ALYNMENT - Private Networks Technology to Business Alignment for Enterprises

Ep # 30: The Practical Approach to Embracing Autonomous 5G Networks - with Aaron Boasman-Patel, TM Forum

Episode 30

AI Ops is a natural fit in the 5G era with software-defined networks that are increasingly programmable. Human supervision of 5G networks is a bit like trying to hop onto an electric train after it has left the station. AI-assisted automation is the holy grail for network managers rattled by a catastrophe precipitated by human errors. While early trials of AI Ops in private and public 5G networks are promising, data capture in network labyrinths is forbidding due to many silos, emergent data standards, and insufficient data to train AI models. How are telcos and enterprise CIOs balancing the risk and benefits of autonomous networks steered by AI? Let’s find out!

Our guest for the podcast is Aaron Boasman-Patel, Vice President of AI & Customer Experience at TM Forum. He is responsible for defining and executing the strategic vision for all AI, manages the cross-ecosystem collaboration projects, and helps to set industry standards. 

In our discussion today, we will uncover a few things, such as: 

-        What are the early lessons from experimenting with autonomous networks?

-        What are the challenges of setting data standards for AI Operations and automation?

-        How does AI Ops benefit private 5G deployments? and

-        How do CIOs prepare for autonomous networks and avoid errors without a way to learn from past mistakes?

So, let us welcome Aaron Boasman-Patel. 



Contact PrivateLTEand5G

  • Follow us on LinkedIn at https://www.linkedin.com/company/privatelteand5g
  • Tweet at https://twitter.com/privateLTEand5G
  • For more resources on Private Cellular Networks, go to https://www.privatelteand5g.com/
  • Email us at ratika.garg@privatelteand5g.com

Ashish Jain:

AIOps is a natural fit in the 5G software defined networks that are increasingly programmable. Human supervision of 5G networks is a bit like trying to hop on an electric train after it has left the station. AI assisted automation is the holy grail for network managers rattled by a catastrophe precipitated by human errors. While early trials of AIOps in private and public 5G networks are promising, data capture in network labyrinths is forbidden due to many silos, emerging data standards, and insufficient data to train AI models. How are telcos and enterprise CIOs balancing the risk and benefit of autonomous networks tiered by AI? Let's find out.

Ashish Jain:

Hi guys, this is your host Ashish Jain, and you're listening to the Alynment podcast where we go beyond the buzzword and connect the dots between technology and its business impact. Meet my guest for today's podcast, Aaron Boasman-Patel, vice president of AI and customer experience at TM Forum. He is responsible for defining and executing the strategic vision for all AI, manages across ecosystem collaboration projects and help to set industry standards. In our discussion today, we will uncover a few things such as what are early lessons from experimenting with autonomous networks. What are the challenges of setting data standards for AI operations and automation? How does AIOps benefit private 5G deployments and how do CIOs prepare for autonomous networks and avoid errors without a way to learn from past mistakes? So let me welcome Aaron Boasman-Patel. Aaron, thank you for joining me today.

Aaron Boasman-Patel:

Thank you, Ashish, it's great to be with you.

Ashish Jain:

So tell us a little bit about TM Forum, your role at TM Forum and involvement in 5G AI automation and overall AIOps.

Aaron Boasman-Patel:

Sure, absolutely. So TM Forum is a non-for-profit member organization and we are working in really enabling CSPs to go from what we call telcos to TechCo. So it's all about digital transformation, embracing an open digital architecture, which is AI and data driven. So CSPs have the flexibility and the automation to be able to go and launch new services and to really survive and thrive in a very competitive market landscape.

Ashish Jain:

Awesome. So let's set the basics, what is AIOps? I mean, is it AI automation? Is it something more or less?

