Clarence Reynolds, TelecomTV (00:11):
The telecom industry stands at a critical inflection point with open ran deployments showing mixed adoption patterns across global markets. As service providers weigh the benefits of virtualization against integration challenges, the promise of AI enhanced networks and new monetization opportunities looms on the horizon. Paul Miller of Wind River joins us to examine the current state of open ran adoption and its trajectory. Hi Paul. Good to see you.

Paul Miller, Wind River (00:39):
Thanks Clarence. Great to be here and looking forward to talking with you again.

Clarence Reynolds, TelecomTV (00:44):
The state of open ran seems to be evolving. We see mixed adoption, some virtual ran, traditional AI ran and of course open ran. So what is your view on the market state of adoption and where is this headed?

Paul Miller, Wind River (00:59):
Well, it's interesting. We find ourselves at a point now where there's actually been generations of open ran or virtual ran technologies, given that we're down the road quite a bit on this technology stack. Initially what we saw was virtual ran and we have some very high scale deployments using that technology Fundamentally, that was what we call a vertical disaggregation, right? The movement from a traditional ransack with proprietary hardware and embedded software to one that used a more cloud native approach with virtualization technology running on typical cut servers and then applications residing on that either as virtual machines or as containers. And so that's taken the industry relatively by storm. It followed the NFV initiatives that happened in the core of the network and we've seen some pretty big deployments using virtual ran technology. Then the next evolution was obviously open ran and we started to see some transformation in what we call the east west or the horizontal disaggregation, the desire to use different radio units from different manufacturers than say the core assets, this sort of thing, and driving interoperability between the different vendors that were happening there.

(02:07):
And of course each of these transformations was driven by a service provider need. They're looking for better commercial power, better TCO in the network, more interoperability, better flexibility to swap vendors out if they are disappointed by them and it's really resulted in that. And then of course finally you have AI ran starting to evolve over the past year or so, and that's really leveraging, I consider that kind of a subset of what we're doing in open ran where the software applications driving the RAN can now take advantage of AI mechanisms such as dynamic beam forming and power saving mechanisms that can be used in the RAN to really optimize the solution for even lower cost and more flexibility. And these are pretty exciting development. It's now everything we just mentioned, virtual ran open, ran AI ran, these are all software defined paradigms and there is a bit of a battle going on between that and the traditional RAN and we're seeing both in deployment today. If we look back in time to 4G, we saw a hundred percent physical ran equipment traditional ran as we've been deploying 5G, that battleground has been going on. The service providers weighing as to whether they'd move to 5G via software mechanism, virtual ran open, ran, or whether they'd continue a traditional or some hybrid approach. As we move towards six G though, I think that that battle is over and we're moving completely towards a software defined universe for telecommunication networks all the way from the core to the edge.

Clarence Reynolds, TelecomTV (03:31):
So what is the actual state of virtual and open ran? Are we seeing at scale deployments with truly mixed vendors and virtualization?

Paul Miller, Wind River (03:40):
Yeah, we actually are, and it's a result I guess we've been doing virtual ran and open ran now for the better part of five or six years and that's resulted in the technology becoming extremely stable. Many people in the industry know about Verizon's huge national deployment of virtual ran and they're also moving to open ran that's in many tens of thousands of nodes that are deployed there, running live production traffic for multiple years very, very efficiently and several of their executives have gone online with public interviews stating the incredible advantages in time to market TCO network flexibility, the ease of deploying new applications and maintaining the network. So I think that's an example of a carrier that's really benefited from a highly scaled virtual ran deployment. As we move more towards open ran, we saw some adoption with Vodafone that we're involved with with our partners in the UK and greater activities there in Europe that continues to progress forward.

(04:39):
That's very, very positive. Vodafone famously ran a year ago at Mobile World Congress, a performance test where they tested traditional, ran against open, ran in the exact same physical location on the exact same days and weather and found that open ran performed better from all KPIs, throughput, latency, et cetera. And so that was really productive to see as well for market adoption. Then back in the US the latter part of last year, our business was fortunate enough to win Boost as a new customer and that's a significant transformation. That's the world's largest open ran network with tens of thousands of sites deployed and that's a true open ran deployment at scale. We've got different vendors for the radio unit as far as the DU and CU elements and the near edge and farge and network all coming from different vendors truly achieving the promise of what Open Ran is slated to provide, and in fact, even our ability to enter into that opportunity was driven by Open Rans architecture, which made it easier for that service provider to swap components out in this case the O Cloud technology and make that transformation in their network.

