The Green Network Summit 2025 - Q&A show day one
By Guy Daniels
Feb 4, 2025
To embed our video on your website copy and paste the code below:
<iframe src="https://www.youtube.com/embed/B3NKnWz4dP0?modestbranding=1&rel=0" width="970" height="546" frameborder="0" scrolling="auto" allowfullscreen></iframe>
Guy Daniels, TelecomTV (00:24):
Hello, you are watching the Green Network Summit, part of our year round DSP Leaders coverage, and it's time now for our live Q and A show. I'm Guy Daniels, and this is the first of two q and a shows. We have another one at the same time. Tomorrow it's your chance to ask questions on energy efficiency and innovation within the network. Now, as part of today's summit, we featured a panel discussion that looked at how to improve network energy efficiency with ai. And if you miss the panel, which was hosted by my colleague ra, don't worry because we will rebroadcast it straight after this live q and a program, or you can watch it anytime you like on demand. Now, we have already received several questions from you, but if you haven't yet sent one in, then please do so now and use the q and a form that's on the website.
(01:21):
Well, I'm delighted to say that joining me live on the program today are Mirko Voltolini, who is VP of Innovation at Colt Technology, Diego Lopez, Senior Technology Expert at Telefonica, and an ETSI fellow, Neil McRae, Chief Network Strategist for Juniper Networks, and Beth Cohen, Telco Industry Analyst with Luth Computer Specialists. Hello everyone. Good to see you all again. Thanks so much for joining us. Let's get straight on to our first audience question. We've got a mix of questions today. Some are AI related, some are more general energy related. So the first question we've got from our audience is given the dramatic rate of advances in ai, what innovations can be applied now to the network to improve energy efficiency and which will not be outdated in a few months time? When should telcos commit? Diego, how about coming across to you and getting your views on this one first for us?
Diego Lopez, Telefónica & ETSI (02:31):
Well, I would say to start with that the pace of advancement, et cetera, probably is the same that we had some time ago or we still have coming with radio interfaces, with optical networks, with processors, with memories with software itself. I don't think that the pace is going to be that much different and that the solution is going to be the same is about using open interfaces, using the possibility of plugging and unplugging the different components. And for this, what we need is to have clearly defined the interfaces of what we use and a way of describing the different elements that we are going to use. The same way that they use other pieces here is probably those pieces have a different nature. Probably the kind of descriptions that we will need would be less physical, more virtual, more oriented to data rather than to that, for example, connectors or tensions or wavelengths or whatever.
(03:44):
But the idea is, as we have done in the past when we're trying to engineering and applying new technologies, is about making the whole thing interchangeable and interoperable. Basically this tradition has been done by formal standards and that has been created by a process of a consensus, et cetera. Right now we have learned to apply methods that converge in a faster way like the use of open source elements or as reference mechanisms, but I do believe that at least if we don't have the ultimate tools for making this happen, we do have the mindset and the methodologies for doing this. And so I don't see any risk in start committing right here right now to whatever the technology can offer.
Guy Daniels, TelecomTV (04:45):
Great. Thanks Diego. So don't wait. Commit now and we'll get some more views from our other guests. Neil, how about I come across to you next? Because AI has got huge media coverage at the moment and we're seeing advances every day. So is it a case of if you commit now you might find yourself with some outdated applications?
Neil McRae, Juniper Networks (05:06):
Well, I think there is a risk of that, but that's the risk of technology. I do think service providers need to be out there experimenting and learning so that they can make the right calls on what they should invest in, what they shouldn't invest in, and also understand where AI is going to make the biggest amount of difference for 'em. And I've said this before, but if you're very good at what you do, probably AI won't have as big an impact as you think, but if you've got a whole, and I've seen this, telcos in the space of inventory of their infrastructure, lots of things that aren't right in the face of customer service tends to take a backseat. And I think that's an area where AI can help. And one of those areas that historically did take a backseat was energy. Thankfully that's front and foremost for just about every service provider, and that's probably one area where AI can do it, but I think if you're sitting in the corner saying, well, we'll wait a few years for ai, there is some merits to that, but I wouldn't be ignoring it entirely.
(06:14):
I think that would be slightly crazy, but I'd also be encouraging your teams. I used to try and find with my teams, I used to try and find the, we'd have 10% of our time, what's out there? What can we build with something that we might not take the normal process for? So definitely experiment, learn, and then when you think you've got something that's got the right impact to customers, the right impact to TCO and generates value, then execute on it, but have a mind to the future because as you say guy, it's not the media is saying, I think the media is behind on how fast AI is actually moving. I think it's moving faster than any of us really understand or realize.
Guy Daniels, TelecomTV (07:02):
Yeah, I'd agree with you that one, it is hurtling along a rapid pace of nots. Beth, lemme come across to you. I know you want to come in on this one as well. I mean, is this question about when to commit to technology to solutions really focus on ai or is it just a general concern no matter what the technology, what the approach is?
Beth Cohen, Luth Computer Specialists (07:20):
Yeah, my opinion is, I mean, telcos are by their very nature, they're engineering companies and yeah, shiny objects are great, but they also need to deliver a service that works. I used to say when you pick up the phone, you want that, you expect that call to go through 100% of the time, and that still applies. You expect your network to be up 100% of the time. So telcos can't afford to just kind of throw new technology in the mix without testing it and making sure that it will actually work. We cannot afford to have a massive outage. So ai, absolutely, the telcos are and should be working with ai, looking at where it makes the most sense. There's been great strides in using the more proven AI solutions such as chatbots and operational, using it to support operational efficiencies. Its support within what I call network intelligence.
(08:41):
I would still say that is definitely a work in progress and there's a lot of things that need to be done before it can be deployed out fully in the field in production. However, that should not stop any of the telcos from digging in and seeing what's there and more importantly, contributing. I know that Diego mentioned open source, there's a number of open source projects in the telco space looking at seeing how AI can be used within the telco space. And there's a number of hurdles that need to be gotten over just to make it more usable and contribute to the algorithms that need to be developed.
Guy Daniels, TelecomTV (09:34):
Great. Thanks very much, Beth. So clear message there is experiment, get involved, don't wait around. Excuse me. Another question from our audience here. I like this one. We hear about the need to break the energy curve to prevent energy rising in line with data traffic growth, that correlation that's been going on for years and years. Surely the only way to do this effectively is by replacing old infrastructure with new and ensuring that new standards focus on efficiency, which is good news for vendors. But can telcos actually afford this? I like the question, Mirko, let's come across to you. Any thoughts on this one? Is that the right approach and can telcos afford to do it?
Mirko Voltolini, Colt Technology (10:27):
Well, the question probably should be more like can telco not afford to do it? I think what we have seen is, I mean definitely is a question of return of investment, but technology that is coming out now is 10 times more efficient, 10 times faster, 10 times cheaper, and also utilize much less pace. So the business case comparing like for all the technology, newer technology is going to be in favor of deployment of new technology. So the operators will do that anyway. I think the key aspect here is also making sure that you actually remove older technology because this is what probably Telco haven't been the good over the years. We keep piling up older technology, new technology and all the technologies isn't going to help you. So I think you need to make a concerted effort to remove legacy network and migrate over time. I think programs can be very efficient, even though they are demanding in terms of resources and time being spent on it. But I think it's definitely an aspect, an area where money can be made and can be saved. So we're actually investing in this. We have an old migration program where we keep deploying new technology, move out. Also all the locations reduce space and reduce power and increase capacity. So there is definitely a case for this.
Guy Daniels, TelecomTV (12:01):
Thank you Mirko. I like that. Can Turco has not afford to do this? Take this approach. Beth, I'm going to come across to you because there was a related question about this energy curve and this relationship as well, and another viewer talking and comparing the telcos with the data center sector and looking at how the data center energy usage is ramping up as a data center. Rollouts are increasing and there's a lot of concern about what these values are and what does it all mean? What's your take? What's take on trying to break this relationship and what are your thoughts?
