5G Factor: OCI, AWS, Google Cloud Make Major Telco Moves, Episode 70

Hello and welcome everyone to the 5G Factor. I'm Ron Westfall, research director here at the Tuum Group, and I'm joined here today by my distinguished colleague, Steven Dickens, our chief technology advisor here at the Futu and Group. And today you Say all the nicest things, Ron, you say all the nicest things.

Well, I appreciate that. Yeah, it's been a, a little while, but it's great to have you back. And that's because we're going to be focusing on the major 5G ecosystem developments that have caught our eye.

And today that includes, uh, for example, uh, the fact that we have Motorola and Google Cloud getting together on a partnership. And naturally we'll drill down on that. Also, we're seeing Oracle Cloud infrastructure, basically throwing its hat into a new telco offering.

And then finally AWS basically rolling out a global roaming capability. And so I think all these things are going to be things that people wanna know more about. And, uh, so with that, you know, thank you again Steven for joining the 5G Factor.

Welcome back. How have you been bearing up with the onset of summer? Well, It's good to be finally in these four walls.

Good to be home and not on the road. We got to hang out recently at a couple of events, which is always good. But yeah, just a busy schedule, not traveling now until the 4th of August.

Well, traveling with the family for vacation, but not traveling until the 4th of August. So very glad to be home. Well, amen to that.

Uh, likewise, I'll be doing a bit of travel in July. We have a network, uh, field Day event coming up with, uh, our tech field day team. Mm-Hmm.

So that's NFD 35 in Silicon Valley, July 10th, 11th. And I'm glad you mentioned August 4th, because this is an opportunity to plug y you'll be on the road, if I'm not mistaken. It's Tech Field Day extra at Share Kansas City.

And that will include, uh, key players such as Popup Mainframe. And I'll hand it back to you, Steven, you know, what else is going on at this Tech Field Day Extra? Yes, it's the first mainframe field day, mainframe field day one.

So bring in the main, the, the field day format to the mainframe space. We've got a few vendors who are looking to commit still, but it, uh, should be a good event, should be a good event, different community than where we've taken field days before, but that community's crying out for it, so it should be good. Well, I agree and I think, uh, the timing is pretty good.

And I know we talked about this before July, August. Yeah. Many people are on vacation, but it's not going to conflict with say, five other events.

The only people who won't be able to join in the real time since are the, uh, people happen to be on vacation during that, uh, two, uh, three days. But the otherwise the people who are on vacation, you're gonna have I think, a pretty good, uh, undivided attention and that will be most welcome. And speaking of something that's most welcome, we saw that Google Cloud and Motorola have announced a new multi-year relationship to bring Gen AI to Motorola phones.

And that includes their brand new series of razor smartphones. I think when you hear Motorola smartphones, you automatically think the Razor brand name, and that's not, uh, just, you know, the smartphone itself. It's really the new flagship Motorola Razor Plus and the Motorola Razor 50 Ultra.

So this is pretty interesting and I think exciting as we talk more about it. But I think what's also important to note here is that Modo AI is playing a leading role here. This is technology that's built using Google Cloud, vertex, ai, Gemini, and Imagine models.

So this is really, you know, I think cementing why the relationship is going to have, I think, legs going into the future. And what is, I think, important detail wise. It's now integrated across native smartphone apps and it also prioritizes bringing users an elevated smartphone experience, which you expect with a Google Gen AI enabled capability.

And also, you know, the ability to complete tasks, uh, in a more efficient manner, provide, you know, insightful suggestions, summarizing information, and so much more. And so, uh, what I think is also important to take away is that Meadow AI is looking to augment that smartphone experience by enabling that context to where personal and collaborative throughput, uh, uh, throughout the phone and the overall ecosystem, and naturally that can improve throughput as well. But, uh, what I think, uh, is going to be, uh, important is that first of all, there is magic canvas, which creates images from basically the user's imagination and can, uh, descriptions into graphics that could be used in messages on social media as wallpapers and more.

So this is actually futuristic stuff. Things like, uh, we've seen on, uh, you know, sci-fi, uh, movies and so forth and like Minority Report and so forth, actually becoming reality today. And this is thanks again to Motorola and Google Cloud's collaboration.

