Tantra’s Mantra Podcast – Episode 35

Prakash Sangam:

Hello everyone, welcome to another episode of Tantra’s Mantra where we go behind and beyond the tech news headlines. I am your host Prakash Sangam, founder and principal at Tantra Analyst.
As we all know, just a couple of weeks back, we had Mobile World Congress, the world’s biggest mobile show. Many of us did our yearly pilgrimage to Barcelona to attend it. I had to unfortunately cancel my trip at the last moment because of some personal emergency. Sorry to have missed the action live from Mobile World Congress, but in fact, I was checking out and following all the happenings online being here in San Diego in the US.

So Mobile World Congress this year was back to its old glory with more than 100,000 visitors as mentioned by GSMA. The event also gave a peek into the future, especially on how the future of 5G is shaping up. All of us were hoping that we could get answers to some of the pressing questions such as when will the most awaited next phase of 5G will arrive? When will all those exotic 5G apps and services we have been talking about for years will come? How and when will 5G operators start to see the return on the enormous investments they have done for their 5G deployment and so on? I’m not quite sure we got all the answers, but we at least got to see where the market is headed and so on.

And to discuss all of that, I have John Smee, SVP of Engineering and Global Head of Research at Qualcomm with me.
John, welcome to the show.

John Smee:

Thanks very much. Great to be here.

Prakash Sangam:
Perfect. So Qualcomm is an undisputed leader in wireless and has been a driver for new technologies and new standards for a long time. What you guys are cooking, John, in your labs today is what we will get in our hands in the near future. So our listeners are really eager to know what you guys are really cooking in your labs right now.
So let’s get started. So let’s start with your background.

I have known John for a very long time, going back to my time at Qualcomm.

So John, I think you spent almost a quarter century at Qualcomm, right? How has been your journey?

John Smee:

Very good. And that’s exactly right. I’ve been here around just over 24 years. So I started right at the very beginning of January 2000, and it was a very exciting time to join Qualcomm when the industry was fully digested CDMA and it was starting to evolve. And so I jumped in, actually, first working on helping bring data into cellular as part of EVDO. And that evolved, obviously, with the ecosystem and WCMA setting the stage after that.

But absolutely, it’s been a fun, long journey for me.

Prakash Sangam:

CDMA 2000, I still remember those good old days. So I know, I mean, the telecom industry, as we have seen, has been in ups and downs for some time and it is seen in great days as well, coming after the troughs. So I think it’s in a little bit of a flux right now, especially after successful 5G EMBB and ANS Mobile Broadband deployment. I mean, it’s worldwide deployment.

People are really enjoying the benefits of 5G speeds and capacity and so on. But we still are not very clear on where the next phase of growth is going to come from, especially when and what some of the applications and services we get to see.
What do you think?

John Smee:

I think we’re starting to see more and more different types of devices. I think even looking at the launch of those very initial 5G smartphones 2019, 2020, and then now fully coming out of the COVID era, I think that was evident also in Barcelona, this that kind of physical energy to the show.

And very sorry to not see you there in person, but it really was kind of back to that industry expansion feel in terms of not only obviously global operators and vendors, but also just from the technology and ecosystem standpoint, a lot of enthusiastic participants saying, hey, where’s wireless going next? Where’s 5G going next?

And so we did see some of that renewed focus on new types of devices. And then also the interplay, I would say between consumer and enterprise. I think that was an underlying theme that in terms of how the global communities are looking at that continued 5G expansion and evolution.

Prakash Sangam:

Indeed. So I think enterprise is going to be key for 5G to go forward, right? I think, of course, consumer is a huge business, but I think the new capabilities that are coming in are really key for enterprises. So from that context, what are some of the key takeaways that you got from Mobile World Congresses?

John Smee:

I would say the key takeaways for me was really this perspective that not only was every single booth, including Qualcomm, is really advocating for the opportunities of AI, but it was also a recognition that AI is really not just a cloud story.

There’s a huge focus of on-device AI, and then that brings in this perspective of what is in the user’s hands? Is it a 5G-connected laptop? Is it a smartphone that has amazing AI capabilities? Is it a wearable pin? Is it something that, as an example from the company Humane, is it something where a smartwatch, a XR, new types of glasses?

So there’s an interesting kind of energy around the fact that we’re all still living in this physical world, and we’re all going to be more connected, but that connectivity is really intertwined with data and AI. And so for me, the kind of takeaway from the show was the infusion of energy around the fact that communications and computing are more intertwined now than ever. So it’s not just about secure connectivity to the cloud.

It’s really more about the fact that there’s going to be a lot more computing on a much wider type of devices, and that people are going to be using those devices in new creative ways because AI itself is opening up new productivity enhancement applications.