Beth Cohen, Luth Computer Specialists (12:42):
So this is a very complex question. So first, the telcos are absolutely interested in reducing energy footprint and they have been quite actively pursuing it in certainly around the globe. However, in the us, the US had got their first, if you will. So there's a lot of legacy equipment out there that obviously needs to be decommissioned still. I mean literally in my town they pulled the copper out about a year ago and making a number of elderly people very mad. But the point is, is that the fact that we still have copper infrastructure out there is an indication that it's time to upgrade and this equipment is long past its end of life. And now let me talk about the telco versus the data centers. So telcos, it's a mature industry. The network traffic is growing, but it's not growing at the same pace as the data center industry, which is growing at what, 45% a year or whatever.
(14:07):
So it's a little bit of a red herring to say, oh yeah, the data centers are using far more energy than telcos or the rise in energy uses far more than the telcos because the industry is newer. In fact, the data center industry has been thinking about the costs of the energy since the beginning pretty much because they're concentrated. So they see the giant electric bills in a different way than the telcos do because the telcos, the energy is spread over a wide area network with hundreds of thousands or millions of nodes. So it's a very complex issue, but I think both industries are working on it, but the answers are slightly different.
Guy Daniels, TelecomTV (15:08):
Yeah. Thanks so much Beth. I think it was good to point out there that these industries are working on this. They're not ignoring it. No evidence of that at all. Neil, did you want to come in on this question about breaking the energy curve and looking at whether or not the approach is to replace existing less efficient kit with new more efficient kit complying to the latest standards? That's the only really long-term sustainable way of fixing this.
Neil McRae, Juniper Networks (15:35):
I don't think it's the only way, but I think it's definitely a way to think about it. I mean, you walk around any central office, you probably find PDH equipment, SDH equipment, first gen optical equipment, an average SDH rings 155 megabits and it probably costs the same amount and power that a 400 gig router is so for sure turn off the old stuff, migrate customers to newer technologies. I think telcos have done an okay job of that. I think they need to do a better job. I think Beth's a hundred percent right on copper switch off in the world. We are so far behind on copper switch off. We need to go way more aggressive at that, not just from an energy point of view, but from a digital enablement point of view. But lemme just mention one thing though. I think there's a nervousness that everything in the network or the data centers is bad when actually data centers have made a dramatic contribution to reducing energy and carbon by enabling people to appear on events like this rather than having to get in a car or fly to a place or travel.
(16:55):
We can all telecommunicate or we can telework or we can do loads of things that in the past just were not possible. Even during covid, we saw one of the vehicle manufacturing teams using edge compute with digital twin to do design and to test stuff in M 3D printing prototypes in their office. So I think when you look at data centers in the network and you say, wow, they use a lot of energy, but don't forget how much value they bring us, which is substantial, but also think about user behavior. I mean, does anyone think about how much it costs in energy when they download that funny YouTube video? Probably not. But it has a cost and I think as sensible users, we may want to think about that a little bit more, although let's not stop watching funny cat videos because they're the highlight of my day.
(17:57):
But there's a sensible approach to this where if we're smart and if we look at things and I've not met a telco that doesn't know what it needs to do there it is just how do you make it happen? And I think that's where AI can potentially help because it can crunch a load of data and help service providers make decisions that are typically really, really difficult. In my past, we were working on copper switch off and it was one of the hardest bits of work to do because you uncover so many things. So for example, how do you do telephony in an elevator if you don't have copper? And that telephony is the emergency for if something happens in the elevator, how do you do looking after those that need help at home if you don't have copper? Because all the things that are out there use copper networks today.
(18:54):
So we've obviously got to do this sensibly, but one of the things I hear and I talk to customers about is network modernization. It's bringing the latest in silicon, bringing the latest in optics, bringing the latest in power management, bringing the latest in security, but also enabling new services. You put those five into a business case, I challenge anyone to see that there isn't a TCO there that is valuable for them, but most importantly brings a capability to their customers that's valuable for them and generates revenue. So don't think of this as just a one directional street that it's all about taking out costs, which is super important, but it's also about driving new capabilities for customers, which is also important and enables your customers to reduce their power and their energy consumption.
Guy Daniels, TelecomTV (19:44):
Yeah, that's the balance. Thank you very much Neil for those comments. Much appreciated. I think we'll move on to another question then. So we've got a few questions in from our viewers already, right? This one's about digital twins as we've just briefly heard. So the question is AI can be used to create digital twins that allow telcos to run virtual simulations of their real life networks enabling experimentation to help improve energy efficiency. But is this just an academic exercise or can it realistically be used to optimize physical networks today? What's the best approach for telcos to take? Well, digital twins have been around for a long time. Beth, do you want to have a crack at this one first and give us your views on the use of digital twins here?
Beth Cohen, Luth Computer Specialists (20:43):
Certainly. So yes, digital twins have around for many, many years I had a lab at Renick and Newman. We were the inventors of the internet. We did several digital twins for the network and that goes back 30 years. So they're not new and they're clearly of great value to the telco industry. So it is definitely not an academic exercise. In fact, I would argue that telcos use them very heavily for all sorts of things, not just energy efficiency. I think that's something relatively new, but certainly for network optimization calculations, for determining best path, all sorts of all sorts of good things. And so adding energy efficiency into those test harnesses is only just goodness.
(21:52):
Digital twins are very valuable to the telco community. I would like to see, most telcos do have them. I would say probably all of them do. I would like to see some more digital twins in the open source community. I know LFN does have a number of platforms that they make available to the community that can be used in this manner. And I think that would help further some of the more standards, some of the more areas where we do need standards and we do need open source to optimize not just the performance of the network, but also the physical and energy efficiency of the networks as well.
Guy Daniels, TelecomTV (22:45):
Great. Thanks very much Beth. We'll come across to Diego in a moment, but I just want to go first to Neil because Neil AI likes to gobble up a lot of existing technologies and claim them for itself. We know this seems to be a case of one of them, but what about this application of digital twins to help, especially with energy efficiency of physical networks?
Neil McRae, Juniper Networks (23:07):
Yeah, look, so data center builders have been using digital twins to design data centers for probably for, I'm going to say more than five, maybe even as much as 10 years. I remember being part of a data center build for a customer where they were plotting the best way of using the space, the easiest way of getting power, making sure that the air conditioning and the flow of AR was going to deliver the yield that they wanted in that data center. So I don't think this is anything super new and I agree super with be on. We've been using digital twins in the networking space for some time. I think the difference AI is going to bring is an ability to run simulations probably much more quicker than we were able to in the past. So I would always get a question, well Neil, when we run the test, why didn't we discover this issue?
(24:07):
Well, it's just because we had so many limited time to run different test criteria. I think ai, I hope, will give us the ability to run more and more test criteria. But I want to point out a piece of work that my colleague Pel did who works in our one group here at Juniper at nano last year. He did a really great presentation on network digital twins. I'd encourage everybody to take a look at, we think this is a super important thing. We're building it into many of our tools for service providers to allow them to do more and more what ifs and why are we doing that? Well, one, to optimize the network to modernize the network, but two, network reliability has just become such a table stake item. You cannot have outages, you cannot take stuff down, which often slows service providers down because they're frightful of the risk of making changes.
(25:04):
Whereas actually what we believe today is by not making changes, you put yourself at greater risk, not updating code so that it doesn't have any bugs or security issues. Getting to the latest platforms that are able to scale more. So enabling service providers to do more by seeing it simulated in front of 'em is definitely a way that we're moving networks all over the world forward. And I'd really encourage everyone to take a look at that. It's on YouTube to take a look at the presentation that CREA gave because it's very good on the whole network digital twin area and shows that we've been using it for some time.
Guy Daniels, TelecomTV (25:44):
Great. Thanks so much Neil. We'll go and take a look at that later. And it's a very popular topic, this because we've just had another question along very similar lines to that question. We're asking at the moment all about digital twins and energy optimization. So thanks for sending these questions in Diego. We're going to come to you and then Beth, we'll pop back to you for a final comment on that. But Diego, what are your thoughts?
