Ray Le Maistre, TelecomTV (00:00:10):
It is now time for our second session of the day, which is titled Leveraging Telco Cloud for Advanced Operations. Now the use of cloud platforms by telecom operators is reaching a new level of maturity with many now adopting a hybrid model. And while the potential and capabilities of public cloud platforms for telco workloads are frequently discussed, how have private telco cloud platforms evolved in recent years? What functions and applications are optimized on a private cloud and which workloads are better suited for public cloud? And how is telco software such as OSS and BSS evolving as a result? We're going to be discussing this in a panel that's going to be coming up shortly, but before we get to that panel, we're going to have a quick one-to-one chat on stage. And rather than look back at the past year for an industry-wide progress report, we thought we'd do something a little bit different for this fireside chat. Instead, we're going to highlight on an interesting service model that was really recently announced in Africa, and I'm pleased to say that joining me is Harker Singh, who is the CEO of Ascend Digital Solutions. So Harker, if you could join me on stage. Let's have a round of applause please for Har.

(00:01:36):
Great to see you. Thank you for joining us at DSP Leaders World Forum and we're going to be taking a quick look at a recent development in Ghana with relation to a planned, shared 5G network. So can you just tell us a little bit about Ascend Digital? What is the company's role in Africa's telecom and digital infrastructure sector just to set the scene?

Harkirit Singh, Ascend Digital Solutions (00:02:00):
Thanks, Ray. I think it's a pleasure to be here. As S and Digital Solutions, we are transforming the whole landscape of digital connectivity in Africa. Starting in Ghana, we are really focused on reducing digital divide and increasing financial inclusions, and over the last about four or five years, we have built a shared rural network providing connectivity to about 3.4 million subscribers. We have recently completed a program, we've plumed the whole country with fiber. We have upgraded the whole optical infrastructure to hundred gig backbone and also the data centers 200 gig simultaneous connectivity have been done and we are the partners to the government institutions. We provide connectivity and service and solutions as part of e-net and we have brought in a lot of digital infrastructure into the rural areas by extending the fiber infrastructure to post offices, police stations, schools, hospitals across the country. And that whole platform has enabled us to actually launch a new initiative that we call as a network as a service. So we are really focusing as a network and putting together a ecosystem and consortium of global technology leaders and MNOs to come together to build a 4G 5G network as a service.

Ray Le Maistre, TelecomTV (00:03:32):
That particular initiative is called NextGen in Fraco at the moment, O-N-G-I-C in Ghana. So can you just tell us about the specific role of Ascend in that and how that's come about?

Harkirit Singh, Ascend Digital Solutions (00:03:45):
So as Ascend, we are the largest provider of neutral digital infrastructure in Ghana. So we have fiber infrastructure, we have our own data centers, we have our own optical backbone infrastructure and a whole software defined core. It's the first what we call as a digital super information gateway that we have set up in whole of West Africa with a period of one year. So as a key consultant partner to NextGen in Franco, we are bringing all of the digital infrastructure and the key part of the component is the modeling around it at how do we work with the industry itself to build up the complementary infrastructure for MNOs so they can actually roll out the services to end customers. So as Ascend, we are providing that base and the platform of neutral infrastructure as a complementary infrastructure to the m and

Ray Le Maistre, TelecomTV (00:04:39):
S. Okay. In what way does this marker a shift in the way that 4G and then soon 5G services are going to be provisioned in Ghana and then in other markets as well? Because this is a single shared network and as I understand it, NGIC has the only 5G license in Ghana, is that right?

Harkirit Singh, Ascend Digital Solutions (00:05:02):
Yeah, that's right. I think to understand what we are doing, I'll just take you back five, six years ago where I visited about 39 countries in Africa and we started with a mission of about creating the biggest digital pipeline and the single digital market in Africa. That's where we started and we had some successes and mostly because of covid we had to stop now looking at these markets, the biggest, and you can say even in the GSMS study recently, the gap is a usage gap. How do you bring more and more usage? How do you bring more and more demand on this infrastructure so that there is enough, you can say applications and content, content goes into the markets. So these markets are usually very depressed in terms of how and what the user should be and how those content and services are coming together to build the right level of services to the end users.

(00:06:02):
So when we look at what we have done here is we are not looking at bringing the boxes and the software and the data center and the fiber and optical, I think that's been done in multiple markets and MNOs have the biggest specialization and the biggest CapEx to deploy those network. What we're changing is the paradigm. So if you take a Ghana as a market, you have two largest operators exiting the market. I'll say that's three. So Vodafone exited the market bought by TerraSol. AAL was sold back to the government, and GLO has also almost exited the market as well. And there is one large operator with a significant majority in the market and the market share. So the tier two, tier three operators are not able to bring that CapEx that allows them to compete. So what we are bringing is a model that brings that network as there along with the network component.

(00:06:59):
What we are bringing is the devices, the whole ecosystem of applications and platform, and this will be the first cloud native infrastructure which we are using. The core network from Microsoft Nokia is bringing up their RAN with the components of Open Ran. And I will say we have also been able to get attraction from new players in the telecom market like Redis, which is a geo platform company who are also bringing up their 5G RAN solutions into the market. With all the collaboration we are able to do bringing these large digital players with cloud native open ran technologies, we are also bringing the whole digital ecosystem that Geo did, which basically Alliance did in India. Okay. So we are using that whole model that was done in India to market, which is the same a RP levels as in Africa. We want to replicate starting in Ghana and go beyond. So the paradigm that's going to shift is bringing the devices, bringing the infrastructure, bringing the applications and the platform to drive just not the connectivity but the usage on this platform. And we have also learned from some of the other initiatives around the world on the Share network and we are based on those learnings, we are actually complimenting and getting the MNOs to be part of the venture as well.

Ray Le Maistre, TelecomTV (00:08:18):
Okay. Yeah, I mean it is quite a tough model to pull off. And you mentioned the company with the large market sharing gone and that's MTN, I mean, is there an expectation in time that they might join and use this shared network or is the expectation that MTN might get its own 5G license?