And the second one is called Style Sync, which syncs your phone's, look to your personal style, and that creates wallpapers and themes that, you know, quite simply matches the user's preferences. And so with that, I don't wanna steal all the thunder, Steven, what are your initial impressions about, you know, Google Cloud and, uh, Motorola getting together on the smartphone side of things? Well, I think there's a couple of key trends that stand out for me here, Ron.

Small models. Yes. We, here are a lot about large language models as the backends to sort of cloud and SaaS services.

I think where I see a lot of, um, perspective going, I'm seeing this with granite from IBM as well, and I was literally just getting a briefing from those guys this morning. I think this is on trend as well from Motorola and, and that, and Google here, the ability to execute a small, maybe that's a two, 3 billion, 3 billion parameter model on the device. Very fast turnarounds.

We're seeing this with Qualcomm and those guys as well with some of the stuff they're doing at the silicon level. So I think the first piece is small language models as opposed to large language models on device. That's the key kind of technology thread that comes through for, for me.

Then you've got a factor in that Motorola is a huge name with a US frame of reference. We think of Samsung and we think of Apple in a pretty big dog fight between those two. You go to other parts of the world, Motorola is number three in market share.

So the, you know, got other parts of the world. It's not an Apple, Samsung, duopoly, Motorolas, you know, some markets got second place. So I think you've gotta sort of take yourself out of the US centric view.

Then the other piece for me that's sort of, yes, this is consumer focus. Yes, there's a consumer component to this and, you know, making these devices cool and interesting is, is fun in a consumer context, but what's the enterprise thread for this? You know, Motorola is part of Lenovo that pocket to cloud type of story.

There's a lot of devices in the Mo Motorola portfolio and the Lenovo Edge portfolio. So the tie up for me strategically important, you know, what are you doing in retail use cases, putting devices in the hands of, uh, associates in retail, what can they do with smart retail smart shelf edges? So whilst this is a consumer focused announcement, and it's great to see, and I think there's a lot of cool things that you can do.

I'm supposed, you know, more an enterprise guy, I'm sort of thinking what does this mean from a tie up between Motorola stroke, Lenovo at the edge and a collaboration with Google? And I think there's some good trend lines there. So I think whilst I need to get my hands on one of these Motorola, uh, sort of razor devices, whoever's watching this in the Motorola team in the Ron and I need some new kit to play with.

But all joking aside, I think whilst the consumer side of, it's really interesting for me, I'm looking at the, the small language model. I'm looking at the pocket to cloud integration from a Lenovo wider edge portfolio, and I'm thinking global market share versus US market share. Excellent point, Steven.

I think it does underline just that, you know, on device AI is gonna be integral. We send That with PC right now. I think you're gonna see it with edge.

This is pushing it to the far edge, if you will, with consumer device. You are gonna see it across the whole. So I think the, the small model versus large model is, is the really, and I mean Qualcomm have been big on this.

Few others have started to come out. We've started to see the first shipments of the copilot PCs. You know, there's a whole trend line here and this is definitely on message.

Oh yeah. Uh, it's definitely gathering momentum, grabbing headlines. And I think to your point, Steven, yeah, it is consumer centric, but I think it immediately impacts, you know, for example, small, uh, business owners, you know, basically who have to use, you know, a razor smartphone to conduct business on a regular basis, Uber driver, you get the whole idea, you know, anybody who needs a smartphone is going to benefit from this.

And it includes, you know, employees of enterprises bring your own device scenarios and so forth. So I think this is something that has, you know, clear upside because Google Cloud can present a full stack gen AI capability, you know, a lot IE it's native and it's all in-house. And the other op uh, concluding observation I'll make here is that I'm seeing Google Cloud stepping up.

Its, uh, telco focus and marketing, and you might ask why that might be. Well, recently we saw Microsoft basically pull the plug, uh, on Azure for operators. And so as a result, this can be an opportunity for Google Cloud, for AWS for Oracle Cloud and others to step up and say, Hey, you know, uh, we can address telco needs.

You don't have to worry about us. You know, uh, basically, uh, pulling back, you know, when some of the going goats tough. But I think the other part is the perception is that those telcos who are relying on Azure for operators, IE the metaswitch and the Affirmed networks, uh, technologies will not be pleased about that development.

And that's, I think, something that's understood. It's just a sales and marketing opportunity for folks like Google Cloud who also, for example, partnering with OCS on its AI and data platform to again, push out gen AI to a telco specific capabilities. And we saw Google Cloud announce, uh, its relationship with iNOS stat.