(05:45):
So we're seeing at scale and benefits that were anticipated from Open ran. Now at the same time, there are difficulties we see competition against traditional RAN incumbent players to move to the cloud-based architectures that open ran drives and so there's a bit of a battle still going on and slower than anticipated market adoption in some areas than what you think you'd see from moving to cloud and software.

Clarence Reynolds, TelecomTV (06:10):
Windriver does business with many service providers globally and as they wrestle with this decision, what factors are affecting the decision to adopt Open ran and how can these obstacles be overcome?

Paul Miller, Wind River (06:23):
Yeah, it's interesting. I think in the early days what we saw was a question as to whether these high performance real-time applications such as software-defined radio, this is what's really running on these high performance compute nodes at the edge of the network and leverage accelerators, relatively advanced Kubernetes technology. The question was could it perform? Could you actually run a 5G radio network on conventional servers using virtualization? That argument has been put to bed Now we now have, as we just spoke about several highly scaled production deployed networks that have been in operation for years and that has allowed not only vendors such as ourselves with the cloud technology, but vendors that provide the ran layer and application software to harden and strengthen the feature set of their application. So I think we've kind of put that to bed. The next problem that kind of emerged was the complexity associated with interop between these different components.

(07:17):
You're moving from in a traditional RAN context, you've got a completely vertically integrated solution with hardware and application software provisioning and maintenance systems all coming from a single vendor. Now that's bad commercially because the service provider is tied to that single vendor and the pricing established by that vendor. They don't have flexibility, but as you move towards the more flexible open ran solution, you then as a service provider often are integrating these components in your network and it's more complex to deal with that. You need different expertises that are more IT enterprise centric as you're dealing with things like IP networking and Kubernetes and servers and virtualization. And so there was a couple years of struggle there as major service providers developed those expertise and understood what the new technology they were grappling with and frankly understood how to integrate these variety of different vendors in their network.

(08:09):
I think we're now past that as well. And so these were really the challenges. And of course the undercurrent as this is all going on is TCO, is this going to cost me equivalent or less than my current network? Are there future benefits to it? Can I create more monetization with this type of network architecture? All of these struggles go into a business justification for building out a 5G network. And I think we're finally at the point as we exited 2024 into 2025 where basically these are completely solved across the board, we're finally at a point in the technology, the vendor maturity, the support, the services that people can bring to bear where any service provider globally can take this technology and then deploy it and it will work right, and it will work very quickly and drive TCO down for them and give them much more flexibility in their network.

Clarence Reynolds, TelecomTV (08:56):
What is the long-term situation in the market? Will six G be the defining transition point to fully virtualized and open ran? And how will the emergence of AI be used in these networks?

Paul Miller, Wind River (09:09):
Yeah, it's a really good question and I think as we mentioned a moment ago, the movement from 4G where you had a completely proprietary implementation right at the edge of the network, you may have had NFV in your core, but your entire edge was fully proprietary. Now with 5G, as we mentioned, there's a bit of a battle going on between open ran and virtual solutions and traditional RAN and the incumbents and whether they make that transformation. And that's starting to tip, right? We're starting to see, and this is very important to a service provider considering this technology now as things move towards more cloud-based solutions, we're seeing the RAN providers in particular pour their investment efforts more into the virtual solutions than they are in the traditional because they can see the transition happening in the market. That means that the newer capabilities like we're talking about with AI ran and that sort of thing are going to now first emerge and maybe not even be back ported to the traditional components.

(10:02):
So in a way, we're starting to leave behind the legacy. We're starting to make that move where the traditional components will stagnate, right? And the new innovation will be on the cloud-based systems. And as we've watched this transition over the last few years, to your question, yes, six GI think now it's the defacto understanding that six G out of the box will be fully virtualized. There will be no proprietary systems there. And so as a result, service providers and vendors need to get their heads around these technologies and transformations. And then AI in the ran that has really emerged as a very interesting impact. We certainly need to support at the edge of the network, the AI-based workloads, and that really means an inference based workload from AI perspective. We need to support the hardware accelerators such as those from Intel and Nvidia and others as well as the enablement from machine learning and other packages and libraries that are necessary to support those AI applications.