Aaron Boasman-Patel:

Yeah, there are the two definitions I think for AIOps. And the one that I really look at is how do you automate your operations? And that's putting AI at the heart of your operations today as a telco. So that's about how do you add intelligence to existing stacks and then to existing applications, but then also, how do you really go to kind of a new state of automation where we start to see things like autonomous networks really being put into place and networks and IT can become self-healing, self-fulfilling and this becomes so important and self-assuring, because when we think about this new generation of services, this 5G services, we are not going to be able to have manual operations because Ashish, quite simply manual operations cannot scale. And the whole idea of going beyond connectivity of 5G is really ensuring that you've got new solutions that can scale and really take advantage of 5G, means you want to get latency as low as sometimes one millibit per second, and also see massive amount of connectivity and devices.

Aaron Boasman-Patel:

We say there's going to be 30 billion devices on the network by 2030, that's a huge amount to try and manage. And we also have to get high liability into the networks to deliver those types of services. So you can only do that if you make your operations intelligent. And by doing that, everything has to be self-assuring, self-healing self-fulfilling.

Ashish Jain:

Okay. So telcos have run their network for God knows how many years, right? So I'm sure they have implemented automation in their network for many years. And automation is, I mean, I also come from the telco space and have seen automation done at many levels. Now you're talking about autonomous networks that is definitely self-healing, self-discovery, self-forming networks, definitely the new paradigm for many, but where do you see the challenges lie in how telcos are doing it today and where they need to be?

Aaron Boasman-Patel:

Yeah, I mean, there's many challenges here. Today, many CSPs are only automating single domains within single business units. What we have to be able to do is add automation intelligence to multiple domains across multiple business units. It's looking for that end to end automation. And of course there's different levels of AI and automation. At TM Forum, we have defined six levels of automation and CSPs need to get a better understanding, first of all, what AI is and what it isn't as opposed to things like advanced analytics and also away from rules based automation to true AI, which is obviously self-learning. And we really have to understand where is it going to make most sense to deploy it. And I think this is where we need to really understand where CSP strategies are and where these need to get to, because it's a great big gap between that, I think today in terms of where we think we are with automation and where we need to be to deliver these new types of services that I get so excited about - everything from industrial automation to healthcare, to smart cities; we've got a long way to go as an industry.

Ashish Jain:

No, I think that's absolutely right. No, but are there certain, I mean, 5G again, it has its own set of broad spectrum of things. I mean, pun intended from spectrum, but it has a lot of spectrum, but there's a lot spectrum of things within 5G. I mean, there is consumer services, there are enterprise service and in the moment you get into enterprise services that opens another can of worms because now every vertical is different. How do you deliver those services? How does automation play in that? I mean it's automation or the AIOps foundational or critical for the success of telcos or plain private 5G?

Aaron Boasman-Patel:

Yeah. Well, let me put it this very, very simply. And I hope everybody listening really understands this Ashish. It's estimated the 700 billion worth of new revenues from the 5G B2B2X space. I can tell you without investing in AI and automation, you will not be able to have any slice of that because networks will just not be able to deliver the amount of reliability. It will not be able to deliver the low latency that we really need to have for these types of services. So AI and automation is absolutely fundamental and critical because the whole thing about delivering B2B2X services technology is only one part of it. The other part is you've got to be able to deliver services at the cost point the market requires. So, you can't just keep adding more operation staff into the network. And also with 30 billion devices connected to the network, you couldn't even do that with manpower anyway. So this is what people need to really understand. We've got to have a new way of thinking about operations, not only in terms of the technology we apply, but also the operating model that sits within it. So it's absolutely critical.

Ashish Jain:

That's a big number, 7 billion, right? And a statement like, okay, you can't get any slice of it, should be intriguing for many CSPs to react. Now, let's peel the onion on that a little bit. When we say for B2B2X services, 5G B2B2X services, and automation is key, can we spell out certain aspects of things that, whether it's a traditional telco CSP, or the new generation of managed service providers that are coming up in the market to offer 5G B2B services. What are the key aspects of automation they need to start with? Or how do they plan this?

