The communications industry is changing at a dizzying pace, so who best to ask for insights into where things are headed, and when?
You could ask a VC, I suppose, but they only get things right 5% of the time. You might ask a service provider instead, but they’ll just start going on about ARPU (which is something to do with money, apparently – not the name of the Kwik-E-Mart owner in The Simpsons). Or you could ask an industry analyst, but they’d only try to sell you a report containing the information they just got from the vendors.
No, if you really want to know what’s coming down the pike, you need to talk directly to someone tasked with anticipating technology change and turning it into viable business solutions for the largest companies in the world.
In other words, you need to talk to the CTO at Ericsson: Erik Ekudden.
Ekudden and his division of Ericsson navigate that tricky territory between customer expectation and technology reality, making sure that Ericsson isn’t caught out by important technology developments, while also developing solution sets that will help its customers save and make money in the long term. That requires diplomacy as much as R&D, especially when the things customers think they want (“cloud stuff,” “robots,” “infinite bandwidth,” “some of that sweet, sweet social media revenue”) may not actually be what they need in order to build survivable businesses in the here and now, and beyond.
Balancing 5G expectations
Standalone 5G (5G SA) is the key example of a technology where Ericsson has had to balance the expectation/reality equation with customers. That job is ongoing, but it is also the key that assures Ericsson’s long-term success.
When the newest G standard became available, telcos (especially in the U.S.) assumed it would work the same magic as 4G, allowing them to make more money by charging consumers for faster throughput (and selling lots of cell phones). But using 5G as a bandwidth booster is like using a banana as a socket wrench: messy, and a waste of a perfectly good banana.
In fact, those service providers had made the fatal mistake of not reading the owner’s manual. It turned out that 5G wasn’t the technology service providers thought it was — it was better.
5G encompasses 104 separate sub-standards, thousands of documents and millions and millions of complicated, hard to understand words. Fun factoid: working in shifts it would take a team of humans two months to read the full documentation. But it would be worth it, because hiding in plain site within the geek coda is a treasure trove of new capabilities (cloud compatibility, network slicing, etc.) designed to enable new types of services, generating new types of revenue in new markets (vertical industries, autonomous driving, smart everything — and, in the U.S., state-sanctioned killbots, but I digress.)
If you’re an Ericsson standards wallah, and you’ve been hanging out for the last decade with your colleagues at Nokia, Huawei, Marvel, Qualcomm, Intel, et al., developing those specs you know this; it’s literally “the point.” And it must be maddening to see your telco customers getting their first 5G rollouts so wrong — let alone having to listen to them complain about how they’re a bit disappointed in 5G, actually.
And now for the good news
This is where the good news kicks in because, despite this hiccup, the 5G revolution is only just getting started, and it is an irresistible force, and eventually it will achieve a level of ubiquity and strategic importance that dwarfs the impact of previous G standards.
There are two reasons for that: first, money. If they are to survive, service providers have no choice other than to explore and eventually embrace the new business opportunities locked within 5G. (It’s just taking them longer than it should have done, because … well, you know … service providers.) Ericsson is laser-focused on helping CSPs do this, via two strategies: the Open Programmable Cloud Ecosystem (OPCE), and the Ericsson Open API initiative.
The second is 5G’s cloud-native design, which equips it, uniquely, to serve as the shim allowing service providers to meld essential new 21st century virtual technologies like cloud infrastructure, artificial intelligence (AI), and autonomous automation onto their pre-existing 20th century telecom infrastructure (antennae, switches and routers, racks, bits of cable).
Cloud-native 5G is the bridge between old and new, and this essential role guarantees that 5G will be a ubiquitous component of world infrastructure for the next decade.
As one of the “big three” 5G vendors (along with Huawei and Nokia), Ericsson stands to be a beneficiary of extraordinary demand for 5G solutions to support the ubiquitous deployment of the technology on a global scale.
That’s one reason to feel positive about Ericsson’s long-term fortunes. The other is the integrated strategy that Ekudden and his team have developed to support its customers in taking what is, unquestionably, the hardest transition that the comms industry has yet seen. It manages the neat trick of being aspirational while still rooted in the pragmatism that telecom is famed for — and for an industry on the cusp of an unbelievable level of disruptive change, this is exactly what is required.