Diego Lopez, Telefónica & ETSI (26:08):
I would like to remark a couple of things that is true AI may bring to the digital twins that we have been, as Neil and Beth were saying for quite a long time in well long in technology terms probably it's not that we have been using them since then the fifties, but in terms of how the technology has matured, it's a long time. One is the fact that the user of AI should help in changing the focus of the twin dynamically. The point is that with ai, with mechanisms of generative ai, what you can do is that you can emulate pieces of the network or certain segments or certain aspects in a match more transparent way that we do right now. Because right now we need models of everything, detailed models we can make right now generative AI models learn the behavior without us having a detailed model of a certain matter.
(27:18):
And we are right now, meaning in our labs, we are using models in which we are emulating, for example, the behavior of the RAN without having a complete detail emulated model of the ran. So we can make some analysis on the impact of certain behaviors in the core network. For example, this is one thing. The other thing is the possibility of submitting the twins or the twin elements to conditions that are unusual, that are difficult to reproduce or that difficult to generate. In other means just by tweaking the mechanisms of the ai, so the parameters of the ai, not the mechanisms itself. So I believe there is a potential of growth there we will witness in the future, I believe a strong evolution that would help us in improving this and again, in a multidimensional aspect, not only with the energy efficiency, but in general with everything that has to do with network management and optimization.
Guy Daniels, TelecomTV (28:38):
Sure. Yeah, thanks very much Diego. And as you say, multidimensional, it's not just one focus on one aspect. We take it more holistically. We'll come across to Mirko next and then we'll pick up extra comments from Beth and Neil. But Mirko, what would you like to add?
Mirko Voltolini, Colt Technology (28:52):
Yeah, I think I couldn't agree more on what has been said. If you think about it, network management systems have had the capability to simulate traffic for a very long time. So we actually had digital twins already many, many years ago. As also be pointed out, I think what AI is bringing now to this simulation capabilities, the fact that you can actually more easily run more complex simulations. If you think about it, the complexity of I look at our net, we have over 500 nodes, a graph of that size if you want to run a net full traffic simulation will take with normal methodologies a very long time to run. And now you can actually run this with theistic coming from AI that would allow us to do much faster. So I think there is a scale improvement that with applying AI capabilities into these type of models.
Guy Daniels, TelecomTV (29:49):
Yeah, absolutely. Thank you very much Mirko, and we'll come across for some extra comments. And Beth, let's start with you. You wanted to add something there?
Beth Cohen, Luth Computer Specialists (29:56):
Yes. So I wanted to pick up on some things that both Neil and Diego referred to, which is, I don't think they use the term, but observability, which is digital twins and the network in general, you can't measure and you can't act on what you don't see. And I think that's where digital twins can really enhance the capability to, you can add the additional, you can tweak things, you can use additional tools to allow you to add that extra layer of observability to drive what you need to look for. And I know that's been a problem in the past for energy efficiency, which is that most of the telco tools in the past have not particularly looked at energy efficiency as part of the levers, if you will, for performance. And that's something I know that the hardware vendors have stepped up on and given us more tools. So that really allows us to do that, the kind of deep experimenting that we really need to do to figure out what's best. And that's where I think AI can really help with that because AI will only produce results for the data that goes in garbage, in, garbage out. So you need to make sure that the data that goes in is the data you want to look at so that the AI can give you results. That make sense.
Guy Daniels, TelecomTV (31:40):
Thanks very much Beth and over to Neil for some extra comments on this question around digital twins.
Neil McRae, Juniper Networks (31:47):
Yeah, I mean just one thing very quickly, the other great use of digital twins is training ML and AI capabilities. So you build a digital twin of your network and you can let the AI learn from it. And I wouldn't underestimate the value of that as we progress through enabling AI to do more in networking. So our MISS system keeps getting better at fixing stuff because it's learning. And we have a product called Marvis Mini that literally enables our AI to create a digital twin of your network. And one of the greatest value points of that is the learning of what's there and then taking that to fix problems or deal with issues in advance so that customers don't see any downtime. So if you have a digital twin, figure out how you plug something in that can learn from it because that in itself can be hugely valuable.
Guy Daniels, TelecomTV (32:47):
Yeah, that's very interesting. Thank you very much, Neil. Thanks everyone for your comments on that question. Before we look at another question, it is time to check in on our audience poll for the Green Network Summit. And the question we are asking you this week is how can telcos most effectively reduce their energy consumption levels? And you can see the real-time results appear to my right here. Well, yeah, the solution appears to be in the hands of the vendors there and certainly not in the cloud, but look, it's early days. Those figures will no doubt change over the coming hours and days. If you have yet to vote, then please do so and we'll take another look at the voting during tomorrow's live q and a show. Right, more questions then. We've still got a bit of time. We have about 25 minutes into the show. Another question here from one of our audience members, it says it was stated at your telcos and AI event last December that there was a disappointing lack of AI deployments in the telecom network, not just for energy and power optimization. What would today's panelists like to see happen this year? Well, Beth, if I can come across to you first, we had our event in December, this topic was mentioned, but we really want to focus on what you'd all like to see happen and develop in terms of the green network this year.
Beth Cohen, Luth Computer Specialists (34:31):
Yeah, I was going to say, what's to say that the telcos have been using AI for a number of years, primarily in operations. To build on what Neil said, if you're a customer and you're placing an order through a chatbot, that's a chatbot. And so the telcos are definitely using ai. However, I think again, we need to see the ROI and to make sure that it works before deploying it in the network as intelligent networking. So I think you'll see more and more AI coming in at the edges and then moving toward the core over the next year or so. And I think that's the right approach because as again, Neil said Telcos need to be risk adverse because that's the promise to their customers that it's five and six nines. And so if AI can help deliver that, great, it'll get deployed. If it can't, then it won't. So I think it's a very practical approach that the telcos need to take toward implementing AI in the core networks.
Guy Daniels, TelecomTV (36:04):
Neil, we're coming across to you because I know you were at the event. What would you like to see happen this year, Neil?
Neil McRae, Juniper Networks (36:10):
Yeah, I think I made the challenge, which is when I look at some industries, I see them actively investing in AI and more automation and taking a step back and looking at their business through the lens of ai. And I would like to see more and more service providers look at that. I'd like to see them do more experiments, figure out how AI can apply to some of their most difficult workloads. If I use one as an example, wireless spectrum planning for wireless service providers is one of the most difficult tasks. And literally the way that we do it today historically is very inaccurate, I would say. And as the world demands more data, we need more spectrum and we need that spectrum to go in places that it hasn't gone before. So for me, I'd really like to see, I'd love to see use cases where people are able to get 15% increase in spectrum utilization because they were using AI in the networking space.
(37:21):
Generally I'd like to see how we're taking out light and see how we're taking out packet loss, how we're making the network much more reliable, more than the five nines that we've always delivered, right? We've been doing that for 10 years. Where's the sixth? Nine? How do we really push the network forward? And I don't say that just to grab a number or a headline. I say it because the more the network's reliable, the more people can depend upon it, the more workloads come to the network. So I really want to see service providers look at that. And then of course the number one use is making it easier for customers to do business with service providers. It's one of the biggest challenges that service providers face. How do they make it easier and easier for customers to order services for customers, to make changes, and for customers to ask for different things That traditionally has been very manual, very knife and fork and very non-digital I think is the word I'd look for of the two words. It's really how do we use AI to make the experience of using the network of buying, the network, of changing the network as easy as it should.
Guy Daniels, TelecomTV (38:36):
Great. Thanks very much Beth. And we'll come to Mirko in a moment, but first of all, Diego, let's get your views on this. Thanks.
Diego Lopez, Telefónica & ETSI (38:45):
Just a note in something that I mentioned the other day is that the fact that there are AI applications that have not necessarily focused on energy efficiency in evaluating agency efficiency or in applying or deciding mechanisms for placements or routing or particular keeping certain gear on or off, but that can translate into a high energy efficiency. And the clear example is when it comes with things that have to do with limiting the number of truck rolls to customer sites to solve problems or mechanisms that would improve the efficiency of maintenance working remote elements of the network. Because we are widespread, we're talking about one ones and the areas can be very wide, and these kind of activities that are in which we are already using AI and telcos have been using AI for this optimization. The only problem is that we are not necessarily measuring the energy impact that this has might be that one of the things that we could do to prove that we're improving our footprint and that we are doing so by using advanced tools like AI could be basically, and this is something I would like to see is about measuring this impact and making our customers and the society at large aware of this.