Aaron Boasman-Patel:

Yeah. And really there's got to really be, what I call, a standardized approach to planning for your AI. And that's why the forum we've even looked at to calling it the AI business canvas, which is very much like a lean business canvas, but you have to understand all the different components across the organization and what's going to be affected. If you think about AI operations, for example, the way you handle service operations is going to be very, very different, everything from instant management, event management, problem management. But really, you've got to make sure that you've got the right data first of all, to enable AI and automation. We often talk about data is the new oil and, I always say, Ashish, oil has been known about for thousands of years.

Aaron Boasman-Patel:

If you go to Texas or into the kingdom of Saudi Arabia, the people who lived there for thousands of years saw this oil, this black stuff, they didn't know what to do with it. But what they did know was that this thing existed. Now with oil, you have to refine it, right? That's the only way it's useful. You can make it into petrol. You can make it into plastics, all these other types of things. But as soon, unless you do that refinement, it's no good. The same is with data in terms of AI, that data is going to be the oxygen which gives life to all of these new services and these new operations that we've got to do. So, first of all, we have to look at the data, make sure we've got the right data, make sure we've got access to it, we can use it, we can do it safely and securely.

Aaron Boasman-Patel:

Then you've got to think about, well, how do I govern this thing? How do we enable the safe deployment of AI? How do we do it at scale and how do we reduce risk? So you've got to look right the way across the life cycle when we go into deploying AI, and these are the key considerations, do we have the right data? Is it in the right format? Can we expose those data streams in a safe, secure way? Do we have things like common data models to make integration much easier? And then do we have the services that require it? So, can we launch the services that we need to invest AI in? Because you don't want to start investing in something, you don't know what you're going to do with it.

Ashish Jain:

No, those are really, really good set of items that you listed here. So I'm going to dig on one of them at least. You said data, it needs to refinement, right? I mean, that's such a valid statement because there is variety of data, actually. If I'm creating it correctly, in my experience with telco, the data comes from so many different sources in so many different formats. So how do you, first of all, normalize it and what is a consistent way of looking at it across the board in terms of being even able to be analyzed? Is that fair?

Aaron Boasman-Patel:

Yeah. No. And I think this is the question that comes up time and time again, Ashish. I think we've got a dream as a telecom industry that will end up having one data model. I'm not sure if that's ever going to be possible. The more I dig into it, the more that I talk to people, they say, "Well, the vendors keep giving us different data models that we have to use for applications." And I don't see much change happening in that way. At TM Forum, we have a standardized data model, which relates to the SID for those people who know us and that's the information framework. And it's about though making sure that there is a common language, at least that when you start to use AI, that you can trace each and every one of those data entities and really understand what it is that you are looking at, because that's the way you're going to get lots of errors in this is if you don't really understand what the data is and how you look at it.

Aaron Boasman-Patel:

I think really huge challenge, I don't think as an industry, we are anywhere close to solving it. And I think that is something that we really have to address, particularly if we want to start working with multiple different industries, because then we're going to go from we have to multiply the amount of data models that we have to look at and that we have to be able to process on our network. And every time you do a translation, even into a common data model that you might have established, it does allow - it increases latency, but also then it increases the amount of room for error. So I would say Ashish, this is still a big warning sign that as an industry, we need to come together and learn how to handle data effectively, safely, and securely, and be able to put it in a way that we can really make the best use of it and then start to use it with partners because that data economy is going to become really, really valuable.

Ashish Jain:

No, I think you're right. I mean, another challenge I also see is, especially in this new area is 5G private networks or 5G B2B services. And going across all these verticals, I mean, cellular has traditionally not played in those environments. I mean, even if it was, it was in a very minimal sense. Is there sufficient data to begin with or are we at a Catch-22 kind of a problem right now, chicken and egg problem, where do we start to analyze it when I don't have enough data to identify a pattern?

Aaron Boasman-Patel:

Well, I think we do have enough data. I think that's a challenge. In some ways we have too much data. If you look at network data is exponentially rising. And I think that's one thing that everybody's very, very clear on. The challenge is what do you do with that data? So recent estimates and there's all sorts of estimates everywhere, but it says we only actually utilize and analyze 1% of all of network data, Ashish. Imagine what would happen if we could analyze even 5% of network data and all the data that goes across the carriers network. So I think we do have the data - it's just that we don't have access to it. And we don't have very clear data strategies. So just before the pandemic, I spent a long time going around a number of CSPs, particularly in Asia, looking at data strategies. And I was amazed at how wild they were between different CSPs. Some CSPs were collecting everything, some CSPs were collecting and storing hardly anything.