(10:59):
The evolution of the core network provisioning with the SMO and the realtime and non realtime rec that allows the deployment and management of these AI applications, that's all very important that enables AI in the ran that has a market impact in the operation of the radio system, right? As I mentioned before, dynamic beam forming and power saving and this sort of thing is the reason behind the AI transformation at the edge. We also see interestingly, a transformation in the core of the network, and this is more around, as we've spoken about, the complexity of adopting open RAN is quite high, right? You've got multiple vendors, complex Kubernetes and cloud systems. How do I manage that as a service provider? One of the ways is with AI in the work that Wind River's doing in this space takes an AI large language model and trains it on the APIs and the underlying systems and allows a human being to interact with natural language with this type of AI large language model, and you're actually conversing with the deployed live network and this takes debugging that would've taken days or weeks and turns it into seconds. And so that further drives down TCO, it makes it easier to maintain and deploy the systems. And so we see AI kind of becoming pervasive from Edge ai, the far edge of the network all the way to the core for operations and maintenance functions.

Clarence Reynolds, TelecomTV (12:17):
Service providers seem to be looking at new use cases for the network. Can you give us some examples and how can those use cases help with monetization? Also, does adopting Open RAN lead to greater monetization of networks for service providers?

Paul Miller, Wind River (12:32):
Yeah, very insightful question and I'll tell you, wind River has a very interesting perspective on this. Wind River has been in business for over 40 years and the foundation of our business was built upon real-time systems and edge systems. So whether it be automobiles or jets, medical equipment, industrial robotics, these things are all running Wind River software. And so as we've been working in the 5G network transforming open ran and virtual ran, we continue to run our business in the areas that is really edge, edge systems and OT systems. Now, what's happened over the past 10, 15 years is all these edge devices have become hyperconnected, whether it be an industrial robot or an automobile. These things are now connected to 4G and eventually 5G networks. What that means is there's an opportunity for monetization for the service providers. Why are these things becoming connected and what use cases and problems does that solve for the consumer?

(13:27):
It's pretty interesting as we look at open ran, and the latter part of your question, why choose open ran, open RAN has kind of a knock on side effect, which is although you're choosing a 5G deployment technology, you're also at the same time vying between do I want proprietary systems at the edge of my network or do I want a cloud system? And the interesting thing about bringing cloud to the far edge of a telco network is now it's a software defined environment. If I'm a service provider and I own that core to Edge cloud, I can now deploy any application I want anywhere in my network with the type of flexibility that Google and Amazon and Azure provides to you. This means that if we take, let's take for example an automotive system. There are concepts like dynamic OTA or Overthe Air update to manage the software in the automobile.

(14:14):
There is cellular V two X or cellular vehicle to anything. This allows a vehicle to talk to infrastructure such as getting information from intersections or civil changes like construction or emergency vehicle behaviors and sending that information to the cars and the cars, even sharing data between each other such as their geodetic location, their velocity vector, their speed, their, these sort of things can be used to enhance safety, comfort and convenience for the automobile sector. Now, what does this mean? This means a tremendous value for automotive providers. The automotive OEMs that manufacture the cars now can have greater software revenue in the systems that they sell to the consumer. The consumers get more interested in these capabilities that are really enabled through real-time systems being connected to the automobile. So the automobiles moving from being a standalone entity to a networked entity and that network is the 5G network. And so if you have deployed this with open RAN that allows you to deploy new monetization applications to the edge of your network, create more revenue from things such as drones or automobiles or private 5G and robotics, if you deploy with the legacy approach, you can't do any of those things. So we really see open RAN as kind of the gateway drug that gets you into a software-defined universe that enables you to generate significantly more revenue on your 5G network. Because of the connected devices such as we've talked about,

Clarence Reynolds, TelecomTV (15:44):
Given when RIVER is heavily involved in cloud deployments for telecom networks, what are you seeing as the trend in service providers? Are they bringing new Ask to the table as they consider upgrading or deploying new cloud infrastructure for ran?