And so we're seeing, I think, you know, Google Cloud quite simply gaining more momentum here. And to underline it all, Google Cloud and Oracle, uh, announced a breakthrough alliance that is multi-cloud capabilities that are native and genuine. IE uh, if you're a Google Cloud customer, you can get Oracle, uh, say database capabilities natively on Google Cloud, and you can use a single interface to do that.

IE answering what the customer actually prefers rather than having to go through the complexity of, you know, multiple interfaces and so forth to get that same capability. Well, there's a couple of things Ron, I'd add to that. Sure.

You mentioned re you mentioned retail. One of the clients that we work with that is a, a, a vendor called Scandi. Yeah.

Really innovative kind of retail, um, software vendor for that does smart data capture. And what they're doing is kind of innovating on how the store associate captures information about restocking a shelf. And that's increasingly becoming AI driven.

It's using the device and largely A-B-Y-O-D-D sort of approach. They, they're sort of give maybe giving the re the retailer associates phone or they're installing an app on a phone that they own. There's an opportunity for innovation here, connecting that, having that sort of AI capability.

Is there a partnership opportunity, a scanned it, a Google a Motorola to create an enhanced experience, do something innovative, get, get that technology into the hands of the store associates. So I think that's a really, we're starting to see that become real. And there's some really interesting things you can do.

I was at A NRF back in January. The innovation in retail is huge, and that's largely innovation at the forage. Yeah.

That, those are excellent insights, Steven and I, I know we can spend more time on it. However, we as promised, we have additional topics, uh, that we want to take a closer look at. And that, again, is along the cloud theme, the fact that we're seeing AWS basically offering an outbound roving capability across its global instr uh, infrastructure.

And I think what's, uh, interesting use case is that they're actually working with sk SK telecom that is, uh, South Korea telecom's proof of concept deployment of user playing function and packet data network gateway user playing on AWS global regions to demonstrate really the practicality and benefits of this approach. And that truly, that includes AWS assets such as the IP exchange or IPX connectivity and as well as, you know, the reference architecture needed to make all this work. So as a starting point, what they're doing as K telecom and AWS is naturally conducting a proof of concept.

And what the test is doing is using Samsung's cloud data 4G 5G, uh, core network of, again, that PGW and UPF capabilities, as well as Samsung's network functions that are compatible with running on Amazon Elastic Kubernetes services or EKS, which uses our friend MTA support for separating networks for each purpose, such as signaling and user traffic. And now that's important because it helps ensure a high availability of network functions as well as EKS worker node groups that are needed to deploy across the multiple availability zones. So if you're roaming, say in Spain and you're, say, you know, from Korea, you can definitely get a much better experience with your roaming.

And that I think is going to be, uh, more popular because, you know, for the inherent reasons people would like to have a roaming capability that they have confidence in that's predictable, that's deterministic. And AWS global coverage, I think can make a big difference here. And there's more, it's not just about, okay, how can I, uh, achieve cost savings through this approach?

It's also, I think, encouraging that the proof of concept results are showing improvement in access speeds to really important global, uh, website such as a 76% plus improvement on Google. 3 seconds to less than one second in order to actually use this capability on in a global basis. 3 seconds.

4 seconds. And so, you know, with those details, Steven, what do you think about this, you know, uh, combination? What do you think is going to really come out of this in terms of AWS being offered, being able to offer a global roaming capability?

Well, I think it's, I mean, I've got some history going way back now. 20 years I'm starting to show my age. I think if you'd have said what you'd have said in a telco context 20 years ago, you'd have been lapped out of the room.

I think the lines are blurring for me for sure. So we've gone from an Ericsson in Nokia sort of centric world with infrastructure in carriers in data centers, you know, in specific data centers in 24 inch racks. You know, going back to those points and we're seeing those lines blur.

So from a cloud infrastructure point of view, we're seeing the lines blur of where is that infrastructure? Is it in the cabinet in a telco provider? Is it in a cloud data center?

Is it, who owns that infrastructure? Going all the way through to something like EKS, which is a, a Kubernetes service. I think what you're seeing and, and you track this space more than I do with sort of something like o ran and, and the, and these sort of, the, generally the trend I see is those lines are blurring between where is that infrastructure, who runs the infrastructure, and where in the services stack is the control point for the, for the operator.