Prakash Sangam:

Indeed. And some of the new use cases in applications, especially 5G is really famous for low latency and such, need standalone, right? And the standalone progress has been really slow.
I think that’s in a catch-22 situation where app developers are looking at wider deployment of 5G standalone and operators who have done enough investments in NSA, looking for enough groundswell of support for the need for 5G standalone kind of use cases in applications to deploy it widely. So how do you think we get out of this gridlock?

John Smee:

I think that was also another positive coming out of the show, because I was challenging many operators in all of our discussions in meetings. And the answer was that there continues to be a lot of expectation that SAE is actually here now. So there was a lot of feedback that there is a lot of investments being made right now and that SAE is indeed coming online in more and more markets.
So I was actually positively excited to hear that that investment in SAE is happening. And as you say, it’s kind of intertwined with the fate of RedCap, which relied on the 5G core. So the good news is with the kind of those initial deployments, getting that ran out there for release 1516, that kind of pivoting back to operators looking more holistically at their networks and saying, hey, we now want to move to the 5G core, recognizing that it brings with it slicing and other capabilities and also just a more modernized architecture.

So kind of the evolution from 4G into 5G on the core side is also just about the overall operational efficiency of these networks going forward. So I do feel that that migration to SA is indeed well underway and that that’s going to continue to open up new interesting opportunities beyond just consumer mobile broadband.

Prakash Sangam:

Indeed. So when I talk to operators, the challenge for them for SA was there are all of these, there’s quite a few of these niche enterprise applications which on their own look really impactful and interesting, but they did not have enough of them to, that will kind of propel the whole network upgrade to stand alone. So that was one of the challenges that they had so far.
And I think RedCap that you mentioned, it’s one of the 5G Advanced features, seem to have that kind of one traction and broad interest wherein it will create the groundswell for the network-wide need for stand alone. And then that might make them switch to stand alone quicker.

What is your view?

And also, RedCap is one, would love to hear the feedback that you guys have gotten from the ecosystem about RedCap. I was actually surprised by how much traction there is, but would love to hear from your point of view and also other 5G advanced features that are seeing some interest from your ecosystem.

John Smee:

Yeah, so it is exactly that point that as the operators are looking forward, they’re starting to say, hey, as I’m moving my network to stand alone, then I’m starting to realize there’s more different types of devices I connect with RedCap. And I would say that’s one of the interesting things from the module side, the power consumption side, and even the very reasonable data rate capabilities, like a couple hundred megabits per second. So this RedCap is kind of fitting a very interesting niche of bringing cellular connectivity to new types of applications.
And so it’s something whether it’s looking at different types of tiers of FWA, different types of tiers of even connecting laptop devices, connecting more tablets, you’re starting to see that price point bring in just more connectivity into cellular itself.

And I think that that’s the part of the energy where, yeah, Qualcomm launched the SDX35 last year, but what was interesting from the design-win standpoint, from the pipeline of interest, we’re starting to see more and more creative solutions entering the market of, hey, I can actually bring cellular connectivity into this application, do we get on a 5G going forward basis, leveraging the fact that we have RedCap. So I think there’s that enthusiasm to be investing into the future. And that’s in some sense what RedCap represents. It’s that carrying on the journey with 5G as 5G continues to evolve and the networks move forward with further improvements that you’re on that 5G RedCap train.

Prakash Sangam:

Yeah, I mean, even in the 3GPP, when the discussion on RedCap started, variables were one of the hero use cases, right? But talking to operators, I mean, they’re using, I was talking to somebody in the Middle East, they’re using RedCap to kind of looking to use, RedCap to even tier the fixed wireless access. So I was really surprised to hear that.

And then, you know, getting 5G into or even cellular connectivity into laptops has been an upward and uphill battle so far.

But with, I think, RedCap, the price points are coming such that integrating connectivity into laptops and making them connected laptops, 5G connected laptops, seem to be fitting the bill up pretty nicely in terms of how cost effective it is and for the performance it can provide and so on, right?

John Smee:

That’s exactly right. It’s hitting this kind of mass tier volume calculation that it’s bringing, as we say, cellular connectivity to applications where you don’t want to just be on Wi-Fi and roaming and logging in. And obviously, as you mentioned, even new markets for FWA where it’s not a gigabit per second market, but it’s more, hey, 200 megabits per second can really support a lot of different application scenarios for FWA as well. So we are starting to see that kind of more and more design wins, more and more creative solutions to execute on the RedCap opportunities.

Prakash Sangam:

Indeed. So another interesting thing in 5G Advanced is the introduction of AI, right? And there are a lot of questions to begin with. And it’s only starting with channel estimation, even there, how the cross learning between the network and the device would work.

How is what is kind of the framework for them to work together, right? Without, you know, not against each other and so on. So how has that been coming along in Release 18?