Harkirit Singh, Ascend Digital Solutions (00:08:38):
So I think that's depends upon how we actually move ahead and the government moves ahead and all the MNOs are part of it. So we have not as NGIC, we want to have the network shared among all the MNOs

Ray Le Maistre, TelecomTV (00:08:55):
To every service provider,

Harkirit Singh, Ascend Digital Solutions (00:08:56):
Every service provider, and the whole idea there is just not the service provider that exists in the country. We want to create an economy which is based on digital entrepreneurs. We really want to bring this open access neutral infrastructure where an entrepreneur can come up and say, I can set up a private network for servicing this industry. I can set up this private network for this industry. So we want to open up that whole ecosystem that has been subdued, subdued a long time. And as part of this whole stuff, we have also got commitments from our partners technology leaders partnership to build the very first technology innovation lab in Ghana that is only going to work on research and use cases for bringing up more and more shared networks in Africa and in the emerging markets. In addition, we are also setting up a software services factory to bring more and more jobs and create localized content and application for the engineers and the graduates. So what we are driving is a purpose built infrastructure that not only drives connectivity digital wide and financial inclusion, it brings to the level where people can start building the applications and IP locally in Ghana starting basically then going beyond in Africa.

Ray Le Maistre, TelecomTV (00:10:12):
And just very briefly to close out the hope, the aspiration is that once Ghana is up and running, that this model will be replicated in other markets. Is that

Harkirit Singh, Ascend Digital Solutions (00:10:23):
Right? Yeah, so the contours of what is there in Ghana are very similar into multimodal markets in Africa, largely because market size are anywhere from 10 to 15 million to 30 million, and there are a few large markets like Ethiopia and Nigeria, South Africa and working on the three or four operators, I don't think the markets can have that level of CapEx going into the markets and being able to compete. So we really see there is a space for this thing there with the right model, commercial modeling and the support from the technology partners bringing the device in application platform that allows the market to actually grow. So yes, so we have been in touch with few countries who are looking into this model in Ghana and I think we will see in end of this year when we launch. So the success of this will become a blueprint for taking it just not in African continent. We are also getting interest from other emerging markets in Forest Asia and also in Latin America.

Ray Le Maistre, TelecomTV (00:11:28):
Okay, interesting. Well, Harri, thanks so much for joining us today and telling us what's going on with Ascend and with these developments in Africa. Really incredibly interesting and keep us up to date as well with how it develops.

Harkirit Singh, Ascend Digital Solutions (00:11:42):
We will thank you

Ray Le Maistre, TelecomTV (00:11:43):
To meet. Thank you. A round of applause for Har please. Thank you very much. So it's time now for a bit more the shuffling of chairs as we invite our next panelists up to the stage. So while we do that process, here's a short video of highlights from our digital support systems summit.

VT-John Abraham (00:12:11):
The biggest benefit of moving to next gen BSS and OSS digital systems is agility. It's the ability to actually participate in opportunities that in the past completely bypassed us a telco. Not

VT-Beth Cohen (00:12:25):
All the telcos are jumping over to the public cloud. We still have substantial resources and will for the foreseeable future in our own data centers, in our own cloud. So multi-cloud and support from multiple platforms is the way to go.

VT-Lingli Deng (00:12:46):
Adopting cloud Native Paradigm does not mean that we have to also move to a public cloud deployment pattern for sales piece. I think especially in my opinion, they need to choose between private, public or hybrid cloud deployment patterns according to their own requirements and also taking into consideration about the local regulation requirements as well.

VT-Andy Tiller (00:13:19):
The migration of applications into cloud native environments is a good opportunity to rethink data architecture. The old approach of having master data management in applications and a big data lake for the data scientists to run BI and operational data stores for other purposes. I think this all needs a rethink

VT-Beth Cohen (00:13:42):
Customer experience is also extremely important. And making that translation between the messy tangle of legacy mishmash that we've been talking about underneath and then presenting it to a customer in a seamless way, there's a lot of work that needs to go on behind the scenes to make that happen.

Ray Le Maistre, TelecomTV (00:14:10):
Okay, fantastic. Now before we meet our co-host and our panelists as we dig into the topic of leveraging telco cloud for advanced operations, as you will know, each session has its own poll. One question, three answer choices. You can only pick one. So here is the question for this session, which should be the lead choice for hosting network workloads as you can see three choices. So go ahead and vote on the telecom TV website, you'll find all of the polls in the DSP leaders world forum section on the agenda page, and we will take a look at how the voting has developed at the end of this session. So I'd love to invite or welcome my co-host for this session. Greg McCall, head of chief Networks officer at the VT group. I think you've, you have been to this event before, but you haven't been a co-host

Greg McCall, BT Group (00:15:09):
Before. I haven't, right. So yeah, looking forward

Ray Le Maistre, TelecomTV (00:15:11):
To it. Great pleasure. Thanks for joining us and of course we have a great panel as well lined up. So let's meet them and let them introduce themselves as starting at the far end with Andrew.

Andrew Coward, IBM (00:15:22):
Hi, good morning. My name is Andrew Coward, I'm the general manager for software networking at IBM.

Mirko Voltolini, Colt Technology (00:15:28):
Hi, morning everyone. I run Innovation and Architectural Cult.

Geoff Hollingworth, Rakuten Symphony (00:15:34):
Hi, I am Jeff Hollingworth, I head of marketing at Anne Symphony

Michele Campriani, Optiva (00:15:39):
Ani. I'm the chief revenue officer of optiva.

Ray Le Maistre, TelecomTV (00:15:43):
Okay, thanks everybody. Well as with all the other sessions, we're going to start with an address from our co-host. So Greg, if you want to make your way to the lectern and give us your thoughts.

Greg McCall, BT Group (00:16:00):
Right. Thank you everyone and thank you Ray. Good morning. It's a real pleasure to see all of you here and for those of you that are online, welcome to I'd like to thank Ray and the whole telecom TV team for inviting me to co moderate today's session. And for those that don't know me, my name is Greg McCall and I'm the chief networks officer at BT group. We have some excellent panelists here today and we are going to discuss the topic of telco cloud for advanced operations. It's a topic that I'm really excited about, but before we get into the detail of that, I'm going to try and give you a little bit of an overview of how cloud plays its part in the network strategy at BT group. Our cloud native transformation is something that we've spoken about a lot over the last couple of years and it really is a key part to our network modernization efforts. But before we get into detail of that, I'd like to give you some of the background behind our transformation.