I think telcos have always been one of the more technical and had the bigger in-house teams, are we starting to see that trend kind of move and then be, not outsourcing, 'cause it's not outsourcing in the traditional sense, but be able to put more confidence in the likes of a cloud provider like AWS you mentioned Google being able to move that li sort of ownership. It's not a telco managing their own Kubernetes stack. It's an AWS managed stack with EKS.

You know, I think you're seeing just some of those trend lines start to move because telcos, do they want to be in the business of running infrastructure or do they want to be in the business of providing value add services to their clients? So I mean, I've maybe stepped back a little bit. You are closer to the actual tech stack in the operator than I'm, but I see this just as part of an overall trend of just cloud adoption, finding its way into telco.

I think the two standout industries for me would be financial services and telco, where they've been more tech heavy. And if you look at cloud adoption and financial services, it's still sub 10% of their total infrastructure. I don't track it so much from a, uh, telco perspective, but I'm assuming it's around the same.

This is just on that broader trend line of telcos being more comfortable to provide services via cloud stack. That's what I'm looking at more thematically at that level rather than the, the sort of deep level, deep in the stack about the IP gateway level that you're talking about. Yeah, no, those are great insights and I think, yeah, it hits on a, a couple of, uh, key points, telco, cloud relationships on off.

It's going to, at the end of the day, come down to the individual operator and how much they want to entrust a cloud partner to, you know, basically run, you know, important, uh, stacks and applications. Neo s we could take a quick snapshot. At and t has already entrusted Azure to, uh, basically run some core capabilities, IE crown Juul type of, um, features, uh, you know, working with them.

Now on the other hand you have Verizon saying, no, we would not ever seriously consider that type of relationship. But I think, uh, what's important here is that as we know telcos, many of them are dialing back CapEx and that is something that's impacting, you know, quite simply how much influence they can have in some use cases in some areas that they can control simply by themselves. Mm-Hmm.

So what that means is working, uh, quite simply more with clouds. I think that's something, there is a consensus there. From my view.

I think that we'll see more of this hybridization, if you will, of telcos and clouds, uh, partners getting together. And I think, you know, there are a couple of underlying, uh, factors that I'll spotlight. There are many of them, but one is, you know, the time to value aspect.

Yes. Uh, working with a cloud provider like AWS with the global infrastructure can save, you know, that ability to get a new application of service out, not only within a, a native, uh, customer footprint, but also potentially be able to, uh, support customers in the other parts of the world with a, a better value proposition. The other part that I think is interesting that it's still misunderstood is that 5G is still in its nascent stage.

And why do I say that? Well, if you look at GSM stats, basically most of the deployments that support 5G are still in their non standalone implementation. That is 4G and 5G are combined together in order to get, you know, at least initial 5G radios out there up and running.

But they're running on, uh, 4G cores. And if I recall, uh, correctly, less than 10% of the operators out there who have a 5G implementation are basically, uh, not on a 5G standalone basis. That is, they're combining the best of 5G core with 5G radios.

The rest of them are still in non standalone, but are vesting and kicking the tires on getting to that 5G standalone. And that's gonna make a difference for something like global roaming and all these other capabilities we've heard a good deal about over, you know, the last few years. Things like network slicing and so forth.

And so I think that actually can help help the SKA telecom and AWS relationship, but also bringing on more operators to say, Hey, global roaming is something that can be, uh, really broadly available and it would work on a, a more efficient, uh, uh, basis with, uh, better value. And, and so I think that also is tying into, uh, the next announcement that jumped out, and that is Oracle announcing the availability of Oracle Cloud infrastructure or OCI Telco Landing Zone. And what that is, it's a deployment tool designed to establish a secure and scalable foundation for, again, telecommunications workloads.

And so it's specifically geared to the needs of the telco. So why now? Well, what it's doing is installing OCI services and telco specific components that are needed to support the setup and operation of telco workloads.

And that includes 5G core cloud native network functions or CNFs, as well as virtual network functions or V nfs, which have been around longer. Now, OCI Telco landing zone deployment and capabilities can make it easier to accelerate project timelines, ensuring, again, that time the value that we touched on. Now of note, the packet core, which in this case comprises the user plane function, has specific requirements that are important to note here.