John Smee:

Yeah, so Release 18 did a really good kind of end to end study on the opportunities for bringing, you know, AI into the air interface, right? Because we’ve always had the ability to use more and more intelligent algorithms, whether it’s smarter networks or, you know, devices that are able to improve modem performance by bringing in AI concepts. You know, I love the traditional transmitter and receiver algorithms.

I think what’s exciting now is that 3GP has really jumped in full steam ahead looking at, hey, how can we actually change the specification, you know, related to, say, improving feedback efficiency. I think we both know that when we look at the spectral efficiency gains of cellular in a world with massive MIMO, in a world with millimeter wave beamforming, there’s also a lot of exchange of side information, channel speed information to enable the actual data communication pipe itself.

And that’s a great example where machine learning can help move from a more stochastic framework of modeling things into more of a predictive approach. And that can really cut down on overhead, which can then enable us to more efficiently use the network resources, the air link resources. So I do fully expect AI to continue to be a hot topic in standardization, even as we pivot all the way into 6G designs and looking at a more AI-native architecture end to end.

Yeah, so with Release 18 starting off, you think you have found a good way how the cross-learning between the device AI, which is running its own algorithms and so on, and the network running its own algorithm coming from different vendors and so on, and usually the device and the network are coming from different vendors, and I’m assuming the vendors have their own models that they’re not public and so on.

Prakash Sangam:

So how is that cross-learning between that working out? You think you guys have a good framework, you work on your stuff, the network guys work on their stuff, you have good channels to exchange information and work towards an additive overall improvement and so on? So that you set a good foundation that the future versions of AI can do even better and much more features and so on?

John Smee:

That’s right. It’s really about, as you say, it’s about partitioning the opportunity for improvement. And instead of having separate optimizations that are happening in a very disjoint way with techniques like Cross Node, and that was actually one of our technology demonstrations in Barcelona was one of our interoperability Cross Node research projects that we did with Nokia.

And it was an interesting one because we were then able to look at some of that dynamics around what’s being exchanged and how that partitioning then of what’s happening on the device side relative to the network side can enable performance improvement and kind of a new approach to this kind of end to end challenge that’s been inherent in wireless communications and information theory. This kind of huge framework around transmitters and receivers and channels and this notion of separability versus kind of mutual solution.

And so I think it’s really an exciting area. And what’s also important to remember is the fact that AI itself is changing so quickly. So not only just generative AI and text to image type applications, but the way in which the field has evolved in a mathematical computer science context, really I think that on the research side of Qualcomm, and I know many other companies as well, it’s something when we’re able to then iterate with our AI teams and come up with new interesting applications all the time.

So one of the kind of fun parts about pushing wireless forward is that the way that wireless and AI are intertwined, that’s actually changing at a fast pace. So even though wireless had a huge amount of improvement from 2G to 3G, 3G to 4G and 4G to 5G, what AI is much earlier in its kind of evolutionary arc, if you will, and that makes the challenges in front of us really about how to effectively combine the two, given the fact that AI and the role of on-device processing, power-efficient, and neural processors, in addition to GPUs and CPUs, it’s changing where data is created, and it’s really also changing the requirements on communications.

And so we are indeed very deeply embedded into that research path, because it’s that coupling between the two that make the problems more interesting.

Prakash Sangam:

So when look at Rel. 19, 20, and beyond, what are some of the things, specifically from AI point of view, that you’re working on?

John Smee:

Think of things like RF sensing, and sensor fusion, and ability to really understand in an environment to create, for example, a digital twin of the physical world. That’s a good example where AI and wireless can really combine to enable a more immersive communications opportunity that is incorporating real-time knowledge of the physical environment.

So I think that what AI brings into wireless is also that ability for wireless itself to move beyond communications into what we call integrated sensing and communications. And so the role of ranging, positioning, versus actually better understanding a 3D physical environment, whether it’s an indoor factory or hospital, or whether it’s an outdoor scenario for connecting vehicles or for outdoor augmented reality.

I think that’s another way in which people will be interacting with their physical environments is also changing quickly from a meta world standpoint. And so that kind of 5G evolution, AI evolution is also on the R&D side, something that we’re taking forward in terms of how we’re looking at the 6G interface.

So 6G, I mean, it’s never too late to talk about next generation. Some people might be asking, we still have to realize the investments that we have done in 5G. But from a R&D point of view, technology development has to start right now for us to see what 6G is in, say, 2030 and beyond. I expect to be a little bit later than that.

Prakash Sangam:

So what are you guys working on as we speak now?