Aaron Boasman-Patel:

And I think that is the problem, it goes back again to what is that data strategy? What data do we need for those services? And then how do we analyze it and get the proper insights that we need. So I'd say actually it's not a data problem. It's a strategy problem.

Ashish Jain:

Yeah. I think that's where I was going with is, I mean, sure, they have tons and tons of data from consumer services. I mean, they've played in that market, they dominated in that market, but when it comes to, okay, I want to start auto providing 5G networks, 5G private networks for factory automation or agriculture automation, or for any reason let's say the education industry. Do I have enough information on that data strategy or do I even have enough understanding to form a data strategy to figure out what level of automation I need to have, to provide, to your third point, services I need to offer them?

Aaron Boasman-Patel:

Yeah. And I think this is something that, I mean, I think CSPs do have it. And I think with their partners, they can absolutely get it, but it is about creating a data ecosystem. And this is what is so, so important because our ecosystem, I mean, mobile network traffic by 2026, we say something 300 exabytes per month is going to be generated. It's huge amounts. It is about working with those partners, understanding what analytic capabilities you have to understand what it is you need for those types of services. So if you think about, if you go to an enterprise, Ashish, they have the knowledge, they all tell you what we know, what information and that they want to have and want to look at. Yes, you'll have to do some tracking of the data. You'll have to do some mapping of that data, which does take time.

Aaron Boasman-Patel:

But it is all there. It's understanding the strategy. What is that service that you want to leverage and that you want to unleash? So if you think about something simple, even like broadcasting of television, for example, from a staging, it's understanding what is that data that you've got to get all the way from the stadium camera or use camera back to the central studio, then to process all of that data and get it back out to TV broadcast. We would understand TV something, an example that telecom has been in for a long time. But they understand from the broadcast companies, what data they need to enable this, how they package it up and the latency and all those things around it.

Aaron Boasman-Patel:

That's what you have to do for every single use case. And that is the big shift I'd say, Ashish. You can't just say, "I'm going to launch a 5G private network," and launch 10 new enterprise services. It doesn't work like that at all. And I've just written a white paper, which addresses how to deal with customer experience in the B2B2X space with 5G. And it's exactly this. It's saying, you've got to really understand what 5G service you want to offer within that B2B2X space, and then really then work with partners, look at multi-sided business models and then apply the technology to it because it's seldom a technology problem. It is often not an understanding of the problem you're trying to solve.

Ashish Jain:

Very well said. Very well said. I think that's the part of the problem, because a lot of what I see is still a technology play, I am a data pipe provider, 5G is yet another services I can offer, here's my edge cloud services. I mean, bringing a 5G network with an edge cloud is what a lot of people are still just like, okay, "Hey, I launched my private 5G services without a context of, okay, what problem I'm trying to solve?"

Aaron Boasman-Patel:

Correct. Otherwise you're just an activity provider, Ashish, are you providing broadbands, wireless instead of wired?

Ashish Jain:

Yeah. And I think unfortunately, a lot of that is still out there. People are still treating 5G as just another great enhancement of mobile broadband and extend that enhanced mobile broadband concept to enterprises, which is in my view, a recipe to failure. It needs to be very verticalized and understanding the problem domain first.

Aaron Boasman-Patel:

It is. And you got to understand, Ashish, CSPs cannot do this alone. I can't see people going out and in hiring something like 30 medical doctors trying to understand a healthcare solution. This is where they're going to have to partner. There's many big hyperscalers who have a lot of speciality in these areas. They have a lot of healthcare or healthcare professionals that work particularly in that sector. And I think this is why partnerships is going to become critical so you can deliver that end to end service because otherwise you're just not going to be able to scale. And that's why I'm so keen on partnerships for delivering these new types of service.