That includes single root IO virtualization or S-R-I-O-V mti, which we already talked about, which is A CNI plugin for Kubernetes and data plane development kit or DPDK an acronym that's been around for a while now. These are, uh, requirements are addressed, uh, during the Oracle container engine for Kubernetes cluster deployment phase of the OCI Telco landing Zone setup. So in other words, Oracle Cloud can basically offload a lot of the pain of actually getting telco workloads onto a cloud.

So lessons are being learned, we're co they're coming up with, you know, ways to streamline and make this easier by including again, you know, the Kubernetes, uh, management piece, the CNF and VNF support all basically having all this more thought through now the key features, the benefits like, okay, great, that's in place, but what will the typical gain from this potentially? Well, first of all, it helps with structural, uh, structured com, compar, mentalization, say that five times. Easy for you to say, Ron, easy for you to say, I would've been better off just saying tenancy.

It's another way of saying Yeah. Yeah. Compar, Compar, mentalization is a, is a stretch for any of us, I think.

Right on. Yeah, we're getting closer to the holiday. You can tell.

And, uh, so this setup ensures, you know, again, that separation of ownership and control over resources. Obviously any, every telco out there wants that control over the resources, over the brand identity and so forth. Also, it's helpful, uh, for, uh, flexible VCN configuration, that's the virtual cloud network capability that it can be done in the standalone mode.

And basically, uh, again, ensuring that there's granular control over network traffic, isolating, uh, you know, all the capabilities and so forth. So again, uh, the telco has the confidence of controlling their own destiny. Working with OCI also, it's important to have the orchestration of cloud native workloads with OKE or oak, that is the OCI Telco landing zone architecture.

That includes that as a core component for orchestrating the cloud native workloads. And that can't be emphasized enough. It's like, okay, it's, uh, basic checklist to have these capabilities, but you have to have an orchestration platform in place in order to ensure that all of these moving parts can work together in an efficient and, uh, you know, capable way, secure way.

Uh, and that is obviously, uh, required. And finally on, uh, performance optimization, this is where it can make a real difference because it's combining those capabilities we talked about, but it's also enabling that efficient network input output that allows for that high performance data processing and lower latency crucial for delivering and a better 5G experience. So this is actually kind of looking at AI in terms of the, uh, back office, the operations side.

Yes. ai, you know, that is non gen ai, AI is making a difference in terms of automating these operations and so forth. And so, yeah, this is, I think a, a very important announce because, you know, we've been following Oracle Cloud, certainly OCI mm-Hmm.

And so from your perspective, Steve, and what do you, uh, find about this announcement that, you know, leaps out? Well, it's interesting you mentioned we're tracking OCI, I mean, I think if you'd have gone maybe three years ago, it was three hyperscalers that we were talking about. Maybe A IBM would've been in that equation in, in certain places.

But I think OC, the, the Oracle cloud infrastructure team are doing a really good job. There's a lot of, um, deployment going on that I've been tracking them from a sovereign cloud perspective. They're doing a good job of sort of bringing those local instances.

But the big trend line that I think is sort of hitting for me with where Oracle is, it's, there's a sort of tension of do I bring the AI to the data or do I bring the data to the ai? I think if you were sort of looking at Oracle over the last 30 years, they've been the home for a lot of the data, though, their database platforms, heat wave, Exadata, exascale, you know, whatever, just Oracle 23, Oracle 23 ai, you know, whatever it is from their portfolio. There's a lot of data.

A lot of telcos will have that data deployed into an Oracle database, whether that's a backend for a billing system, an O-S-S-A-B-S-S system, whatever it is, there's likely to be an Oracle database in there somewhere. So as these telcos look to kind of expose AI to that at a really high level, it's, well, do I take all of that data and move it to a public cloud service? Egress fees, ingress fees, you know, there's a lot of data sovereignty.

Where am I gonna move it? How am I gonna secure it? I think Oracle's benefiting from the overall trend line here, which is let's move that AI toward the data.

Maybe we need it connected to the cloud. 'cause that's where we may be some of the inference and training is happening. But if we're gonna run the database and we're gonna put that AI from maybe more from an inference point of view against that data, I think there's a trend line here.