John Smee:

When we’re looking at 6G, we’re actually really looking at it also in the context, you know, of pushing 5G advanced forward. I think even what we’re studying in 3GP right now on topics like AI and sensing, looking at new ways of using different frequency bands, some of that’s going to be applicable to 5G advanced commercial deployments that happen within the next five years, and some of them will be incorporated, you know, into a new interface targeting, you know, 6G deployments that begin in 2030.

So when I look at 6G, it’s really about understanding, hey, what can we do to improve the overall operational efficiency, the energy efficiency, the spectral efficiency, not only in new bands, but also in legacy bands, you know, FDD and TDD bands, you know, bands below 1 GHz, bands, you know, from 1 to 7 GHz, what are some interesting things that are happening there, versus what are some new ways to bring in new frequency bands, you know, between 7 and 8 GHz or 10 and 12 GHz.

So there’s a lot of interesting work in 6G that’s about making sure we’re really improving the cost-effectiveness of deploying and operating a cellular network. And I think that’s what’s going to help the industry also move beyond communications into a world where there’s the opportunity to bring value and monetization into sensing, into AI, into new types of more immersive communications part.
So if I had to kind of summarize it, it’s around kind of two major core themes of improving TCO and OPEX from a communications efficiency, resiliency, kind of green network energy consumption side. And then there’s the part about new opportunities and AI and compute integrated sensing.

And so what’s interesting in that broader context is, as we say, it’s only 2024 now. And so when we look at 6G technology that’s going to be 2030 to 2040, there’s a huge amount of interesting opportunities that lie in front of that from a more connected smart city.

And going back to our earlier chat here about the role of enterprise relative to consumer. So 6G is not really just happening because consumers need a new smartphone in 2030. There’s going to be so many other things that are part of that overall efficiency improvements in the cellular networks.

Prakash Sangam:

Yeah, that’s exactly right. So far the cellular networks were designed and standardized for public consumer network and use. I think going forward, it will be even the network architecture and the deployment point of view has to change as well. It will be more custom-built, maybe one for enterprises, one for consumers, one for others and so on.

I think the whole concept of building very large, country-wide, single kind of application networks will have to change as well, right? Because there’s so much of variance in what people use this technology for, isn’t it?

John Smee:

That’s right. And it’s interesting, because what we kind of talk about is the surface area of 5G. And if we look at the surface area of cellular in a 3G era versus a 4G 2010s era to a 5G 2020s era, absolutely, the surface area of cellular is so much bigger, which is one of the reasons that whether you’re talking automotive or you’re talking industry or you’re talking embedded IoT, there’s a lot of interesting intersections that map back to, hey, I want to have cellular connectivity.

I want to think about my network in a different way. I want to be connecting and bringing value of that connectivity into new end-to-end applications. And I think it’s that energy and enthusiasm that in some sense underscores why there’s that global interest in continuing to drive from 5G into 6G, both in an academic context, in an industrial policy context, and obviously for companies like Qualcomm, in a huge innovation, pushing technology forward context.

Prakash Sangam:

You think 6G will have a new modulation scheme, or you think we are good at OFDMA for another generation?

I think OFDMA has a lot of inherent advantages from an efficiency standpoint, a time-frequency slicing, spatialization opportunity standpoint, in terms of that. So I do think the waveform will largely be OFDM based.

We often look at some tweaks to that, very high frequency bands, where we are really concerned with peak to average. So we look at single carrier waveforms for that sort of scenario.
But OFDM is a real workhorse, 4G, 5G and likely into 6G, because of that flexibility that it affords from that time-frequency standpoint.

So short answer is OFDM will absolutely be part of 6G.

Prakash Sangam:

That was my last question. Any closing comments?

Yeah, so in summary, I think we’re at that interesting point of inflection where 5G is out there today, operators are focusing on that transition into SA, and in SA, the standalone mode bringing with the 5G core opens up new opportunities for reduced capability, new kind of end-to-end efficiencies for connecting more devices via cellular.

John Smee:

And I think that’s the interesting reality of the cellular ecosystem, the device side, the network side, the operator side, and then that consumer enterprise side. So we have an exciting five years ahead of us as we’re continuing to see more applications move into wireless and at the same time beginning that foundational research into 6G.

Prakash Sangam:

Excellent, John. We can spend hours discussing this, but still have not covered everything. Thank you very much for your insights.

Indeed, it is a testing time for operators and the industry as we move from really successful 5G to the next phase and looking for where this new killer app and use cases are coming from.
And it is interesting to watch how the industry moves forward. Thank you again, John.

John Smee:
Sounds great, Prakash. Thanks again for having me.

Prakash Sangam:

So folks, that’s all we have for now. Hope you found this discussion informative and interesting. If so, please hit like and subscribe to the podcast and whatever platform you are listening this on. Don’t forget to check out the other content on website www.tantraanalyst.com I’ll be back very soon with another episode putting light on another interesting tech subject.

Bye bye for now.