Ashish Jain:

I think that's going to be key. Now let's look at, from this particular space. And I also see a different type of an activity where it's not always telco CSP driven, I know TM Forum focuses a lot on CSPs, but there is a parallel, I would say, environment getting created where enterprises are working directly with either hyperscalers or new generation of MSPs that are providing private cellular networks or private networks for their enterprise use cases, where they are not coming with a lot of such baggage, whether they're greenfield CSPs, or they are just an independent managed service provider, which are not as big as Verizons and AT&Ts of the world, but they have the telco expertise to offer an end-to-end private network services, maybe because they have played a role in managing their DAS, managing their wifi, managing some small sales in the past. And now they're able to bring in this whole ecosystem system of private 5G where the run is local, the core is local and they would still be able to provide cloud based managed services around it.

Ashish Jain:

Is the problem domain for them different than what a CSP has to be successful in the private 5G?

Aaron Boasman-Patel:

No, I think it's very similar. I think it really about understanding what that delivery is going to mean. We've seen a massive boost from those sectors really wanting to digitize everything from manufacturing, healthcare, understanding that they need to digitize themselves. I think that's the one positive side if there is one, Ashish, of the pandemic. Is that everybody's aware that they need to digitize, they need to do their business differently. Retail, for example, really went fundamentally online and you've started to see them bring in their own in-house capabilities to manage the scalability that they need. So this is why if CSPs don't wake up today to this opportunity, you will find that people in retail, healthcare, manufacturing, whatever starts to manage and look after their own networks, maybe with partners in some other aspects. So this is the risk that we have if we don't move fast enough, is that you're seeing huge, big IT connectivity departments being spun up within these own enterprises.

Ashish Jain:

No, I totally agree. I think in any, in any industry, somebody will find a way to solve a problem. And the longer you wait, the higher risk it is for you to lose that opportunity. And I think we are seeing that already. And to be honest, in this particular sector, I mean, I can't even tell you how many new companies I've seen that have emerged from nowhere in playing a role in providing private 5G services.

Aaron Boasman-Patel:

Exactly. And a good example, Ashish, sorry, but there, if you look at JCB, they're a big tractor manufacturer, make all sorts of heavy machinery, they've set up their whole ecosystem to get the machinery connected. Now, who would ever have thought that a tractor manufacturer would start developing an ecosystem, which is based on connectivity? So there's real examples out there and it just goes to show you how it can be done.

Ashish Jain:

Exactly. Exactly. So talking about examples, have you come across an example that you think would be a setting stage for many others in terms of adoption of AI automation and solving some of these challenges in a good way? I mean, you see a lot of service providers, what they're doing and what other players in industry are doing something that caught your attention.

Aaron Boasman-Patel:

Yeah. I mean, and I think unfortunately in a lot of ways, it tends to be more of these newer and greenfield operators because they don't have their legacy and they've got different needs. And also they've been spun out and born out of the enterprise space, I would say. So if you think about people like Jio in India, that's obviously Reliance Industries. So they service already a large part of their own industries, but it shows you how, if you develop a platform business, you can actually start to do that and apply it to other areas. Now they've got great success within their own reliance industries, but again, what they're probably going to be suffering with is how do they adapt it to people who aren't within the reliance group? So I'm going to be very interested to see how they can start to adapt with people who don't have the same mandates as they do.

Aaron Boasman-Patel:

And then the same can be seen with Rakuten in Japan, again, born out of an enterprise business. So they know kind of what the requirements are to work within those enterprise and platform types of businesses. We hear a lot of stories of other companies, trying to shift towards that platform models. So, I think Axiata are doing a very good job. They're setting up a platform business, or being exposed through open APIs. They've been able to do some really successful things through doing that, amazing using existing services they have, for example, hairdressers, for example, they've been able to launch an app so hairdressers can do all their booking systems through Axiata, quite a simple use case, but just shows you how, if you have a modular open architecture, which is AI driven, you can start to launch these new services quickly. And I think the likes of Vodafone and Orange as well are really starting to develop these platform type businesses ready for enterprise.