(00:17:04):
We are all witnessing a technological evolution that is really creating huge demand for connectivity. Firstly in providing a core service that enables these technology advancements to handle the increasing data loads that we are all seeing, but also ensuring that we can deliver fantastic customer experience. And secondly, in providing value added services that can help our customers achieve new levels of operational efficiency. The key to being able to do all of this is about building a fully programmable network. And that means it's a network that is not only programmable, but it's always on demand, real time AI first and supported by security capability and people that we've never seen before. So that is why we believe Cloudification is essential to the modern day telco at BT group. Our network cloud strategy, we've moved most of our network functions now into what we call our network cloud.

(00:18:14):
Doing this and leveraging the benefits of this really brings with it the key foundations behind our strategy. And while it's an ongoing journey, what we have seen is real benefits already. We multi-tenant today and that brings an increased level of efficiency, it brings an increased level of resilience and it enables us to scale our network functions independently from each other. As more and more of our network functions become cloud native, we are better positioned to deliver scalable on-demand connectivity. And also as we are doing this, it'll enable us to deliver new enterprise use cases. Additionally, we've recently finished the build and then the migration of all of our customers off our old Huawei core onto our new cloud enabled mobile core. This is the same core that will support our 5G standalone network, which we plan to launch later this year. But for us the mobile cloud core is much more than just a compute and cloud environment. It's a network that is built and been optimized for real-time communication and supporting traffic at real scale. The concept of continuous optimization is something that is ongoing and will be continuous. The other area that we are seeing a huge amount of value of Cloudification is supporting our network operations. Our network cloud enables us to deliver capabilities, services, network functions and lots of other capability across our network. In hours, these same activities and functions used to take us months if not years to deliver in the past.

(00:20:03):
A great example of how Cloud is supporting our automation AI efforts, which like all of you I'm sure is becoming even more important across the whole of BT is how we are using this to enable and upgrade and improve our network in real time much faster, in much more efficient and in much more reliable ways. We are also using it to monitor, detect and to fix issues before our customers even see this. And this is really critical as part of our operational resilience efforts. On top of all of this, we've also made a number of organizational changes and this is to help us to simplify our processes and we've brought much more closely integrated teams together from design development operations and we really started to see the benefits of having a full DevSecOps environment. And this is allowing us to focus on delivering smaller chunks of functionality to upgrade our network more frequently and ensuring that we do that in a highly quality way, but just as important as how Network Cloud is helping our operations. It's also helping us to think about how we drive product innovation and ultimately the combination of an AI enabled DevOps team and the cloud native 5G SA network. We believe this is the ingredients that we need to be able to offer the type of capabilities and services that our customers need both now and into the future.

(00:21:40):
But the final thing that we get out of this is we are able to create a platform that gives us real opportunity to expose our network. 5G era networks are massively software based, they're rich in capabilities and historically companies like ours, network operators used to keep their functionality to themselves. But through cloudification APIs we can now make that our network functions available to the developer communities and hopefully we'll be able to leverage their brilliant expertise to deliver new functions, new products, new services to our customers beyond core connectivity. And in turn we'll be able to make a fair return on that as well. So hopefully that gives you a little bit of an overview of how we are using cloud in BT to really drive our operations and drive our business forward. Thank you for listening. And with that I'll hand back to Ray.

Ray Le Maistre, TelecomTV (00:22:40):
Great, thank you Greg. Thanks very much for setting the scene and giving us the overview of where BT is now with its network cloud. I want to give an opportunity now for the panelists just to pick up and maybe ask any questions that they want to ask based on anything that you brought up in that opening speech.

Andrew Coward, IBM (00:23:06):
I have a question. It is fascinating the journey that you've taken with the 5G core and the move to standalone, but there's a lot more applications behind that, right? Would you even be able to guess? Is it thousands of applications that run across the telco cloud large across bt?

Greg McCall, BT Group (00:23:23):
It's hundreds rather than thousands and that is our destination for all new network functions. So that is our starting point and then we look at could we run it there or should we be running it elsewhere? But the majority of our new applications are running

Andrew Coward, IBM (00:23:38):
On online. Presumably you're on a journey to move each of those applications into this new environment essentially.

Greg McCall, BT Group (00:23:43):
Correct. We've made quite good progress the last 18 months.

Michele Campriani, Optiva (00:23:50):
Maybe if I could McKinney, of course there's the network but also the IT side and sometimes a mix and match, but how do you see also that side, especially applications like BSS and so on, how do they evolve and are part of your transformation?

Greg McCall, BT Group (00:24:08):
Yeah, so in BT we've organized ourselves where we've got a digital division and we've got a networks division and we've got multiple cloud environments. So we've got an enterprise cloud environment, we've got a network cloud environment. Our enterprise cloud environment is where a lot of our enterprise applications run and we also are running a lot of our enterprise applications in the public cloud as well. So we are multi-domain and run hybrid clouds. What we are going through right now is how do we evolve all of our cloud environments to be consistent so that it doesn't really matter what use case or what application people want to run, they can choose their destination and obviously the network cloud is closer to our network and therefore where there's latency driven applications that will be the destination. And then equally in the UK we've got the TSA, the telecoms security act where we are obligated to keep some of those functions on hardware that sits closer to the network.

Geoff Hollingworth, Rakuten Symphony (00:25:18):
Since this is question time for Greg, how do you define and measure success for moving to cloud native?

Greg McCall, BT Group (00:25:29):
It's a really good question and I think from a KPI point of view, I think the journey started with us really thinking about what is our long-term cost base, but also what are we trying to do in the future? And where we got to was if we continued in a vertical stack proprietary way, our cost base is going to get out of our control. So we took the decision four or five years ago to invest quite heavily in what we call the network cloud. And I think it's starting to pay off. Just the time it takes for us to deliver functions now is much, much shorter. We've all seen what's happened with VMware recently. I think that has turned out to be a really good decision of us moving off that as well. And then from a KPI point of view, it really is how are we enabling the business to deliver capability where we can make money rather than worrying about hardware, worrying about internal processes.

Ray Le Maistre, TelecomTV (00:26:39):
I just have a quick follow up question as well, and not just for you Greg, but for the panel as well. There's been a lot of telco cloud strategies over the years and right back at the beginning it started with we'll build all our own data centers and do everything ourselves. And things have changed obviously a lot, a lot since then. But do you feel, and I want to start with you Greg, because not so long ago you were talking about this to media analysts and you mentioned that other operators were now getting in touch with you to find out how BT got to where it is. Do you get a sense now that quite a few network operators have figured out what the optimum strategy, telco cloud strategy is for them and now they're looking to see who's done this model and that model and trying to pick up best practices?