Guy Daniels, TelecomTV (40:36):
Thanks very much Diego. And let's go across to Mirko as to your thoughts on what you'd like to see.
Mirko Voltolini, Colt Technology (40:44):
Yeah, so Lance have this in a generic way rather than just specifically on energy efficiency and also bring our experience. If you look at what has been happening over the last few years, we have invested a lot in experimentation. There is a number of use cases that we have been developing internally. What I like to see is that more and more of these use cases actually make it to prime time. So the rich maturity and we're able to industrialize them. If I look at, it's kind of like a funnel, we have had so many ideas and different initiatives, a few of them have made it to full production, but still quite a lot of them are either confined into specific areas or maybe it's just still in the lab. And so I think getting some of those out of the lab and into full implementation will be beneficial to do that. You have to put in place, I mean working in a lab environment is kind of easy enough when you start to deploy things in production, you have to put in place safety mechanisms, you have to industrialize, you have to train operations, you have to have to basically scale it to a way that that can should be used in full production. So I think that is my wish for this period now.
Guy Daniels, TelecomTV (42:00):
Great, thanks and thanks everyone for those comments. I'd like to thank our audience as well for sending in questions, but we've got a load of questions just received here in the past five or 10 minutes. There's about eight or nine questions, which unfortunately we will not have time to get through today, but I'm going to look at those later and we'll allocate those to tomorrow's live show. So tune in tomorrow because some of these late arriving questions will hopefully get to tomorrow. We still have time though. Possibly one more question today. So let me read this one out. Do neutral host networks represent the best pathway for a sustainable delivery of 5G sa, especially with regards to energy and emissions and should this approach be mandated through policy? And I'll add to that another question we received about 10 minutes ago, which says, is there any room for policy makers to contribute to make telecom networks greener and more efficient? So neutral host networks and the mandated policy aspect. Diego, I'm going to throw this one at you first. I'm sorry about that, but what do you make of this?
Diego Lopez, Telefónica & ETSI (43:15):
I'm happy regarding neutral hosts, I'm not sure we don't have the measures and the evidence that this would simplify, reduce the energy consumption for sure, it would reduce the cost of deployment in some cases, that's for sure because we would have a single one. I'm not sure whether the requirements for serving several competing requests would not translate in over provisioning in some cases and over consumption in others, we would need evidence and measurements that we don't have so far as far as I can tell. But for the second question about whether this should be mandated or be part of regulations, et cetera, my answer is very short is no. The telco sector is extremely overregulated. This translates into higher costs and lower innovation rates. And the last thing I believe we need is by any means more regulations in whatever the sense and in the particular in this case, for sure not.
Guy Daniels, TelecomTV (44:35):
Thanks very much Diego. I think you'll have a lot of our audience agreeing with you on that last point. This is our last question looking at the time. So let me go across to some of my other guests and Beth, lemme come across to you next.
Beth Cohen, Luth Computer Specialists (44:47):
So I have slightly different take on it. First of all, neutral host is not a new idea. I mean we've been having peering networks for decades. So neutral host is not that different a concept. And I think we should take that as the model to provide the neutral hosts should they be mandated. I think the telcos need to work together to do it because many of the telcos will say, Hey, I'll be the neutral host and I'll be able to kind of run things. But it really does need to be appearing proper peering relationship. On the second question, I don't think it should be mandated so much as it needs to be encouraged. And sadly I will use the US whip whipsaw, if you will, of the energy policies that's going on right now where the US has literally gone back and forth over the last 10 years and literally in the last month at the speed of light, if you will, in changing what its priorities for energy efficiency are. So it's not so much that it needs to be mandated, but it needs to be encouraged. And that's two different things. So energy efficiency is in my mind sort of the obvious thing that we should pursue just for cost savings and good for the planet and all that good stuff, but it doesn't need to be regulated to death either.
Guy Daniels, TelecomTV (46:35):
Yeah, thanks Beth. And it's a good differentiation there between those two approaches. Neil, what are your thoughts on this question about neutral host and also the policy side of it?
Neil McRae, Juniper Networks (46:48):
Well, look, I'm with Diego on this. I don't think we get any more policies. I think let innovation roll out and if it needs regulation then regulate it. But the moment our industry's scared to do anything for fear of regulating, and we obviously need sensible regulation, but I don't know how you can sens be regulated until you've got enough of knowledge about something and let's take neutral hosts. So I think the devil's in the detail of this question because look, neutral hosts are out there. You go into quite a few malls or stadiums, you'll probably find neutral hosts, but it's a very specific question about 5G SA. And of course 5G SA is the kind of what we call the full fat 5G where it's got latency, it's got, sorry, low latency, it's got high availability, it's got all the features that we thought would make a difference in the world.
(47:47):
And I think to do that on a neutral host looks challenging. It's not impossible, but I think it's hard and I think it's hard to guarantee it if you're selling a service. So imagine operator builds a neutral host in an airport, but then the airport wants to buy service from a service provider that isn't a neutral host provider. How do you square that circle in ensuring that you've got the right service levels for a 5G SA application? I think that's pretty hard. So is neutral host a tool that we need to use more of a hundred percent. There's very many places where neutral hosts where using a neutral host to provide coverage in a car park isn't going to differentiate you with customers, but it's going to bring value add and it's going to keep CapEx low. So that's a great place to use it in transportation hubs, ensuring that when people are traveling or there's a mass collection of people that you make the most out of the infrastructure.
(48:49):
But that's a balance. So does somewhere like Heathrow Airport need more than one operator rather than a neutral host? I would say yes, but does LEED Bradford need more than one operator? I'd say probably not. And for those of you watching outside, he throws this massive place and leads us this tiny little place just in the north of the uk. So you got different scales going to drive different requirements for needs. But neutrals definitely are an area of investment and I think we're going to see a lot more of it, particularly in places like beach towns or holiday locations where it's too expensive for service providers to build big macro infrastructure. But if you have a neutral host, we can share that neutral host and we can give great service to our customers whilst they're on holiday and they only needed maybe three or four months in a year, at least in the uk.
(49:49):
And I think the same will be true in other geographies. So very many use cases for neutral host. Is 5G SA the strongest use case for it? Maybe, maybe not. I think we need to experiment and trial things a bit more, but certainly sharing infrastructure generically where it doesn't differentiate us, I would think this is something of value, but I think we also need to be able to build infrastructure and have that competitive element that drives our service providers to build better networks every day. And if we make everything the same, we'll just end up with a lowest common denominator, which I don't think is for the industry.
Guy Daniels, TelecomTV (50:28):
Great. Thank you very much, Neil. Well, I think that's all the responses for that question, which is indeed our final question of today. We are out of time now, so thank you to all of our guests who joined us for this live program. Do you remember though, to send in your questions for tomorrow's live q and a show as soon as you can? Don't leave it too late. As I mentioned earlier, we've already got about half a dozen or so good ones that we're going to look at for tomorrow's program. And please do take part in the poll. There is still time for you to have your say. Well, you can find the full agenda for day two of the summit on the telecom TV website, and it includes a panel discussion on the benefits of building a green radio access network. Remember, you can watch that on demand from tomorrow morning and for our viewers watching live today, in case you missed the earlier panel discussion, we are going to broadcast it in just a few moments, so don't go away. We'll be back tomorrow with our final live q and a show, same time, same place. Until then, thank you for watching and goodbye.
Hello, you are watching the Green Network Summit, part of our year round DSP Leaders coverage, and it's time now for our live Q and A show. I'm Guy Daniels, and this is the first of two q and a shows. We have another one at the same time. Tomorrow it's your chance to ask questions on energy efficiency and innovation within the network. Now, as part of today's summit, we featured a panel discussion that looked at how to improve network energy efficiency with ai. And if you miss the panel, which was hosted by my colleague ra, don't worry because we will rebroadcast it straight after this live q and a program, or you can watch it anytime you like on demand. Now, we have already received several questions from you, but if you haven't yet sent one in, then please do so now and use the q and a form that's on the website.