Ashish Jain:

Of course, I know. And Dish is trying to do the same thing in the U.S. Hopefully they will, at some point, if they follow what everyone else in the industry is doing. Talking about programmable networks, so definitely network programmability is key. And a lot of people, there's a lot of misunderstanding in the industry, also what that means. And a lot of time programmability, to many means, "Okay, I have exposing API for SMS, and I've been doing that for years. I mean, for 20 years, I've offered SMS services at API. I'm already programmable." What is that next level of network programmability that the CSPs need to think about now?

Aaron Boasman-Patel:

Well, that's a very exciting area and a great question because it's something that I'm working on at the moment with our members and it's about intent and intent in autonomous networks is going to become a big topic and that's critical. And what it means is, is that for any service that you have, or your stack up your network already knows what it needs to do, how to provision the service, how to make sure the service assurance surrounded the provision decision right away across that landscape, because you'd be able to say, right, for this service requires, this is the way to build for it. This is the way the amount of latency you have, this is the bandwidth that you need. And as soon as you stack up in all the different instances, it's going to be able through intent to be able to program itself. And this is what we mean by self-programmable networks, which is based on intent. And that is going to be revolutionary in terms of networks, because it means you'd be able to stack up new services in a matter of minutes, as opposed to days, weeks, months, because intent driven autonomous networks will mean you'll know what the intent is when you launch a new service and no matter what the different service scenarios are on different networks, it will be able to drive that intent and deliver it almost automatically.

Ashish Jain:

And to put an example to it, if you're trying to do predictive surveillance using computer vision, you need a different type of intent versus if you're just doing plain voice calls, you have a different intent.

Aaron Boasman-Patel:

Correct. Absolutely. And then you'll have what you call intent managers, Ashish, which will manage all of that.

Ashish Jain:

Perfect. Perfect. Now there are definitely, we see a large shift in the data landscape as we talked about earlier, the data is also being captured in terms of newer applications, which are likes of natural language processing, emotion detection, computer vision, as I just mentioned. How does the data captured on the program from the application side is used to optimize the network on an ongoing basis?

Aaron Boasman-Patel:

Yeah, I mean, well, it's going to depend where that data's captured, what part of the network it's being done on because obviously with edge computing, a lot of the data we're hoping is going to be much more nearer the device. So I think depending on the scenario will impact how the network's going to be affected by it. Sometimes we might have to do we, we can do proactive learning where it's in real time. So particularly on the edge and the device, but other times it might actually have to be looking at a network pattern within a day or a week to understand how the network is responding. I don't think yet we're at the point where we can have a single approach for this because we're going to have mesh networks.

Aaron Boasman-Patel:

We are looking at different types of cloud hybrid cloud environments, and that's all going to impact the way that data is utilized to inform what the network does. So I think we're getting to this point where we're getting much more influence for data and the network patterns and understanding, which service is going to recall, which bottlenecks may be on the network or understanding what the implications are going to be. But I think this is something where more research is going to have to be done as we genuinely launch new services. I think one of the problems we have, Ashish, is that at the minute we're doing a lot of theorizing and we're doing a lot of academic study into how this is going to affect the network. This is a bit like using synthetic data for AI deployments. Until it's running on the network with the exact network conditions, that's when the really exciting developments are going to take place with that.

Ashish Jain:

Yeah. I think it goes back to your point, I mean you need to collaborate. You can't do this in isolation and hire 30 doctors to figure it out. You have to partner with your enterprise customers and all other partners and ecosystem partners to get that data and form a data strategy. I think that's a very important point if I bring it back into your suggestion here. One of the last things I want to touch upon is the hybrid cloud. I mean, there's so much talk about hybrid cloud, edge cloud, fire cloud, this cloud, that cloud. How does automation play in that? What is hybrid cloud automation? I mean, I'm sure there's a lot of confusion out in the industry on this one.