Greg McCall, BT Group (00:27:36):
Yeah, I think it's probably fair to say Ray, we weren't the first to start it but one of the first to have completed it, which I think is a credit to the team. It's a credit to the vision we had and the blood, sweat and tears that people went through to get us there. And I think specifically to your question, we get a lot of people coming in and talking to us. A lot of people are asking us what learnings can they take from us to ensure that they can accelerate their migration onto a network cloud as well. And I always start with this's probably two or three things. The first is you've got to have a mindset change. You've got to really think about a lot of our team we're very used to taking stuff off vendors and then beating up vendors to deliver the capability or the functionality.

(00:28:31):
You've got to change your mindset and you are the system integrator. You've got to work with the whole ecosystem to bring it to life. I think the second one is being able to collaborate and influence and make sure that you are flexible in what you're doing and what you expect from others. And then I think thirdly, it really just does take a lot of effort and also rethink on the kind of skills that you've got within the organization to bring it to life. So we've had to hire different skills and this kind of merge of software and networking skills coming together is where we've seen a real benefit.

Ray Le Maistre, TelecomTV (00:29:11):
Now obviously as we know BT has been able to do this and we're going to get into the options shortly, options for different operators and maybe look at what some of the battle models are, but maybe we can move on now and maybe dig into what it actually means for a network operator to run its own cloud platform and what this means in terms of operations and infrastructure and how it's managed as part of a hybrid cloud approach. Maybe here I can come to Merko first and ask about the cult experience because obviously cult is a company that's been working with cloud partners and running its own cloud platform as well for quite a few years.

Mirko Voltolini, Colt Technology (00:30:00):
Yeah, yeah, that's true. So we're a fixed operator. We don't have a mobile core, so I think has been maybe more difficult for certain type of applications to be moved to the cloud and doesn't make sense for instance, for optical core or route switching core to have to move entirely to the cloud. But the control layer for those networks is best place to run in a telco cloud environment, whether it's public cloud or private cloud. We can debate on that. If I look at the operational side here, how do we run applications in the cloud? I can look at two scenarios. We have a multi-cloud, hybrid cloud environment, a number of applications that we buy off the shelf from vendors that think about CRM applications or inventory that recently we introduced a new partner, a run from a cloud. And I think the model there is evolved from on-prem to cloud and has got lots of advantages.

(00:31:02):
You don't have to take care anymore of the development of the application is done by the entirely, you get capabilities overnight without worrying about deployment of new releases. And maybe a challenge here is the move from a more CapEx based model to a subscription model. So as you shift in the way we spend money for that, then the other side of this alongside of the shelf application is what we develop inhouse. Yesterday were talking about cloud native applications. So all the control layer, the stack they were building to control the network, SDN, the orchestration layer is also something we moved to TECO cloud. And I think that is a different story in the sense of you don't have the benefit. I mean you need to keep developing those applications actually more intense in terms of effort than the previous model. And also you have to rethink about your entire stack. I don't see anymore in the future the strict separation of O-S-S-P-S-S and network more about the control layer that applies on top of the network being part of this open environment we're building.

Ray Le Maistre, TelecomTV (00:32:18):
Okay, I want to come to Jeff briefly here because Rakuten simply, you obviously have great insight into what's happening at Rakuten Mobile, built everything from scratch and basically everything runs on a private cloud, doesn't it?

Geoff Hollingworth, Rakuten Symphony (00:32:41):
I mean maybe I like the reason I asked Greg that question. It's always interesting when you'd have these technical discussions because maybe the fundamental question of just asking why are we doing cloud at all? 15 years ago we weren't doing cloud, now we are doing cloud and there tends to be always two answers to that for whatever it is. Either you're doing X in this case cloud because you want to be more efficient, you want to save money or you're doing X because it gives you capabilities to move faster, do more, have more agility, beat the competition. And I just looked Google before I came on stage just out of interest. So this year the telco cloud is going to spend, it's predicted 23 billion, there's going to be 23 billion spent on cloud in the industry to vendors, loads of vendors out there, really happy about that and they've set up organizations to celebrate that moving, but that means that we're already 23 billion down in the hole.

(00:33:50):
So I haven't seen a metric and the devil is in the details. That's why it's interesting to get inside various scenarios because when you aggregate at the highest level, you lose all specificity of actual progress. But in general, what's happened over that last 10 years is as an industry, our revenues have maintained or declined, but our operations have actually gone up. They haven't gone down. So we have to look at ourselves then and ask ourselves why are we doing cloud? And I liked the way that I summarized what Greg was saying, cloud isn't something you buy. Cloud is something you do. That's why it's so hard. We are an industry that enjoys buying things and what we rarely buy is not technology. We buy insurance and we pay a premium of that so that we can go and drink dry Martinez. And if something goes wrong is like I'm sorry you down down the hole, you go and fix that.

(00:34:54):
So having that intent or knowing why you're doing something I think is really important. So now if we go back to the story that I always tell about Rakuten entering telco, it didn't choose to do these things. They were cool and somebody told them to do it. They chose for two reasons. One, it's what they do everywhere else in their business, the most efficient way to execute a digital business. Everything's digitized or it starts with the processes. It doesn't start with the technology. And the second one, which was more important, they couldn't afford to do it a different way and they couldn't afford to do it a different way in terms of time and in terms of money because really the core thing about is that everything becomes programmable, which means that anything that you do twice, you write once in software and it repeats and it repeats and it scales.

(00:35:47):
And that's where you get this tremendous hyperscale operational effect where you can deliver exponentially more but your operations always remains linear or you can maintain your delivery but your operations exponentially decreased or ideally you flow off that way. But I think one of the interesting questions that always comes in my mind is if we can decide what those KPIs are, and I believe my biggest fear of an industry is this pivot a little bit to 5G SA because 5G has gone cloud native by birth, by design. If we're not careful, we're going to take that complexity and deploy it in a 4G static operating model, which means you've plugged it together and so you've got all the cost of complexity, but you haven't got any of that dynamic sharing or freedom or anything. And my last thing from 10 years ago when we first started doing cloud in telecom in a previous life we were working with an operator who we did a NFE packet call and they wanted to be first and they rolled it out and we got this really angry call and they said, well this is shit, it's just 30% less capacity.