(01:21):
Well, I'm delighted to say that joining me live on the program today are Mirko Voltolini, who is VP of Innovation at Colt Technology, Diego Lopez, Senior Technology Expert at Telefonica, and an ETSI fellow, Neil McRae, Chief Network Strategist for Juniper Networks, and Beth Cohen, Telco Industry Analyst with Luth Computer Specialists. Hello everyone. Good to see you all again. Thanks so much for joining us. Let's get straight on to our first audience question. We've got a mix of questions today. Some are AI related, some are more general energy related. So the first question we've got from our audience is given the dramatic rate of advances in ai, what innovations can be applied now to the network to improve energy efficiency and which will not be outdated in a few months time? When should telcos commit? Diego, how about coming across to you and getting your views on this one first for us?
Diego Lopez, Telefónica & ETSI (02:31):
Well, I would say to start with that the pace of advancement, et cetera, probably is the same that we had some time ago or we still have coming with radio interfaces, with optical networks, with processors, with memories with software itself. I don't think that the pace is going to be that much different and that the solution is going to be the same is about using open interfaces, using the possibility of plugging and unplugging the different components. And for this, what we need is to have clearly defined the interfaces of what we use and a way of describing the different elements that we are going to use. The same way that they use other pieces here is probably those pieces have a different nature. Probably the kind of descriptions that we will need would be less physical, more virtual, more oriented to data rather than to that, for example, connectors or tensions or wavelengths or whatever.
(03:44):
But the idea is, as we have done in the past when we're trying to engineering and applying new technologies, is about making the whole thing interchangeable and interoperable. Basically this tradition has been done by formal standards and that has been created by a process of a consensus, et cetera. Right now we have learned to apply methods that converge in a faster way like the use of open source elements or as reference mechanisms, but I do believe that at least if we don't have the ultimate tools for making this happen, we do have the mindset and the methodologies for doing this. And so I don't see any risk in start committing right here right now to whatever the technology can offer.
Guy Daniels, TelecomTV (04:45):
Great. Thanks Diego. So don't wait. Commit now and we'll get some more views from our other guests. Neil, how about I come across to you next? Because AI has got huge media coverage at the moment and we're seeing advances every day. So is it a case of if you commit now you might find yourself with some outdated applications?
Neil McRae, Juniper Networks (05:06):
Well, I think there is a risk of that, but that's the risk of technology. I do think service providers need to be out there experimenting and learning so that they can make the right calls on what they should invest in, what they shouldn't invest in, and also understand where AI is going to make the biggest amount of difference for 'em. And I've said this before, but if you're very good at what you do, probably AI won't have as big an impact as you think, but if you've got a whole, and I've seen this, telcos in the space of inventory of their infrastructure, lots of things that aren't right in the face of customer service tends to take a backseat. And I think that's an area where AI can help. And one of those areas that historically did take a backseat was energy. Thankfully that's front and foremost for just about every service provider, and that's probably one area where AI can do it, but I think if you're sitting in the corner saying, well, we'll wait a few years for ai, there is some merits to that, but I wouldn't be ignoring it entirely.
(06:14):
I think that would be slightly crazy, but I'd also be encouraging your teams. I used to try and find with my teams, I used to try and find the, we'd have 10% of our time, what's out there? What can we build with something that we might not take the normal process for? So definitely experiment, learn, and then when you think you've got something that's got the right impact to customers, the right impact to TCO and generates value, then execute on it, but have a mind to the future because as you say guy, it's not the media is saying, I think the media is behind on how fast AI is actually moving. I think it's moving faster than any of us really understand or realize.
Guy Daniels, TelecomTV (07:02):
Yeah, I'd agree with you that one, it is hurtling along a rapid pace of nots. Beth, lemme come across to you. I know you want to come in on this one as well. I mean, is this question about when to commit to technology to solutions really focus on ai or is it just a general concern no matter what the technology, what the approach is?
Beth Cohen, Luth Computer Specialists (07:20):
Yeah, my opinion is, I mean, telcos are by their very nature, they're engineering companies and yeah, shiny objects are great, but they also need to deliver a service that works. I used to say when you pick up the phone, you want that, you expect that call to go through 100% of the time, and that still applies. You expect your network to be up 100% of the time. So telcos can't afford to just kind of throw new technology in the mix without testing it and making sure that it will actually work. We cannot afford to have a massive outage. So ai, absolutely, the telcos are and should be working with ai, looking at where it makes the most sense. There's been great strides in using the more proven AI solutions such as chatbots and operational, using it to support operational efficiencies. Its support within what I call network intelligence.
(08:41):
I would still say that is definitely a work in progress and there's a lot of things that need to be done before it can be deployed out fully in the field in production. However, that should not stop any of the telcos from digging in and seeing what's there and more importantly, contributing. I know that Diego mentioned open source, there's a number of open source projects in the telco space looking at seeing how AI can be used within the telco space. And there's a number of hurdles that need to be gotten over just to make it more usable and contribute to the algorithms that need to be developed.
Guy Daniels, TelecomTV (09:34):
Great. Thanks very much, Beth. So clear message there is experiment, get involved, don't wait around. Excuse me. Another question from our audience here. I like this one. We hear about the need to break the energy curve to prevent energy rising in line with data traffic growth, that correlation that's been going on for years and years. Surely the only way to do this effectively is by replacing old infrastructure with new and ensuring that new standards focus on efficiency, which is good news for vendors. But can telcos actually afford this? I like the question, Mirko, let's come across to you. Any thoughts on this one? Is that the right approach and can telcos afford to do it?
Mirko Voltolini, Colt Technology (10:27):
Well, the question probably should be more like can telco not afford to do it? I think what we have seen is, I mean definitely is a question of return of investment, but technology that is coming out now is 10 times more efficient, 10 times faster, 10 times cheaper, and also utilize much less pace. So the business case comparing like for all the technology, newer technology is going to be in favor of deployment of new technology. So the operators will do that anyway. I think the key aspect here is also making sure that you actually remove older technology because this is what probably Telco haven't been the good over the years. We keep piling up older technology, new technology and all the technologies isn't going to help you. So I think you need to make a concerted effort to remove legacy network and migrate over time. I think programs can be very efficient, even though they are demanding in terms of resources and time being spent on it. But I think it's definitely an aspect, an area where money can be made and can be saved. So we're actually investing in this. We have an old migration program where we keep deploying new technology, move out. Also all the locations reduce space and reduce power and increase capacity. So there is definitely a case for this.
Guy Daniels, TelecomTV (12:01):
Thank you Mirko. I like that. Can Turco has not afford to do this? Take this approach. Beth, I'm going to come across to you because there was a related question about this energy curve and this relationship as well, and another viewer talking and comparing the telcos with the data center sector and looking at how the data center energy usage is ramping up as a data center. Rollouts are increasing and there's a lot of concern about what these values are and what does it all mean? What's your take? What's take on trying to break this relationship and what are your thoughts?
Beth Cohen, Luth Computer Specialists (12:42):
So this is a very complex question. So first, the telcos are absolutely interested in reducing energy footprint and they have been quite actively pursuing it in certainly around the globe. However, in the us, the US had got their first, if you will. So there's a lot of legacy equipment out there that obviously needs to be decommissioned still. I mean literally in my town they pulled the copper out about a year ago and making a number of elderly people very mad. But the point is, is that the fact that we still have copper infrastructure out there is an indication that it's time to upgrade and this equipment is long past its end of life. And now let me talk about the telco versus the data centers. So telcos, it's a mature industry. The network traffic is growing, but it's not growing at the same pace as the data center industry, which is growing at what, 45% a year or whatever.
