Ericsson's CTO on What Really Goes Into Rolling Out 5G
In the next year, US carriers will begin rolling out 5G networks nationwide. The transition will take some time, but the technology promises to jumpstart a new era of telecommunications where everything from AR/VR headsets and Internet of Things devices to milions of self-driving cars will be connecting and interacting with network infrastructure in real time.
AT&T, T-Mobile, and Verizon are making headlines with their 5G plans, but networking and telecom giants are supplying them with the tech to do it. One of the biggest players is Ericsson, which—alongside competitors Nokia and Huawei—is suppying carriers across the US, Europe, and Asia with hardware and software as they race to deploy 5G networks.
Ericsson CTO Erik Ekudden appeared at the Techonomy conference in New York City for panels on global connectivity and demystifying 5G. He also sat down with PCMag to discuss the company's 5G infrastructure, playing in the US market, Ericsson's broad machine intelligence strategy, and what the future holds once 5G networks realize their full potential.
PCMag: You have a lot of experience with 5G. What are the most game-changing things the faster, lower latency technology brings to the table?
Erik Ekudden (EE): We're building 5G to meet the requirements of the world around us. When you start to move workloads to the network platform or the network cloud, you're not routing them only in a few central places in the world. In fact if you have a national or even local network, you need the bandwidth. You need the lower latency. 5G is the answer to that. 5G infrastructure really supports the trends that we're already seeing in terms of how we modernize enterprises, and how we modernize industries.
Ericsson has supported 5G capabilities in its hardware and software since 2015 and you've been working closely with major carriers in the US, Europe, and Asia. Over the last few years as the competition to get 5G deployed has ramped up, what does that look like logistically for Ericsson as you provide infrastructure to all these carriers racing to roll out 5G by the end of this year or early 2019?
EE: Yes, [in the US] the four carriers are sort of having a public dance. For us, this is something we've been preparing for for some time, partly because we were the one leading and driving the industry towards completing the standards by the end of last year. Now we're building the hardware that we're releasing [in the] first half of this year. That is going to be combined with the 5G standard decided on in December, and then we have the software coming [in] the second half of this year.
Many people may not realize all of the different levels of the infrastructure that you need to make 5G work. For Verizon, for instance, Ericsson is providing the core network, the radio access network, and all the transport services, correct?
EE: Yes, the transport services, but also what's called the OSS/BSS, the management system, and the modernization system. So how do you charge for those IT services or mobile bands, and then again how do you manage so that you get what you order? It's about guaranteeing bitrates and guaranteeing resilience.
Ericsson is also leveraging this kind of virtualized infrastructure with AI and intelligent algorithms to do things like dynamically allocate resources on the edge and as-a-service. Can you put that context with the company's broader machine intelligence strategy?
EE: All our networks and software will be AI-driven. We run networks for about a billion subscribers in the world, predominately emerging markets, and we can get significant efficiency gains by automating many of those processes. The idea is to leverage all the operations data that we have with these networks, where we already have a solid base of radio and core networks, and transport nodes. We've upgraded that software with advanced algorithms that use the operational data each node has to improve performance with self-optimizing networks.
Machine intelligence or network AI from Ericsson affects all our businesses, all our products, and operations with services. And we are in a fortunate position that we have access to a lot of operational data to be able to improve. I'm talking about the operational data of course, not user data.
Talk about how Ericsson is using that operational data and machine intelligence for predictive modeling.
EE: We're using the data from those automated operations to help technicians when they're out servicing. It's something we call radio assistance, or cutting down the cost and time of servicing. All of this, again, is part of one platform across companies that leverages all these huge cases for operations. When you go into more advanced user interface questions, then we're partnering up with companies that are specialized in those areas. For example, advance context-adapted word processing. We build that from components and leverage that with one set of screening data, layered across all those networks around the world. We build the knowledge band systems on top of that.
Then we also have collaborations with partners that are using AI for their applications, whether it's advanced management functions, supervision or surveillance applications, etc. They're using our infrastructures to put advanced analytics and advance machine learning algorithms all the way out at the edge. That's why we have this distributing cloud infrastructure that can host any of those workloads.
Going back to the telco landscape in the US, have some of the moves from the current administration changed the equation at all? Not only in an emboldened push for 5G, but also in singling out and banning the hardware of competitors like Huawei.
EE: We have a strong situation here in North America working with all the national carriers. We also work with the government and have discussions with the regulatory side to make sure that [the] US has [a] competitive situation when it comes to possible progressions in technology. Now, I can't comment on what some of the competitors are doing or not doing here in North America, but I think the important thing is that we face this competition globally. We have to have the best and most efficient products wherever we're operating, whether that's in Europe, China, or North America.
North America is a very important early market when it comes to LTE, which Verizon was driving. If you look at the penetration, North America is still ahead of the world. And the same will happen in 5G. So we've had to develop five different stations, advanced multi-use micro-technology antennas, and all of those things early, because we are present in North America.
You talked about this during your panel earlier, but why will the transition to 5G becoming the predominant network technology be slower than people realize?
EE: It'll be a combination of 4G in existing bands and 5G in new bands. I think it's more like a three- to five-year gradual migration. The US is launching 5G this year, and we'll have the first devices and networks in 2019 and certainly 2020. Then we're talking about a full-blown rollout.
Let's put this all together. When you talk about the sort of next-gen networks that telco infrastructure providers like Ericsson as well as all the carriers are building, what will that ultimately look like? How does a 5G network with adaptive machine intelligence built into it change how we think about connectivity?
EE: Think about the smartphone in front of you or the glasses on your face. For consumers in the next five or 10 years, the processing on these devices will be done on the network and the edge and your glasses will have functionality like AR/VR because of 5G's latency and bandwidth. I don't deny that 5G's a fantastic vehicle for consumers, but the point is that some of those things you can start to see already in the coming years with upgrades to LTE and the early 5G systems. For applications like gaming and more advanced user interfaces, 5G will make a big difference in the next two or three years.
What about the IoT side with real-time machine-to-machine (M2M) communication?
EE: It's not about M2M in the classical sense of connecting small devices to the network, gathering data, and doing analytics and optimization. We do connectivity management around the world. 5G systems with really advanced network slicing in transportation, logistics, smart manufacturing; it starts with the upgrades to LTE happening today, and the next step is when consumer experiences are 5G-powered to the extent that the form factors of the devices will change. Advanced network edge computing will allow this [points to smartphone] to become thinner, cheaper, and more capable because of the connections in it.
What about 5G networks and self-driving cars?
EE: The challenge there is on the regulatory side and the maturity of autonomous systems in general. So we've worked a lot with the US car industry, with European car makers, with Toyota...to make sure self-driving cars are leveraging global infrastructure to create a point-to-point solution between cars. What we're doing with Toyota is their initiative to introduce the 5G edge cloud so Toyota can gather data from all its car brands and create infrastructure to use its network as a 5G cloud.
This article originally appeared on PCMag.com.