(00:37:09):
And what they'd done is basically reassembled the hardware dedicated with a virtualization layer with a virtualized packet core on top. So all that was happening was 30% capacity as something fixed up there, spoke to something fixed down there and they didn't realize that. So they'd operationally recreated the past, which is what we do without thinking. That's why it's so hard actually. Again, there's not one size fits all and it's not a technology problem

Ray Le Maistre, TelecomTV (00:37:39):
Actually. Yeah, that was one of the learning curves from the early days of NFE, wasn't it? Just don't replicate what already exists. Andrew, I want come to you because if we're looking at how it's done and how success can be measured, you need to look at the different aspects of what comprises a telco cloud, right?

Andrew Coward, IBM (00:38:03):
That's right. That's right. And at IBM, as we take our telco customers through this journey, there are three really big building blocks if you like, that have to be consistent. The first is platform and the platform you're going to run everything on. You can't have multiple platforms. You need the absolute consistency of that to the point my earlier around VMware obviously Red Hat and IBM are taking advantage of that migration pretty strongly as you might imagine into Kubernetes right now. That's the first building block. The second building block is at modernization. It's not that long ago since telco apps were virtualized in the first place. And so now that next stage of putting them in containers, spitting them up, enable 'em to scale out is a pretty heavy lift in some cases. And then the third element is orchestration and automation of that. And when you do that and you do that properly, it means that you have this flexibility so that any application can be placed in to cloud any area. If you go one step further inside, I can allow that into the public clouds as well. It's a kind of normalization effect that you get when you do orchestration. And I'd say right now that's the weakest area that we identify in telcos. They've had multiple goes at it. I think there's still a lot of work to be done there.

Ray Le Maistre, TelecomTV (00:39:26):
So these are all good pointers and good trends for the operators with the scale and the resource to be able to do this and take advantage of this kind of platform. But that's not the case for everybody. Is it Mackay?

Michele Campriani, Optiva (00:39:41):
Yeah, we are a vendor, so that's why I was asking the question. We are BSS vendor. We've been there for 20 years and of course we participate to this transformation. Well first of all, we'd done the first thing, which is doing it our own. Six years ago we decided we were going to go cloud native and we did that for our own efficiency. So we did this but also to participate to the transformation, of course we see different levels of adoption. It depends also on what kind of network operator we're talking about. There's tier ones like bt, but there's also small operators that don't necessarily can afford to invest in their own telco cloud. So if they want to go as they want to go into the cloud and take advantage of that, they usually use the public cloud and at a little bit more audacious because BSS is a mission critical software and we know that operators not always feel comfortable with a public cloud, especially there's privacy and data issues with all of that. But in general what we see is that the smaller operators are driving that adoption of the public cloud and they're taking advantage of the agility that comes with it. And of course because you adopt the public cloud, you can actually focus on what is the BSS function, which is bringing agility on creating pricing plans and then go to market and then you get marketing teams that can do that without worrying about the actual infrastructure. And in terms of the actual evolution, we've been 20 years. So of course on-prem was the way it was done.

(00:41:43):
Nowadays what we see is that the 10% of our pipeline of opportunities are still on-prem, but that's going to die very soon. Those kind of unique cases, 65% is actually private cloud and 25% is public, but that 25% is growing and is growing fast. So what I foresee at least from a BSS perspective or IT overall environment is that public cloud will become the default choice. I think the smaller operators are driving that, especially because as I said, they can't afford their own telco cloud.

Ray Le Maistre, TelecomTV (00:42:28):
Yeah, absolutely.

Greg McCall, BT Group (00:42:30):
I mean I think you've got to look at it through three lenses, don't you? It's that control where do you need absolute control? And we are not going to put mission critical applications in the hands of someone else. We run the 9 9 9 service for the uk. We need to have real control there. The second bucket is quality. How do you drive quality? And you can drive quality both public and private, but what are your KPIs that you're looking at to drive that quality? And then the third is cost. If you can run a use case in the public cloud cheaper than you can run it in the private cloud and you've met the other two criteria, then that's what you do. However, not all use cases are cheaper to run in the public cloud than in a private cloud. So I think you've got to get those balances right.

Ray Le Maistre, TelecomTV (00:43:18):
And this brings us quite neatly onto what exactly would an operator do with a private telco cloud and where does it make sense? Is there a clear and obvious way to think about the types of workloads that would run on a private telco cloud? Are we talking here about key service orientated systems like a content delivery network platform? 5G core obviously has been mentioned and you mentioned that as well, Greg and customer data and you mentioned the Kaylee there about there are sovereign data issues here, but are there some real obvious examples? Maybe we can start with you Greg, and get a few examples of what you think runs best and well for the customers and makes the CFO happy as well in terms of being able to deliver profitable services to customers leveraging that network cloud platform.

Greg McCall, BT Group (00:44:25):
Yeah, so Ray, I think sometimes we get the weeds of the technology, which I think is what we're talking about. Our customers don't care whether we service their product through the public cloud, the private cloud or some appliance. They

Ray Le Maistre, TelecomTV (00:44:43):
Want champions

Greg McCall, BT Group (00:44:43):
League. Yeah, all they want is they want the best network, they want the best service and they want the best products. So I think when we look at it, it's not really thinking about too much more than how do we deliver an outstanding outcome for our customers? And sometimes we know that latency is really, really important. So we try to put those applications of those network functions closer to where the customer's going to be consuming them. So that is a good use case. You talk about CDNs, we in a number of discussions right now with CDN providers around could we leverage the compute that we've already deployed to host some of that capability because we know that that is the type of thing that is closer to the customer. But equally I think we've got to think about the network intent and what we're trying to achieve and then work out where best to do that rather than the other way around. Yeah,

Andrew Coward, IBM (00:45:43):
I think it's really interesting to the performance discussion. We've had this mantra in IBM of slow is the new down and it's really interesting when you think about that from a customer facing application perspective. If you have a three second delay in a response time, there's a 56% drop off in customers. So if that's your main webpage or your provisioning system, then that's a problem. Conversely, if your website is just 0.1 of a second faster, there's a 10% uplift in sales as a consequence. These are fascinating statistics that for our customers are just driving the obsession with performance. And so to that kind of where do you place the app, then that most often is the primary consideration for anything that touches a customer.