(14:07):
So it's a little bit of a red herring to say, oh yeah, the data centers are using far more energy than telcos or the rise in energy uses far more than the telcos because the industry is newer. In fact, the data center industry has been thinking about the costs of the energy since the beginning pretty much because they're concentrated. So they see the giant electric bills in a different way than the telcos do because the telcos, the energy is spread over a wide area network with hundreds of thousands or millions of nodes. So it's a very complex issue, but I think both industries are working on it, but the answers are slightly different.
Guy Daniels, TelecomTV (15:08):
Yeah. Thanks so much Beth. I think it was good to point out there that these industries are working on this. They're not ignoring it. No evidence of that at all. Neil, did you want to come in on this question about breaking the energy curve and looking at whether or not the approach is to replace existing less efficient kit with new more efficient kit complying to the latest standards? That's the only really long-term sustainable way of fixing this.
Neil McRae, Juniper Networks (15:35):
I don't think it's the only way, but I think it's definitely a way to think about it. I mean, you walk around any central office, you probably find PDH equipment, SDH equipment, first gen optical equipment, an average SDH rings 155 megabits and it probably costs the same amount and power that a 400 gig router is so for sure turn off the old stuff, migrate customers to newer technologies. I think telcos have done an okay job of that. I think they need to do a better job. I think Beth's a hundred percent right on copper switch off in the world. We are so far behind on copper switch off. We need to go way more aggressive at that, not just from an energy point of view, but from a digital enablement point of view. But lemme just mention one thing though. I think there's a nervousness that everything in the network or the data centers is bad when actually data centers have made a dramatic contribution to reducing energy and carbon by enabling people to appear on events like this rather than having to get in a car or fly to a place or travel.
(16:55):
We can all telecommunicate or we can telework or we can do loads of things that in the past just were not possible. Even during covid, we saw one of the vehicle manufacturing teams using edge compute with digital twin to do design and to test stuff in M 3D printing prototypes in their office. So I think when you look at data centers in the network and you say, wow, they use a lot of energy, but don't forget how much value they bring us, which is substantial, but also think about user behavior. I mean, does anyone think about how much it costs in energy when they download that funny YouTube video? Probably not. But it has a cost and I think as sensible users, we may want to think about that a little bit more, although let's not stop watching funny cat videos because they're the highlight of my day.
(17:57):
But there's a sensible approach to this where if we're smart and if we look at things and I've not met a telco that doesn't know what it needs to do there it is just how do you make it happen? And I think that's where AI can potentially help because it can crunch a load of data and help service providers make decisions that are typically really, really difficult. In my past, we were working on copper switch off and it was one of the hardest bits of work to do because you uncover so many things. So for example, how do you do telephony in an elevator if you don't have copper? And that telephony is the emergency for if something happens in the elevator, how do you do looking after those that need help at home if you don't have copper? Because all the things that are out there use copper networks today.
(18:54):
So we've obviously got to do this sensibly, but one of the things I hear and I talk to customers about is network modernization. It's bringing the latest in silicon, bringing the latest in optics, bringing the latest in power management, bringing the latest in security, but also enabling new services. You put those five into a business case, I challenge anyone to see that there isn't a TCO there that is valuable for them, but most importantly brings a capability to their customers that's valuable for them and generates revenue. So don't think of this as just a one directional street that it's all about taking out costs, which is super important, but it's also about driving new capabilities for customers, which is also important and enables your customers to reduce their power and their energy consumption.
Guy Daniels, TelecomTV (19:44):
Yeah, that's the balance. Thank you very much Neil for those comments. Much appreciated. I think we'll move on to another question then. So we've got a few questions in from our viewers already, right? This one's about digital twins as we've just briefly heard. So the question is AI can be used to create digital twins that allow telcos to run virtual simulations of their real life networks enabling experimentation to help improve energy efficiency. But is this just an academic exercise or can it realistically be used to optimize physical networks today? What's the best approach for telcos to take? Well, digital twins have been around for a long time. Beth, do you want to have a crack at this one first and give us your views on the use of digital twins here?
Beth Cohen, Luth Computer Specialists (20:43):
Certainly. So yes, digital twins have around for many, many years I had a lab at Renick and Newman. We were the inventors of the internet. We did several digital twins for the network and that goes back 30 years. So they're not new and they're clearly of great value to the telco industry. So it is definitely not an academic exercise. In fact, I would argue that telcos use them very heavily for all sorts of things, not just energy efficiency. I think that's something relatively new, but certainly for network optimization calculations, for determining best path, all sorts of all sorts of good things. And so adding energy efficiency into those test harnesses is only just goodness.
(21:52):
Digital twins are very valuable to the telco community. I would like to see, most telcos do have them. I would say probably all of them do. I would like to see some more digital twins in the open source community. I know LFN does have a number of platforms that they make available to the community that can be used in this manner. And I think that would help further some of the more standards, some of the more areas where we do need standards and we do need open source to optimize not just the performance of the network, but also the physical and energy efficiency of the networks as well.
Guy Daniels, TelecomTV (22:45):
Great. Thanks very much Beth. We'll come across to Diego in a moment, but I just want to go first to Neil because Neil AI likes to gobble up a lot of existing technologies and claim them for itself. We know this seems to be a case of one of them, but what about this application of digital twins to help, especially with energy efficiency of physical networks?
Neil McRae, Juniper Networks (23:07):
Yeah, look, so data center builders have been using digital twins to design data centers for probably for, I'm going to say more than five, maybe even as much as 10 years. I remember being part of a data center build for a customer where they were plotting the best way of using the space, the easiest way of getting power, making sure that the air conditioning and the flow of AR was going to deliver the yield that they wanted in that data center. So I don't think this is anything super new and I agree super with be on. We've been using digital twins in the networking space for some time. I think the difference AI is going to bring is an ability to run simulations probably much more quicker than we were able to in the past. So I would always get a question, well Neil, when we run the test, why didn't we discover this issue?
(24:07):
Well, it's just because we had so many limited time to run different test criteria. I think ai, I hope, will give us the ability to run more and more test criteria. But I want to point out a piece of work that my colleague Pel did who works in our one group here at Juniper at nano last year. He did a really great presentation on network digital twins. I'd encourage everybody to take a look at, we think this is a super important thing. We're building it into many of our tools for service providers to allow them to do more and more what ifs and why are we doing that? Well, one, to optimize the network to modernize the network, but two, network reliability has just become such a table stake item. You cannot have outages, you cannot take stuff down, which often slows service providers down because they're frightful of the risk of making changes.
(25:04):
Whereas actually what we believe today is by not making changes, you put yourself at greater risk, not updating code so that it doesn't have any bugs or security issues. Getting to the latest platforms that are able to scale more. So enabling service providers to do more by seeing it simulated in front of 'em is definitely a way that we're moving networks all over the world forward. And I'd really encourage everyone to take a look at that. It's on YouTube to take a look at the presentation that CREA gave because it's very good on the whole network digital twin area and shows that we've been using it for some time.
Guy Daniels, TelecomTV (25:44):
Great. Thanks so much Neil. We'll go and take a look at that later. And it's a very popular topic, this because we've just had another question along very similar lines to that question. We're asking at the moment all about digital twins and energy optimization. So thanks for sending these questions in Diego. We're going to come to you and then Beth, we'll pop back to you for a final comment on that. But Diego, what are your thoughts?
Diego Lopez, Telefónica & ETSI (26:08):
I would like to remark a couple of things that is true AI may bring to the digital twins that we have been, as Neil and Beth were saying for quite a long time in well long in technology terms probably it's not that we have been using them since then the fifties, but in terms of how the technology has matured, it's a long time. One is the fact that the user of AI should help in changing the focus of the twin dynamically. The point is that with ai, with mechanisms of generative ai, what you can do is that you can emulate pieces of the network or certain segments or certain aspects in a match more transparent way that we do right now. Because right now we need models of everything, detailed models we can make right now generative AI models learn the behavior without us having a detailed model of a certain matter.