And where I see OCI sort of more generically sort of progressing is 'cause they're tethering those exascale and exudated platforms to the OCI platform. And it's really saying to, uh, enterprises don't move that data towards the ai, move the AI towards the data. So I think when you overlay that with a telco lens and say there's some huge systems from an OSS and BSS perspective that have got Oracle backends, you can take that overall train trend line of what's got going on with OCI and moving data and moving AI to the data.

And then you can put a telco lens on and go, one of the biggest sources of data is these big telco companies. So it just makes more sense for me. And, and that's what I'm seeing from OCI, and I know you track the telco space more than I do, but we kind of, one of the areas we overlap is around Oracle.

Um, so I see that kind of trend line coming through and I know we've done some good stuff recently with our coverage of heatwave and Oracle 23 AI and those types of things. So I see, I don't track the telco space as much as you do, but I see from an Oracle and Telco and a data and AI perspective, I see that same trend line that I see across a lot of the rest of their announcements. Excellent points.

Yeah, I think, you know, these capabilities and the Oracle portfolio basically impacts every enterprise organization out there and telcos are no exception mm-Hmm. So it's good to see them spotlighting the telco capability, certainly the, uh, the new capability that came out, heatwave gen AI that could be distributed across the Oracle portfolio across different platforms and basically turbocharge the price performance, you know, benefits from using gen AI with that existing data to your point about, you know, bringing AI to the data and so forth. And so I, Well, we've seen that with stuff like that, bringing Vector capability, rag capability, not having to change the underlying database structure, but bringing new AI feature function to that existing data that can be huge for teams.

Is it, do I deploy a new vector database to get AI or do I bring Vector into my traditional, uh, Oracle database? That's huge for a lot of people moving that once you have to move data, it's almost game over from a security and cost perspective. If you can bring that capability into the existing database, you're in such a better shape.

You know, it, it's painful and yes, you're right. A basically it's vectored data based done right. You know, and so this is something I think will, uh, again, prove a, uh, a differentiator and a competitive advantage.

But I think what's also interesting about this announcement is, you know, the and part, you know, I think this is gonna be intriguing, you touched on the fact that OCI has definitely invested in sovereign cloud, so I anticipate that there will be quite simply, you know, situations, you know, countries where there'll be overlap of OCI Telco landing zone with, uh, Oracle Sovereign Cloud capabilities. So exactly, uh, stay tuned on that. But also what I think is, uh, uh, interesting is what about, you know, the role of Oracle communications, you know, the, uh, unit that, uh, is developing, you know, the, uh, telco applications capabilities as well as, you know, the partnerships, uh, you know, with, uh, multi-cloud, uh, implications that you already touched on, you know, the Google Cloud relationship, but also, uh, they have the same thing with Azure.

And I think that's, to your point, Oracle Cloud has been a welcome entrant into the overall cloud space because they basically, I, I believe have, you know, energized the segment by saying no egress fees. Well guess what? Now almost all the hyperscalers are not doing egress fees anymore because that pain that we talked about in terms of moving data around and having to pay for your own data, like know, running a storage unit and having to pay every time he takes something out of the storage units, it's like, you know, that that was pretty, uh, uh, egregious for some.

Glad we've seen that change for sure. The market certainly needed that E Exactly. And we touched on the multi-cloud aspect.

Really Oracle's been integral in driving that because customers say, Hey, why can't we have, you know, multi-cloud implementations that are commonsensical and that work, you know, like, uh, you know, enterprise networks and so forth. And so I think we're, uh, much closer to a, I don't wanna say universal multi-cloud, but certainly a broader multi-cloud ecosystem that is responding to customer demands. And with that, thank you so much Steven for joining.

It's always great to have our cloud guru on and, you know, providing insights on, you know, the latest developments in the 5G ecosystem. You always say the nicest things. Ron, I love coming on your show.

No, it's a pleasure to be here. I think for me, some of the intersection of telco and cloud's crucial to talk about. So always a pleasure.

Yeah, I think we're already queued up our sequel here, you know, what about next steps? Always glad to come on your show, buddy. Always glad to come on your show.

Well always most welcome. And with that, everybody, thank you for joining the show. Don't forget to bookmark, you know, the 5G Factor as well as the future and group in terms of being able to, you know, uh, view our, uh, show with, uh, convenience.

And, uh, with that, everybody have a great 5G and multi-Cloud Day.