Ray Le Maistre, TelecomTV (00:46:36):
And there's also the consideration here as well for if you're delivering, and we heard earlier about the importance of low latency during the API session and if you're going to be able to deliver low latency services, then you're going to need a distributed cloud, aren't you? There's going to lead towards edge deployments, isn't

Andrew Coward, IBM (00:46:59):
It? Right? I mean another thing to think about, no application is an island and so you are only as good as the slowest link in that. So if that application has to call a backend database that takes two and a half seconds to respond, then that's why your application is now slow. So it's not just where the application lives, it's what else it has to talk to in order to function. And that complexity can be quite mind boggling. In fact when you think about how many different ways your systems are connected together. And so understanding that and managing that in real time and making sure that you are managing obsessing over the latency for every single one of those connections that app needs, that's what you need to focus on.

Ray Le Maistre, TelecomTV (00:47:44):
And Merko from a cult perspective, you've got this platform, you have customers, how does a company like Colt go about monetizing what it has in terms of its cloud assets and the kind of services it can deliver? Is it, well, I mean what kind of services are people looking for in 2024?

Mirko Voltolini, Colt Technology (00:48:10):
A couple of examples again has been fixed network operator. SD one has been massive success over the last few years. We run our SD one core and also our edge, some of our edge deployments for customers in our own private cloud. That is how technology has been built natively works very well also allows you to get in proximity of the customers. So often the customers, and I'm talking about locations here more than actually sitting in the public cloud, but running from data centers, you can deploy edge endpoints for SD one in proximity of where the customer actually wants to terminate the traffic. So it's also that location environment where you actually deploy in proximity of where the customer needs to be. So that's one example. The other example I'll bring up is network as a service. We built a control layer for our connectivity services, which runs again from our TechCo cloud, allows customer to consume services in on demand fashion, real time wherever we have capacity, which is many, many data centers and keep expanding that capability. There is a way to monetize and it also changes how do we charge customers the actual consumption model move to as a service, more usage base than you buy and pay all in one goal. So those are a couple of examples. Been significant successful with that model. So it's growing quite fast.

Ray Le Maistre, TelecomTV (00:49:49):
Jeff, did you want to come in there?

Geoff Hollingworth, Rakuten Symphony (00:49:52):
I was just contemplating, I was contemplating if we view what we're speaking about on this stage as a profession, so a cloud native professional and you're looking at the landscape of how to do that the best way then I think cloud native is not to do with infrastructure, it's to do with the applications and lifecycle deployment and management of it. So then not all applications are the same. We agree with that and the workloads are different and obviously then they have characteristics that must tell you where their placement is best. So some of the characteristics that have immediately jumped out to me in the past is if your application has complete predictability in load, it's always running at the same load. It's consistent and that's a good place to put it on private infrastructure, no point in renting a flat if you can buy it, if you're going to live there 365 days a year, no point in doing that if the application isn't changing very much, if it's highly static and it doesn't require any other services.

(00:50:59):
So there's no agility built in. Again, great for being kind of on that core infrastructure. A lot of the core applications in telecom hit those requirements and also for other reasons are very suited to private cloud. Now if you are experimenting, if you are trying to launch a new service and you have no idea, it can go from one to a million overnight. Public cloud is really good because you only pay for the opportunity and time, but make sure you can get it back if you need to because there's a great saying that we have that says failure is really cheap in public cloud, but success is really expensive so you pay a lot forever. And then the last thing that just to add almost a BP oil spill in all of this, and we haven't mentioned it once on this part, I thought

Ray Le Maistre, TelecomTV (00:51:56):
Oil firms are available,

Geoff Hollingworth, Rakuten Symphony (00:51:59):
IT AI and data is a game changer. So data gravity suddenly we maybe come onto to this time.

Ray Le Maistre, TelecomTV (00:52:06):
We are going to come onto this in a moment, but I want to come to the audience to see if we've got any questions, but I want to come to McKayley first as well just to maybe comment on what Jeff just said there about a cost because we hear lots of different anecdotes

Michele Campriani, Optiva (00:52:28):
About the cost. Yeah, if you look at the BSS, it has that mission critical aspect of it, so I understand what Greg is saying, but on the other side it needs the agility that we were talking about and you want to try out new plans and usually an operator takes what, six months to come up with new plans because it is a problem. You have to ask them the ability to do that and the billing team and so on. Now if you want to become agile and launch new pricing plans and meet new promotions and then maybe you decide that actually it didn't work out, you want to take it out and you want to try something else, well that's where it's all happening in the BSS area. So BSS should be agile and should be capable of supporting that and the cloud is definitely the right environment for that. And of course us as providers have to be cloud native in order to support all of that. Yeah.

Ray Le Maistre, TelecomTV (00:53:36):
Okay. Now I think we do have a question one at the back here and another one at the front two. Put your hands up and let our colleagues know. So we'll start at the back there. If you can state your name and your company

Audience-Francis Haysom (00:53:53):
Francis has Apple Research. Greg, good comment you made about your team needing to become more aware that they are a system integrator. I wondered, and it's across all the panel, one of the challenges that I'm a former vendor and you find that a lot of relationships with a telco are a bit of a zero sum game in terms of offloading of risk, offloading the manufacturer. And I just wondered as you move more to a cloud, a sort of cloud approach to this, a more system integration approach, are you finding not just that you're acting as a system integrator, but you're finding that the relationships, the ecosystem that you're having is changing in terms of the commercial environment? Is it driving a change and I wonder what the others are also seeing in terms of that rather than just being a kind of it's all your fault.