(27:18):
And we are right now, meaning in our labs, we are using models in which we are emulating, for example, the behavior of the RAN without having a complete detail emulated model of the ran. So we can make some analysis on the impact of certain behaviors in the core network. For example, this is one thing. The other thing is the possibility of submitting the twins or the twin elements to conditions that are unusual, that are difficult to reproduce or that difficult to generate. In other means just by tweaking the mechanisms of the ai, so the parameters of the ai, not the mechanisms itself. So I believe there is a potential of growth there we will witness in the future, I believe a strong evolution that would help us in improving this and again, in a multidimensional aspect, not only with the energy efficiency, but in general with everything that has to do with network management and optimization.
Guy Daniels, TelecomTV (28:38):
Sure. Yeah, thanks very much Diego. And as you say, multidimensional, it's not just one focus on one aspect. We take it more holistically. We'll come across to Mirko next and then we'll pick up extra comments from Beth and Neil. But Mirko, what would you like to add?
Mirko Voltolini, Colt Technology (28:52):
Yeah, I think I couldn't agree more on what has been said. If you think about it, network management systems have had the capability to simulate traffic for a very long time. So we actually had digital twins already many, many years ago. As also be pointed out, I think what AI is bringing now to this simulation capabilities, the fact that you can actually more easily run more complex simulations. If you think about it, the complexity of I look at our net, we have over 500 nodes, a graph of that size if you want to run a net full traffic simulation will take with normal methodologies a very long time to run. And now you can actually run this with theistic coming from AI that would allow us to do much faster. So I think there is a scale improvement that with applying AI capabilities into these type of models.
Guy Daniels, TelecomTV (29:49):
Yeah, absolutely. Thank you very much Mirko, and we'll come across for some extra comments. And Beth, let's start with you. You wanted to add something there?
Beth Cohen, Luth Computer Specialists (29:56):
Yes. So I wanted to pick up on some things that both Neil and Diego referred to, which is, I don't think they use the term, but observability, which is digital twins and the network in general, you can't measure and you can't act on what you don't see. And I think that's where digital twins can really enhance the capability to, you can add the additional, you can tweak things, you can use additional tools to allow you to add that extra layer of observability to drive what you need to look for. And I know that's been a problem in the past for energy efficiency, which is that most of the telco tools in the past have not particularly looked at energy efficiency as part of the levers, if you will, for performance. And that's something I know that the hardware vendors have stepped up on and given us more tools. So that really allows us to do that, the kind of deep experimenting that we really need to do to figure out what's best. And that's where I think AI can really help with that because AI will only produce results for the data that goes in garbage, in, garbage out. So you need to make sure that the data that goes in is the data you want to look at so that the AI can give you results. That make sense.
Guy Daniels, TelecomTV (31:40):
Thanks very much Beth and over to Neil for some extra comments on this question around digital twins.
Neil McRae, Juniper Networks (31:47):
Yeah, I mean just one thing very quickly, the other great use of digital twins is training ML and AI capabilities. So you build a digital twin of your network and you can let the AI learn from it. And I wouldn't underestimate the value of that as we progress through enabling AI to do more in networking. So our MISS system keeps getting better at fixing stuff because it's learning. And we have a product called Marvis Mini that literally enables our AI to create a digital twin of your network. And one of the greatest value points of that is the learning of what's there and then taking that to fix problems or deal with issues in advance so that customers don't see any downtime. So if you have a digital twin, figure out how you plug something in that can learn from it because that in itself can be hugely valuable.
Guy Daniels, TelecomTV (32:47):
Yeah, that's very interesting. Thank you very much, Neil. Thanks everyone for your comments on that question. Before we look at another question, it is time to check in on our audience poll for the Green Network Summit. And the question we are asking you this week is how can telcos most effectively reduce their energy consumption levels? And you can see the real-time results appear to my right here. Well, yeah, the solution appears to be in the hands of the vendors there and certainly not in the cloud, but look, it's early days. Those figures will no doubt change over the coming hours and days. If you have yet to vote, then please do so and we'll take another look at the voting during tomorrow's live q and a show. Right, more questions then. We've still got a bit of time. We have about 25 minutes into the show. Another question here from one of our audience members, it says it was stated at your telcos and AI event last December that there was a disappointing lack of AI deployments in the telecom network, not just for energy and power optimization. What would today's panelists like to see happen this year? Well, Beth, if I can come across to you first, we had our event in December, this topic was mentioned, but we really want to focus on what you'd all like to see happen and develop in terms of the green network this year.
Beth Cohen, Luth Computer Specialists (34:31):
Yeah, I was going to say, what's to say that the telcos have been using AI for a number of years, primarily in operations. To build on what Neil said, if you're a customer and you're placing an order through a chatbot, that's a chatbot. And so the telcos are definitely using ai. However, I think again, we need to see the ROI and to make sure that it works before deploying it in the network as intelligent networking. So I think you'll see more and more AI coming in at the edges and then moving toward the core over the next year or so. And I think that's the right approach because as again, Neil said Telcos need to be risk adverse because that's the promise to their customers that it's five and six nines. And so if AI can help deliver that, great, it'll get deployed. If it can't, then it won't. So I think it's a very practical approach that the telcos need to take toward implementing AI in the core networks.
Guy Daniels, TelecomTV (36:04):
Neil, we're coming across to you because I know you were at the event. What would you like to see happen this year, Neil?
Neil McRae, Juniper Networks (36:10):
Yeah, I think I made the challenge, which is when I look at some industries, I see them actively investing in AI and more automation and taking a step back and looking at their business through the lens of ai. And I would like to see more and more service providers look at that. I'd like to see them do more experiments, figure out how AI can apply to some of their most difficult workloads. If I use one as an example, wireless spectrum planning for wireless service providers is one of the most difficult tasks. And literally the way that we do it today historically is very inaccurate, I would say. And as the world demands more data, we need more spectrum and we need that spectrum to go in places that it hasn't gone before. So for me, I'd really like to see, I'd love to see use cases where people are able to get 15% increase in spectrum utilization because they were using AI in the networking space.
(37:21):
Generally I'd like to see how we're taking out light and see how we're taking out packet loss, how we're making the network much more reliable, more than the five nines that we've always delivered, right? We've been doing that for 10 years. Where's the sixth? Nine? How do we really push the network forward? And I don't say that just to grab a number or a headline. I say it because the more the network's reliable, the more people can depend upon it, the more workloads come to the network. So I really want to see service providers look at that. And then of course the number one use is making it easier for customers to do business with service providers. It's one of the biggest challenges that service providers face. How do they make it easier and easier for customers to order services for customers, to make changes, and for customers to ask for different things That traditionally has been very manual, very knife and fork and very non-digital I think is the word I'd look for of the two words. It's really how do we use AI to make the experience of using the network of buying, the network, of changing the network as easy as it should.
Guy Daniels, TelecomTV (38:36):
Great. Thanks very much Beth. And we'll come to Mirko in a moment, but first of all, Diego, let's get your views on this. Thanks.
Diego Lopez, Telefónica & ETSI (38:45):
Just a note in something that I mentioned the other day is that the fact that there are AI applications that have not necessarily focused on energy efficiency in evaluating agency efficiency or in applying or deciding mechanisms for placements or routing or particular keeping certain gear on or off, but that can translate into a high energy efficiency. And the clear example is when it comes with things that have to do with limiting the number of truck rolls to customer sites to solve problems or mechanisms that would improve the efficiency of maintenance working remote elements of the network. Because we are widespread, we're talking about one ones and the areas can be very wide, and these kind of activities that are in which we are already using AI and telcos have been using AI for this optimization. The only problem is that we are not necessarily measuring the energy impact that this has might be that one of the things that we could do to prove that we're improving our footprint and that we are doing so by using advanced tools like AI could be basically, and this is something I would like to see is about measuring this impact and making our customers and the society at large aware of this.
Guy Daniels, TelecomTV (40:36):
Thanks very much Diego. And let's go across to Mirko as to your thoughts on what you'd like to see.