Paul Miller, Wind River (16:00):
Yeah, they certainly are. And as we started this journey back six, seven years ago, looking at virtual RAN and eventually open ran, we had a kind of duality. The core of the network was deployed with legacy cloud technology, I guess we'd call it with OpenStack and virtual machines and that sort of thing. And then at the edge of the network it was all greenfield. They were deploying Kubernetes based systems for 5G. Now, as time has moved on, the core functions in the network have started to move towards containers as well, and that means that a transformation of the core cloud is happening at the same time. So what does that mean? We're seeing service providers come to us now and say, look, it's no longer just a decision about what I run at the edge. I need to transform my core. I need to transform my IT and enterprise workloads.

(16:46):
I think the industry is very well aware of what's happening with VMware and Broadcom and the opportunity that's creating for transformation of cloud and the IT enterprise end of the business. So everything from back office, telco core to Far Edge is now looking to transform. And that means we've changed from a point of use cloud technology for core and edge to a service provider saying, look, I'd rather adopt one technology all the way from the edge all the way to my core. That's a pretty significant change. And the interesting thing there is we've seen as a business solving the challenges at the far edge of the network is much more difficult than what's at the core. You're looking at real-time radio functions, hardware accelerators, low TCO pressures, low overhead, making sure your overhead software runs at the minimum number of cores. So if you've performed that optimization at the far edge with your technology such as Wind River has, as we move back to the core, that becomes ever easier to do and that's resulted, people are largely aware in the market of our win with Boost, establishing them as a new wonderful customer that we have great relationship with them.

(17:51):
That is a core to edge topology like I'm talking about now. Literally every single component in that network from core to edge is transitioning over to Wind River software and that gives us the ability to run everything from IT enterprise and back office all the way out to the 5G Open, ran on a single cloud technology and maintain all of that from a single pane of glass. Operational benefits of that obviously are huge. So this kind of transformation, the core to edge transformation and the desire to move every type of workload no matter where it is into a cloud, is something that we're really seeing and we're embracing with our offering.

Clarence Reynolds, TelecomTV (18:26):
And as Wind River, CTO, what do you see as the future of open ran and telco networks?

Paul Miller, Wind River (18:33):
Yeah, that's certainly an interesting thing. I think we've seen the core to edge transformation that we just spoke about. I think that's going to continue to propagate. We'll have a single technology being used from the core to edge networks, and that'll take the next five years or so. As that matures, we'll see the explosion of Edge ai. I think we're only at the very beginning of what AI is capable of providing for a service provider for network automation and cloud technology. We tend to look at that from an AI ran perspective, but as I mentioned in the core of the network in operations, in tooling, we've seen for example the ability to have an AI trained on the source code looking at log files and within seconds point to where the issue is in the source code and even suggest a fix for that. So it's really, AI is a transformation that's bringing power tools to everybody.

(19:25):
No matter what your job is, whether it's administration and requirements driven back office, you're operating and maintaining the network, you're writing software, AI is going to be a pervasive transformation, much like literally IP was back in the day. I think that's a pretty big impact to the future of all these networks as we talked about with the monetization angle. But really the transformation of the OT environment is connecting to the IT environment. We call this ot IT conversion and convergence. And what that really means is that all these connected devices and now we're going to become synonymous with connectivity and integrated with backend applications. I don't think we think anymore whether we buy a TV or car that it's not connected. And having that connectivity to backend cloud systems is something that's transforming every aspect of our daily lives. And seeing the automobile go through that, for example, and then become connected to 5G is transforming a lot of business models and considerations about whether we should use software technologies versus proprietary technologies. I think we're going to continue to see that shift for the next five to 10 years where everything moved more towards open source and software-defined configurable systems rather than proprietary systems. So a lot of things happening there, a lot of transformations over the next few years, and it's a very exciting landscape for the service provider.

Clarence Reynolds, TelecomTV (20:45):
Well, Paul, thank you so much for these valuable insights.

Paul Miller, Wind River (20:49):
Thanks. It's great to talk to you again, Clarence.