Aaron Boasman-Patel:

No, absolutely. I think it's getting more complicated with cloud. There we all thought this cloud can be to make things easier with virtualization. And now we've got public, private, hybrid cloud, edge cloud, fog, you name it, we've got it. And it's like data lakes, streams, ponds, et cetera. It's the same thing is happening in the space. But the role of automation there will be to link all of the data sets between the different clouds and where that data is sitting in different applications. Now that has to become automated. And you're going to have to use AI to do that. Now, the reason for that is because there's going to be different legislation about different data, depending where it is stored, whether it's near the device, on the device, in the public cloud, in a private cloud, in a hybrid environment, depending where that data is stored means very different rules will be regulated.

Aaron Boasman-Patel:

It means you'll be able to do very different analytics and get very different insights. And also you're only going to be able to move and share certain amounts of data. Now that's going to have very, very wide ranging consequences, particularly if you think about national borders, international borders and the different sovereign rules that you have around that data. So that is where AI is going to become very, very important, because it will understand the rules around that data. And it will speed up the data exchange, which you can have, because as we said, for a lot of these services that we're going to launch is we've got to use real time data. So we haven't got time for everybody checking manually can this data be used in this instance? And what if something goes wrong? So AI is going to have to monitor that. And we are creating something at TM Forum, which is called an AI data governance engine, which is exactly designed to do that. It's to make data governance automated so that between those different environments that you can look at and store the data differently and utilize it.

Ashish Jain:

A lot of good stuff TM Forum is doing around this area. So that's really good to know. We're coming to an end, Aaron here. So I'm going to throw you a tricky question here to end. So let's say I throw five buzzwords at you, 5G, AI, automation, edge and private networks. How would you connect the dot between them for an enterprise IT leader?

Aaron Boasman-Patel:

Yeah. Yeah. Well, I say you need all of those. You need all of those. You're always meant to end these things easy, but never mind. No well, you need all of those things, but the one thing that drives them all together is automation. And that's what I would say. Automation is going to be based on AI. You've got to have your cloud and those hybrid environments to store and utilize your data. So it's all just different names of the next generation of technology or different parts of it. The way I would look at it, Ashish, I've just thought of this. So I hope it works, it's a bit like a fruitcake. A cake's a cake, but you've got lots of different fruits in it, and that's what to an enterprise you've got to understand is that all these different components are going to make the cake. They're going to make the network and all of them are equally important. And you've got to see though, the end to end view. If you want to make a cake, if you want to make a Victoria sponge, you need your eggs, your flour, your sugar, put it together, the right ingredients, you get the right cake. The same with these new enterprise type services, you've got to understand what service you want at the end. And then you've got to assemble the right components and right technologies to come together.

Ashish Jain:

Perfect. Perfect. I think this is really new domain for many, and I'm sure we will see a lot of exciting updates. Definitely the work you guys are doing at TM Forum, putting all these governance and standards and helping the community shift from being a traditional telco to digital providers is definitely a great, great effort. So I really applaud you guys for that. And once again, thank you so much for sharing your insights and being available for us to record this.

Aaron Boasman-Patel:

Thank you, Ashish. It's been an absolute pleasure and hopefully I'll get to come back soon and give you some updates on how the industry's doing.

Ashish Jain:

Absolutely. We need to do this again for sure.

Aaron Boasman-Patel:

Thank you very much.

Ashish Jain:

Thanks.

Ashish Jain:

Great discussion, Aaron, your assertion on the criticality of AI and automation for the success of 5G is compelling. Losing any piece of the $700 billion 5G B2B market is not worth risking. I'm sure CSPs will learn a great deal from your insights and the work TM Forum is doing to help them transform their operations. Thanks everyone for listening, please subscribe to the Alynment podcast on your favorite platform. It's A-L-Y-N-M-E-N-T. I hope you will continue the conversation by asking questions and sharing your thoughts on the autonomous networks and AI in the 5G context. Feel free to reach out to me at ashish.jain@kairospulse.com or drop me a note on my LinkedIn. Wish you all a very happy new year. Until next time, have fun. And don't forget to set your goals for 2022.