Mirko Voltolini, Colt Technology (40:44):
Yeah, so Lance have this in a generic way rather than just specifically on energy efficiency and also bring our experience. If you look at what has been happening over the last few years, we have invested a lot in experimentation. There is a number of use cases that we have been developing internally. What I like to see is that more and more of these use cases actually make it to prime time. So the rich maturity and we're able to industrialize them. If I look at, it's kind of like a funnel, we have had so many ideas and different initiatives, a few of them have made it to full production, but still quite a lot of them are either confined into specific areas or maybe it's just still in the lab. And so I think getting some of those out of the lab and into full implementation will be beneficial to do that. You have to put in place, I mean working in a lab environment is kind of easy enough when you start to deploy things in production, you have to put in place safety mechanisms, you have to industrialize, you have to train operations, you have to have to basically scale it to a way that that can should be used in full production. So I think that is my wish for this period now.
Guy Daniels, TelecomTV (42:00):
Great, thanks and thanks everyone for those comments. I'd like to thank our audience as well for sending in questions, but we've got a load of questions just received here in the past five or 10 minutes. There's about eight or nine questions, which unfortunately we will not have time to get through today, but I'm going to look at those later and we'll allocate those to tomorrow's live show. So tune in tomorrow because some of these late arriving questions will hopefully get to tomorrow. We still have time though. Possibly one more question today. So let me read this one out. Do neutral host networks represent the best pathway for a sustainable delivery of 5G sa, especially with regards to energy and emissions and should this approach be mandated through policy? And I'll add to that another question we received about 10 minutes ago, which says, is there any room for policy makers to contribute to make telecom networks greener and more efficient? So neutral host networks and the mandated policy aspect. Diego, I'm going to throw this one at you first. I'm sorry about that, but what do you make of this?
Diego Lopez, Telefónica & ETSI (43:15):
I'm happy regarding neutral hosts, I'm not sure we don't have the measures and the evidence that this would simplify, reduce the energy consumption for sure, it would reduce the cost of deployment in some cases, that's for sure because we would have a single one. I'm not sure whether the requirements for serving several competing requests would not translate in over provisioning in some cases and over consumption in others, we would need evidence and measurements that we don't have so far as far as I can tell. But for the second question about whether this should be mandated or be part of regulations, et cetera, my answer is very short is no. The telco sector is extremely overregulated. This translates into higher costs and lower innovation rates. And the last thing I believe we need is by any means more regulations in whatever the sense and in the particular in this case, for sure not.
Guy Daniels, TelecomTV (44:35):
Thanks very much Diego. I think you'll have a lot of our audience agreeing with you on that last point. This is our last question looking at the time. So let me go across to some of my other guests and Beth, lemme come across to you next.
Beth Cohen, Luth Computer Specialists (44:47):
So I have slightly different take on it. First of all, neutral host is not a new idea. I mean we've been having peering networks for decades. So neutral host is not that different a concept. And I think we should take that as the model to provide the neutral hosts should they be mandated. I think the telcos need to work together to do it because many of the telcos will say, Hey, I'll be the neutral host and I'll be able to kind of run things. But it really does need to be appearing proper peering relationship. On the second question, I don't think it should be mandated so much as it needs to be encouraged. And sadly I will use the US whip whipsaw, if you will, of the energy policies that's going on right now where the US has literally gone back and forth over the last 10 years and literally in the last month at the speed of light, if you will, in changing what its priorities for energy efficiency are. So it's not so much that it needs to be mandated, but it needs to be encouraged. And that's two different things. So energy efficiency is in my mind sort of the obvious thing that we should pursue just for cost savings and good for the planet and all that good stuff, but it doesn't need to be regulated to death either.
Guy Daniels, TelecomTV (46:35):
Yeah, thanks Beth. And it's a good differentiation there between those two approaches. Neil, what are your thoughts on this question about neutral host and also the policy side of it?
Neil McRae, Juniper Networks (46:48):
Well, look, I'm with Diego on this. I don't think we get any more policies. I think let innovation roll out and if it needs regulation then regulate it. But the moment our industry's scared to do anything for fear of regulating, and we obviously need sensible regulation, but I don't know how you can sens be regulated until you've got enough of knowledge about something and let's take neutral hosts. So I think the devil's in the detail of this question because look, neutral hosts are out there. You go into quite a few malls or stadiums, you'll probably find neutral hosts, but it's a very specific question about 5G SA. And of course 5G SA is the kind of what we call the full fat 5G where it's got latency, it's got, sorry, low latency, it's got high availability, it's got all the features that we thought would make a difference in the world.
(47:47):
And I think to do that on a neutral host looks challenging. It's not impossible, but I think it's hard and I think it's hard to guarantee it if you're selling a service. So imagine operator builds a neutral host in an airport, but then the airport wants to buy service from a service provider that isn't a neutral host provider. How do you square that circle in ensuring that you've got the right service levels for a 5G SA application? I think that's pretty hard. So is neutral host a tool that we need to use more of a hundred percent. There's very many places where neutral hosts where using a neutral host to provide coverage in a car park isn't going to differentiate you with customers, but it's going to bring value add and it's going to keep CapEx low. So that's a great place to use it in transportation hubs, ensuring that when people are traveling or there's a mass collection of people that you make the most out of the infrastructure.
(48:49):
But that's a balance. So does somewhere like Heathrow Airport need more than one operator rather than a neutral host? I would say yes, but does LEED Bradford need more than one operator? I'd say probably not. And for those of you watching outside, he throws this massive place and leads us this tiny little place just in the north of the uk. So you got different scales going to drive different requirements for needs. But neutrals definitely are an area of investment and I think we're going to see a lot more of it, particularly in places like beach towns or holiday locations where it's too expensive for service providers to build big macro infrastructure. But if you have a neutral host, we can share that neutral host and we can give great service to our customers whilst they're on holiday and they only needed maybe three or four months in a year, at least in the uk.
(49:49):
And I think the same will be true in other geographies. So very many use cases for neutral host. Is 5G SA the strongest use case for it? Maybe, maybe not. I think we need to experiment and trial things a bit more, but certainly sharing infrastructure generically where it doesn't differentiate us, I would think this is something of value, but I think we also need to be able to build infrastructure and have that competitive element that drives our service providers to build better networks every day. And if we make everything the same, we'll just end up with a lowest common denominator, which I don't think is for the industry.
Guy Daniels, TelecomTV (50:28):
Great. Thank you very much, Neil. Well, I think that's all the responses for that question, which is indeed our final question of today. We are out of time now, so thank you to all of our guests who joined us for this live program. Do you remember though, to send in your questions for tomorrow's live q and a show as soon as you can? Don't leave it too late. As I mentioned earlier, we've already got about half a dozen or so good ones that we're going to look at for tomorrow's program. And please do take part in the poll. There is still time for you to have your say. Well, you can find the full agenda for day two of the summit on the telecom TV website, and it includes a panel discussion on the benefits of building a green radio access network. Remember, you can watch that on demand from tomorrow morning and for our viewers watching live today, in case you missed the earlier panel discussion, we are going to broadcast it in just a few moments, so don't go away. We'll be back tomorrow with our final live q and a show, same time, same place. Until then, thank you for watching and goodbye.
Please note that video transcripts are provided for reference only – content may vary from the published video or contain inaccuracies.
Live Q&A Discussion
This live Q&A show was broadcast at the end of day one of the Green Network Summit. TelecomTV’s Guy Daniels was joined by industry guest panellists for this question and answer session. Among the questions raised by our audience were:
- What AI innovations can be applied to the network now to improve energy efficiency, which will not be outdated in a few months time?
- How do we break the ‘energy curve’? Surely the only way is by replacing old infrastructure with new, but can telcos afford to do this?
- Are AI-powered digital twins just an academic exercise or can they realistically be used to optimise physical networks today?
- In terms of AI deployments in the telecom network, what would the panellists like to see happen this year?
- Do neutral host networks represent the best pathway for sustainable delivery of 5G standalone, especially with regards to energy and emission?
- Should green and energy-related matters be mandated through government policy?
Recorded February 2025
Beth Cohen
Telco Industry Analyst, Luth Computer Specialists, Inc.
Diego R Lopez
Senior Technology Expert, Telefónica and ETSI Fellow
Mirko Voltolini
VP of Innovation, Colt Technology
Neil McRae
Chief Network Strategist, Juniper Networks