I talked to Ekudden about 5G, of course. But we also covered 6G (no rush on this one, he told me); the move from best effort to tailored performance; AI, in its embedded and generative forms; autonomous networks; the advent of open network infrastructure as a path to new communications service provider (CSP) revenue streams; and the coming epic transformation of enterprise and public sector infrastructure (a transformation that will eventually change the fabric of global society in unprecedented ways).
Read the full transcript of my interview with Ekudden below.
Steve Saunders: Erik, hi, I’d like to start with a really big question if I may. There’s so much change in our industry at the moment, what with cloud, and AI, and automation, and orchestration and all of that coming together. It’s a huge challenge for everybody. How are you shepherding Ericsson forward to be able to support the transition to the “new new”?
Erik Ekudden: I think you’re right. This is a big change for us but also for our customers, because the changes that you mentioned are profound, both on the technology and architecture side, but equally on the business model and operational side.
Our approach is to continue to be early in new technology. We are very committed to continuing to invest heavily in R&D, both on the product side as well as research into long-term technology solutions.
This puts us ahead but also gives us a chance to work with leading customers who help define the blueprint that others can adopt and benefit from.
But it’s also very much about working with partners in the ecosystem, and standardization still plays a huge role in that by aligning the industry so that it can still scale.
Saunders: You mentioned the business model. You are giving service providers new ways to make money via the Ericsson Open API initiative. How is that being received by the service providers and are you confident that they will be able to transition to new ways of making money?
Ekudden: Around openness, yes. And I have seen you highlight its importance in your own analysis, Steve. There is clearly an opportunity to make networks more available, programmable, open.The Open API Initiative is based on using technology from Vonage as a way to open up the network to developers and provide an aggregation layer that really makes developers empowered by network capabilities, no matter where their applications show up in the world… that is a big change, I believe.
And the interest is very much there, and you see the first commercial cases appearing from operators in Japan, and Europe; from AT&T, Verizon, AWS and others. The operators can see what they need to make the networks available in a granular, dynamic way. And it’s not always about getting the highest performance; it’s about tailored performance…
Saunders: About what now? Tailored performance?
Ekudden: Yes. Or, sometimes we call it differentiated performance. The mobile broadband subscription that a consumer has for their smartphone is good enough for many, many things in their daily life. But for connected homes, and residential access, and fixed wireless access, you need to be able to deliver guarantees in terms of the bandwidth and latency for those applications.
And this is of course also true for businesses, enterprises, and verticals like smart manufacturing and public safety… they all make very big demands of the network. So you need to tailor the network to deliver what’s required on a dynamic basis when and where it’s needed. So networks are not just best effort, they are …
Saunders: … Not just faster, smarter?
Ekudden: Yes. And this goes beyond bandwidth. If you add advanced positioning, if you add security and authentication services, if you add resilience — these are all capabilities that come at a premium. And service providers need to offer premium service.
Saunders: This move away from best effort is very interesting to me as it relates to what’s happening globally with big public infrastructure projects and government policy.
In the U.S. and the U.K., the governments are still really fixated on delivering broadband speeds to their electorate. Bread and circuses and bandwidth. So, they are still very much in that best effort mindset.
But elsewhere in the world, governments have started to embrace the idea that by adding a layer of smart cloud — with AI, automation, robotics — to physical infrastructure, everything changes. There are some extraordinary examples of that emerging all over the world now. Is this an important trend for you?
Ekudden: I think your smart cloud term captures this. We see countries where the governments have the foresight to build out digital infrastructure to support specific sectors. There is a lot of benefit for the public in doing this with remote education, for example, or remote health. It’s happening in Singapore, Taiwan, maybe to some extent parts of Eastern Europe, and in the US with public safety. I’m also very encouraged by what we see in India now with the big success their identity authentication and payment systems, which all runs on 5G.
Saunders: 5G, yes. It seems to play an important part in all of this.
Ekudden: It does. If you have national 5G standalone [availability] it gives you the platform you need to create these differentiated services. The virtual 5G network can sit on top of the same physical 5G infrastructure and now you’re creating efficiency gains, network slicing become a tool, and network APIs becomes a dynamic way to engage with the network and ultimately pay for the service that that you need.
