that ecosystem as well, because that is where innovation comes from, and because those edge use cases are still percolating. People are talking about all kinds of things, but the applications are not defined yet. That is why we need that consistency for the developer community. And why we partnered with Orange on this 5G Innovation Lab in France, where Orange is bringing the network and we are bringing the edge, and we are together inviting enterprises and developers to test their ideas in this sandbox.”

To be continued…

Note, this article is taken from a new Enterprise IoT Insights report, called The Role of Hyperscalers in 5G and Edge. To read the full article – including about Google Cloud’s go-to-market strategy with operators, the breakdown of use cases across the cloud-edge, and roles and responsibilities in the edge-5G ecosystem – grab hold of the report, which is out later this week (check back here for availability).

Previous post
IoT is easier than ever, and on the verge of greatness, at last (TTI on 2022)
Next post
Open virtualised RAN will accelerate adoption of private 5G (Mavenir on 2022)