Greg McCall, BT Group (00:54:52):
Yeah, I think it's a really good question and what we saw is a reluctance to change initially both internally and with our vendor community expecting us to build this kind of cloud environment and then we can just put applications on top of it and they would just all work and it'll all be fine just to bring it to life. The model that we ended up with is we had an ongoing, we don't call it a war room, it's not really a war, but a war room

(00:55:26):
Going 24 7 that basically it was an ongoing teams channel that basically all the vendors and everyone within the ecosystem could ask any question at any time of the day. And that's how we started to make real progress. People started to talk to each other and what you've got to recognize is the majority of people, regardless of what company they work for, they want to do a good professional job and you've got to create the environment where you can do that. Now that does lead you onto the kind of commercials, but fundamentally what we got to was we all have to be successful. The enemy isn't between us and the vendor. The enemy isn't between two vendors working together. Fundamentally, it's an industry as we said earlier where we are not awash with a massive growth other than on data which people don't pay for. So what we have to do is find better and more cost effective ways of the whole ecosystem working together. And I think we've broken through that to be honest. I think we've got great relationships with all of our vendors in the community and the relationships we've got is people are looking out for each other to make each other successful rather than to prove that they're not the ones that have caused the problem, which is where we were probably four or five years ago

Geoff Hollingworth, Rakuten Symphony (00:56:44):
And I'd just like to complete a thousand percent, that was the Rakuten experience, getting people into a room and everyone said we're going to make this work and the goals that they were set were traditionally insane. The moment people came together and it felt it was weird because you didn't have management of risk on paper, it felt riskier to some people, but it's not the risk kind of disappeared and then people felt that and then you get this flywheel. It is a real interesting human almost like experiment of how people forget kind of the problems and they look for solutions suddenly.

Ray Le Maistre, TelecomTV (00:57:25):
Okay. We have a question down here. If you can state your name and your company.

Audience-Alfredo Musitani (00:57:30):
I'm Fred Ani from Tele Argentina, and just one question that came across is about the dependency on the cloud software in particular that you're using. In particular, what is happening in all the other industries? I don't think it's the same in the telco industry, but about VMware problem and it is going across every industry and I think what happened is that problem happened with AWS software as a service, not the cloud provider or maybe other one. And if you rely too much in one or the other. And what is the impact in VMware specifically because it's a real problem VMware right now. That's a question to everybody and other here.

Andrew Coward, IBM (00:58:29):
Yeah, if I may, we'd see across different industries. Exactly that point of it is becoming a legal requirement to be in multiple clouds. The banking world is going through that right now. Every industry is thinking about resilience and the challenge with that comes to complexity because if I have to run an application in AWS and Azure or IBM Cloud and Oracle, then I now have this challenge of how do I normalize that and make it simple and which users go to which cloud because I can't just have one in backup. I have to run them both simultaneously and share. And so there's a journey that most of our customers are on what I would call cloud normalization, which is how do I take an application and run it on private cloud, run it on edge if I want to run it in AWS, run it in Azure and normalize that. You can imagine that the hyperscalers are not too keen on that idea because it commoditizes their cloud. But the reality is regulation is forcing the move to normalization and as a result there's a whole host of tools, some of which IBM provide that are enabling that normalization work to take place and allow the migration of applications in and between all these clouds.

Ray Le Maistre, TelecomTV (00:59:47):
Okay.

Mirko Voltolini, Colt Technology (00:59:48):
I would add to that, I mean in general if you look at the specific of VMware, it's kind of like the same across all technologies. As a large operator, you have to have a dual vendor approach as a minimum. So you have also multi-cloud approach. I think I a hundred percent agree what Andrew said about portability of applications. We're not there yet. In the short term, most of us run dual stacks, so you have to be able to have a backup plan when something happens with one of your vendors.

Greg McCall, BT Group (01:00:23):
The final word I'd say on this is it is about having optionality and I think that's what a harbor cloud approach gives you is optionality and VMware remain a strategic partner of ours, but there are some use cases that it's better for us not to run on VMware to run on either private or public cloud.

Ray Le Maistre, TelecomTV (01:00:49):
We are getting towards the end of the session and I do want to make sure that we address what Jeff very neatly called the oil spill situation. And this is to focused on what kind of impact has the emergence of generative AI had on telco cloud strategies because this is the kind of scenario where the traffic volumes could get massively out of hand and have lots of other implications as well. There's lots of talk about the energy implications as well and the chip vendors are all scrambling now to try and mitigate this. But we'll maybe start with you Jeff, as you already mentioned it in the past year, what's happened? How can it be dealt with and what have we learned?

Geoff Hollingworth, Rakuten Symphony (01:01:51):
I think I'm actually friends with the chief privacy guy at Uber. If we want to know how to design this, it's a good idea to go and look at the massive scale internet companies and just watch their journey. And what I think we'll discover is that they've built a network to support the operations of their network. So there's a separate data network that just is there dedicated to collect and distribute telemetry data across the whole operation. And the volumes of that data are mind boggling. Now, we always claim in telecom that we have this data, but really it's exhaust fumes. It is not machined, industrialized, prepared or managed as a complete system and that's what you need to get to. So we were fortunate in Rakuten because we built everything up from a distributed data center point of view. So it already has that kind of flow of data gravity to the data centers where the functions are and we collect it all back.

(01:03:05):
The second part though is that we do use public data sets for the automation services that we actually do and some of the modeling that we do. So we actually ingest that into almost like a private cache that we run our models on in the real time. And if it gets blown up, that's fine. We'll repopulate the cache over time, but the design of that data network is not something that you again leave by chance. It's a system driven operation and it really drives then from that point of view, it drives the center of gravity. If you want to start doing self-driving networks, then that data has to be in the network. It has to be where the network is because you haven't got the latency or the time to actually bring it to different places. So it's a discipline again, in its own right. It is fascinating and when you work at that scale, I mean that's, that's how we would get 20 something joining our industry to be honest, be so turned on by managing that amount of data at that scale and doing things that are that important. If we started marketing it telecom a bit that way, I think we'd get a different recruitment company.

Ray Le Maistre, TelecomTV (01:04:29):
Greg, what kind of in the past year, I'm sure there's been asked probably lots of times, how are you using AI to improve your processes, but at the same time you are probably seeing some interesting traffic statistics movements as well and how's that playing out and how does having the cloud platform you have enable you to do things better? Or does it not make any difference at all?

Greg McCall, BT Group (01:05:01):
Well, I think Jeff makes a great point. I think as telcos we've undervalued the brilliance and the vast amounts of data we've got and we haven't used it either to monetize that data nor to improve our operations well enough. So I would start answering your question, but our network cloud is quite a vast estate. We've got seven control plane sites, 10 user plane sites, and two tool sites over 30,000 BGP entries. So you can't manage that unless you're collecting all the data and you've got all that data in real time and you've got that data where it's timestamped and you're not working off old data. So I think from a mindset point of view, we've had to realize that data is the key to running this complex network and that has driven a different set of people have come in and they are then challenging us to say, you've got all of this data, why don't we use it to do different things?

(01:06:10):
And I think within our business, and I'm sure many other people, AI has been a thing for some time machine learning automation. But what we are now starting to do is challenge ourselves on the standard things we used to do. Why do we have to do them in that way and how can we use AI to drive it better? And I always think about gen AI as the kind of little angel on the shoulder of your first line operators where they can ask any question and they can get an immediate answer and be upskilled in real time. And I think that is really the killer use case that we can start driving.

Ray Le Maistre, TelecomTV (01:06:47):
Okay. And McKay, I want to ask as well, from your perspective, and this is just from a sort of a capability point of view, has optiva been able to take some of the potential that's come from the upsurge in Gen AI and be able to more quickly perhaps integrate that and make use of it?

Michele Campriani, Optiva (01:07:13):
Yeah, absolutely. I mean, again, the BSS environment is very conducive for the use of gene AI in the sense that we have the data that you are talking about and the use of gene AI can do several things. One is again, helping out operators to understand what kind of products and pricing you can offer and specifically even personalize it for customers so that you can help out in understanding what could be the best offers and specialize on specific segments. And Gene AI helps you right away in doing that or it can help you in your operations. Again, in looking at how efficiently you can build these plans and come up with new ones and just even conversationally interact, it goes back to the marketing team that is using is the end user of this function. And again, maybe being able to ask what would be the best pricing plan and can you configure that for me and let's launch it tomorrow, which is a dream right now, but it could actually happen if you have a gene AI environment to support you. Okay.

Ray Le Maistre, TelecomTV (01:08:44):
And yeah, Andrew, and then we'll finish with Merck.

Andrew Coward, IBM (01:08:50):
I was going to say, I know we've been playing buzzword bingo for the last day and a half. The one we haven't actually used, if you familiar with your cards, is edge compute, which starts to come into the fore with AI as a mechanism of aggregating, collecting, responding. In this AI world, and I'll give you an example, at IBM, we work a lot with what we call quick serve restaurants, which as fast food restaurants and in the quick serve business, AI is a very important element because it's taking orders. So as you drive up or walk into a store, you can now speak your order in what might be three or four different languages and that will get interpreted and you don't want to have that compute if you put it in the cloud, it's too far away. So you need this low latency response. So there's an edge compute opportunity that's starting to play out in this edge. Compute will also be important for the aggregation of the vast amount of telemetry that we're talking about because you can't push all that upstream. And so you'll need in the same way that cdn, so the number one use case for edge compute right now, the inverse of collecting and aggregating and shrinking and normalizing that information will be the other key use case that will play out as a consequence of ai.

Ray Le Maistre, TelecomTV (01:10:04):
Okay, and MoCo finally has the past year, have you as COL had to kind of rejig its strategy to deal with the demands for these new and emerging applications?

Mirko Voltolini, Colt Technology (01:10:17):
Yes. So I second again what Tan just said here, we're looking at the opportunity of running inference at the edge and I would expand your questions to include machine learning. It's not just about gene ai, but it's the whole set of data that you can leverage to build smart solutions. End to yesterday we were talking about the fact that telecom is just a piece of the overall puzzle. So delivering outcome to customers. Think about combining connectivity with AI at the edge, maybe iot, which is a typical use case for the industrial environment. Creating that type of smart industry type of applications is one of the opportunity. Then maybe going back to operations, I think I would like to offer another perspective is about how do we manage the data. What we're seeing happening and is a challenge is that we are collecting now data in multiple different places.

(01:11:08):
I think that the risk that we have is that we move from an environment we had on-prem where we databases scatter over the place with data, not in a good place to a model where we have multiple clouds with data all over the place again, and I think that problem hasn't been solved yet. We will like to avoid duplication of data. Imagine we collecting data from our CM environment, from our inventory environment and maybe sit on different clouds plus our TECO cloud. Think about how to bring this together is one of the challenges we have

Ray Le Maistre, TelecomTV (01:11:41):
And actually we're going to get a little bit into that topic this afternoon in the automation session there as well. But for now we are out of time for this particular session, so we do need to end the discussion there. But before we end, we're going to take a look at the poll results we got so far. Just a reminder, the question we asked was, which should be the lead choice for hosting network workloads? And here is how the voting has gone so far. Perhaps predictably, hybrid cloud has come out on top, but private cloud 29%, I dunno, that's a bit higher than I thought I have to say. And public cloud obviously still quite application specific it would seem at the moment. So thanks very much for voting there. Just a reminder as well that these polls aren't closed, you can continue to vote on the telecom TV website. So that's the end of our sixth session. We're going to be back this afternoon at 2:00 PM UK time, 1400 hours for session seven, which is improving network optimization through automation. In the meantime, for those of us here in Windsor, we have a nice lunch break, which is going to be served in the foyer. And don't forget the pinball and the all important charity tournament. Remember, this is your last chance. The lunchtime break is your last chance to get your high scores into Alex. And

Geoff Hollingworth, Rakuten Symphony (01:13:15):
Anyone beat a Neil yet,

Ray Le Maistre, TelecomTV (01:13:16):
Ray? Yeah. Wow,

Geoff Hollingworth, Rakuten Symphony (01:13:20):
This

Ray Le Maistre, TelecomTV (01:13:21):
Cup awaits somebody in this room. I mean, who wouldn't want this? You can touch it. Greg,

Geoff Hollingworth, Rakuten Symphony (01:13:29):
Just before we finish, can I make a public announcement? So SpaceX is launching Starship in about three quarters of an hour broadcast live. So while we've been solving cloud, like Elon's launching a rocket, see if he can get it actually back to earth. So if you want to watch that, that's on their website, about 20 past one.

Ray Le Maistre, TelecomTV (01:13:49):
Okay. You can be having your lunch and watching the next generation of internet Connect

Geoff Hollingworth, Rakuten Symphony (01:13:54):
People leaving us

Ray Le Maistre, TelecomTV (01:13:56):
Now for our online audience. Don't go away. This isn't the end for you. As we're about to start our next extra shot program, Sharla is going to be discussing the issues around Telco cloud with her guests. So don't miss that. Stay with us. But for now, let's thank Greg as our co-host and our panelists for this session. Thank.