As 5G networks go nationwide it really makes sense to combine the digital infrastructure plans of a country with whatever applications they want to drive over that.
Saunders: How long do we get to enjoy all of the benefits of 5G before we have to start thinking about 6G? I was at an event recently where senior execs from BT and Telefonica both said they didn’t want 6G to get here for a long time because they need to monetize their investment in 5G, and they think there’s a huge opportunity there.
Ekudden: Actually, I share their perspective. Ericsson has been around for a while and one thing we have learned to respect is the need for technology innovation to happen continuously, but 6G is still in the early research stage for us.
There’s really no rush, Steve, because we are sitting on a fantastic technology in 5G, and we are just starting to open the door to the opportunities that exist beyond consumer and best effort services.
Over the coming years, we will be really focused on this open 5G evolution, and this is a cloud-native journey that takes us both into the core and out to the RAN, and increasingly is powered by evermore powerful AI.
Of course, eventually we will get to 6G. We’re talking about 100 times higher performance so that uplift will address the need for high performance to help solve bottlenecks, and it will also serve some really advanced new use cases, like Gen AI at the network edge, or satellite integration, or distributed MIMO. And eventually, it will blur the boundary between physical and digital worlds.
Saunders: You mentioned AI. Do you think there is a danger that it is being over-hyped at the moment, and that it will suffer the same backlash that we’ve seen other buzzy technologies go through before we find out what they’re really good for?
Ekudden: When technology matures, it rarely happens in a linear way. At some point it becomes the topic to focus on at industry events, and part of a broader discussion in society, and expectations can get a bit silly.
On the other hand, we have to consider the results that what we call embedded AI has already delivered in network operations in terms of energy efficiency and performance in both the advanced radio antenna systems, but also in the core and the cloud. Those gains are truly important, and we would not be able to get to those without inbuilt AI.
Of course, what we’re also all talking about now is generative AI, and for that we’ve showed examples of how to improve the interaction between the network and network engineer, using a telecom specific vocabulary. I think this is the first step in something that will have a profound impact on the daily life of people working in telecom and enterprise.
So AI has multiple roles to play in the network.
One is to create a better performing network using embedded AI. Another is explainable AI, helping an engineer find the root cause of a problem, explaining the rationale behind a recommended fix, and then taking that learning and applying it to running and preventing outages in other networks as well.
The foundational technologies are similar, but the use cases and implementation are somewhat different. That’s why we have launched a number of products in this space, and also are showing a path towards fully cognitive networks where you have full transparency about what’s happening today, but also the intelligence to know how you could improve it in the future.
Saunders: I wonder at what these cognitive networks will operate in a way that is fully autonomous, without human oversight. Service providers have told me they are uncomfortable with the idea that there won’t be a human being having the last word on network decisions but given the scale and speed and complexity of today’s networks, and the potential of AI to help automate them, that seems absurd to me.
Ekudden: We are certainly moving towards fully autonomous networks. Some of the leading operators have the ambition to [implement] Level 4 autonomy by 2025. We think that’s broadly feasible, but not all functions will be automated overnight. There needs to be great confidence in this automation, and it takes time to build the trust that it actually does what it’s intended to do and that when you need to you can certainly override the system.
So full autonomy will happen at different layers, with closed loop at the service level, and at the highest level of business intent and interactions with the users. That’s where we are going and it will move fast, to your point, and with the proper controls it will give very, very significant gains.
Saunders: Erik, we still call it the communications industry, but what we’re talking about today isn’t really about communications anymore, is it? For me, what you’re building now is the global infrastructure of everything, a digital world where technology is integral to supporting and enabling all of life on the planet. Surely this is the biggest transition that we’ve ever been through as an industry.
Ekudden: I think so. When we first started to talk about this a few years ago, it may have felt a little bit futuristic. But now we are seeing governments, countries, industries following this blueprint and moving in this direction. It’s super exciting.
I hope that others around the world look at the first movers and take note and use them as a a blueprint that they could also adopt, or that they could build on if they want to move even further or faster.
Original article can be seen at: