Networks are being counted on more and more every day, to underpin and help realize the promise of everything from 5G to Industry 4.0 plus whatever the imagineers of the world come up with next. Through the pandemic, networks have literally kept us going — not just technologically, industrially, or commercially — but as human beings. Behind this is the IP and optical network fabric that connects it all and keeps the network that heals and protects us, the network that empowers and delights us, beating strong.
In this interview Sasa Nijemcevic, Head of Nokia IP/Optical Network Automation, gives us an overview of one of the 3 foundational pillars for service providers to master the unexpected: automation and tools. Watch this short video to learn about the benefits of network automation and how Nokia can help you implement it throughout your organization.
- Learn more about mastering the unexpected through innovation: https://www.nokia.com/blog/mastering-the-unexpected-through-innovation/
- Learn more about the IP and optical networks that have kept our world going: https://www.nokia.com/networks/networks-keep-us-going/
- Listen to a service provider’s experience: https://www.nokia.com/networks/automation/measurable-benefits/#network-automation-makes-for-a-smooth-operator
Kevin Gray: Hi, my name is Kevin Gray, publisher at Fierce Telecom, and I'm here today with Sasa Nijemcevic, VP and GM of the Network Automation unit at Nokia. Okay, Sasa, let's go ahead and get it right into it. I'm going to start off with two related questions here. Automation is obviously top of mind for every operator right now, what is Nokia doing for them, and what exactly is your role in all this, Sasa?
Sasa Nijemcevic: Well, thanks for having me, Kevin. You're absolutely right, automation has become a high priority for pretty much every operator there out, right? The webscalers have taken the lead years ago, showing the world that networks can be deployed, operated, ran in a fully automated fashion. And service providers are definitely taking that lead and looking at the ways they can improve the way they operate their networks by investing in automation. As you can imagine, Nokia is leading the network infrastructure vendor, and we provide network infrastructures to pretty much every single network on the planet. And with those network infrastructures, we provide tools to manage, control, and automate those networks.
My team is primarily focusing on automation of IP, IP optical networks, with the fulls kind of gambit of capabilities around service fulfillment, assurance, optimization, automation, up to and including 5G and 5G slicing, right? And our customer base is pretty large, over 1,000 customers around the globe, ranging from the largest [inaudible 00:01:41] service providers, pretty much every single one of them, all the way down to the small verticals, transportation, energy, public sector customers. And all of them are investing in automation, automation in one way or another, albeit sometimes with different priorities and with different needs that they're trying to fulfill through automation.
Kevin Gray: Yeah, that makes sense, and that's probably why network automation, just [inaudible 00:02:07], is actually one of the highest performing things on our websites, over the past year here. So it's obviously something [crosstalk 00:02:13]-
Sasa Nijemcevic: That's good to know.
Kevin Gray: ... looking into, right? But if you could kind of help us understand that, why is automation so important for the network?
Sasa Nijemcevic: Well, from a customer's perspective, if you are running in a network, right? There could be multiple drivers, one of them could be just responsiveness as a society, as humans we are so used to instant gratification, and we are so used to getting our services immediately, right? So that responsiveness in delivering new services, rolling them out is super important. On the other hand, we also very used to this world where it's services are always on, there's no downtime of any of the services. If Instagram goes down for an hour and the entire world knows about it.
And so, all of our operator service providers are very well aware that they need to invest in keeping those services up and automation is one way or two, to achieve it. And last but not least I would say is the cost, right? Even though it's very obvious, but cost sometimes is a very important driver for automation, and has been across many industries in the past, and it continues to be in ours. So I think when you look at it from service provider's perspective, they have no choice, but to invest in automation, if they want to stay competitive, really, right?
Kevin Gray: Yeah, and I've heard some of those benefits before, right? Like the ability to get new services out faster, and I'm just really curious though, how much an operator can actually get from automation. Are there any specific stats that help quantify this type of thing?
Sasa Nijemcevic: Yeah, and that's really interesting because years ago when we are starting on this journey, right? You can intuitively know that automation should [crosstalk 00:03:44], but if you don't have enough data, or data points to quantify, it's harder to build business cases around it. It's harder to understand what will be the return on this investment because investment is needed and so on and so forth. So it's kind of quite interesting that now years later, as we have a lot of projects under our belt, we can see real data. Analysys Mason recently did the study and I think the link is available as well, for everybody who wants to take a look at the details. Then the study was done using the real operator data from the networks in where they've implemented automation projects, focusing on basically three areas, areas of service fulfillment, network life cycle management, and assurance. And the results are quite impressive, right?
They found that the labor data... So labor time or, okay, investment in that labor has been reduced by 68% for those projects that they've studied, right? And they've also found that, let's say, for example, time to revenue, introduction of new services went down by 88%, very significant improvement, right? When you think about, if a definition and introduction of new service used to take a 100 days, now it takes 12 days, it's a very significant improvement, right? And furthermore, they then found that, say error processing, and meantime to repair, those improved by ranges of 70, 80%, right? Very, very significant improvements for sure.
Kevin Gray: Wow okay, interesting. So taking a step back, you mentioned a number of automation areas such as service fulfillment and assurance there, I think, right?
Sasa Nijemcevic: Yeah.
Kevin Gray: I guess operators can't automate everything overnight, I'm guessing. And then you start somewhere, right? So you've been in a lot of these conversations, where do you think operators should start?
Sasa Nijemcevic: Well, that's a very good question. I think it'll depend on the priority that they have in their organization, and sometimes the priorities might differ from one organization to another, even within a single single operator. For example, if their priority is increased revenue, the most natural place to start it on automating would be a service fulfillment. Although we have seen in Analysys Mason study that, area of service fulfillment is probably the one where the opportunity for savings is the least, because the investment in that area is less than in other areas. On the other hand, if the priority is on customer satisfaction, for example, the area of assurance might be really interesting area.
For example, we have one really large customer in Australia that has a vision of automating every single aspect of the way they run their network. And the first phase they've started years ago, was the investment in automating assurance, and they finished that in, I think in 2020. And since then, they've done an analysis of all of their operations teams, and tried to quantify the benefit of investing in automation of assurance. And what they found was results very similar to those of analysis based study, which is savings of around 80, 90% when it comes to alarm triage, troubleshooting, meantime to repair, identifying root causes of outages and things like that, and removing the manual mistakes that often accompany manual configurations of the network. So they've seen a great benefit of investing in assurance. On the other hand, I think Analysys Mason study has shown that, if you're looking at a cost, and cost savings, probably the best area of investment would be the life cycle management, network life cycle management, and automating those aspects.
Sometimes they're not very sexy, but it upgrades, restores... Upgrades backups, restores, roll out of new devices is very, very good opportunity for savings, right? So for example, recently I listened to an interview with one of our customers at Middle East do, who have talked about the automation projects that they're doing with us, actually. And they're focusing a lot on upgrades, for example. They brought their upgrades down from eight hours, it took to upgrade a single router, which meant for them, extended maintenance windows, even to do that one router upgrade. So they brought it on from eight hours to below one hour, which for them is a significant improvement and ability now to, in a single maintenance window, do four routers, which means that they can roll out new capabilities in the network much, much faster and so on, and so forth.
So depending on the priorities, I would suggest you may choose different areas or to start with, that said, I think the important message there is that whichever area one might chooses to start with, just starting is important. And even if it's a small increments of automation, will yield the results along the way, it doesn't have to be 100% automated in order for it to be beneficial for the organization.
Kevin Gray: I'm going to take a wild guess here and assume that Nokia provides tools for all those areas, correct?
Sasa Nijemcevic: Correct.
Kevin Gray: Can the operator just take all the solutions though, and just deploy them, and manage them on their own, do operators even have the skillset for this?
Sasa Nijemcevic: Yes and no, some of them do, most don't. As I said, we have over a 1,000 customers, and I would say, my guess would be that, probably 80% of them do not have the skills. Those are really, really hard skills to acquire, right? As we know, you need engineers who understand the networking, they understand the software, right? Because they become a programmers, because now we are talking about this DevOps NetsOps world where you're writing infrastructure as a code, you're writing software, right? So the skills are pretty high demand, but I think even before you get to that, we need to understand what kind of capabilities of the platforms that are used for automation are needed, right? Those platforms have to be highly programmable and open, you cannot expect automation to happen in the code that is released as a new software release, right?
It has to be hot plugable, it has to be intense that the delivered run time or adapters that are plugged in into the running network runtime, so that they can be quickly updated as the needs evolve, as the new service is introduced, or the new versions of the [inaudible 00:10:10] introduced, and things like that. So for that, you clearly need to have high skilled engineers in order to perform those customizations. And in our experience, the best approach is trying to industrialize as much as you can, so that the degree of customization for the specific customer environment is minimal, maybe 10, maybe 20%, but certainly not 90%, right? So the less customization you do, less skilled labor you need in order to do the customization.
Now, some of our customers are very capable at hiring those engineers and training them, and have a really strong engineering departments. And they do a lot of work themselves, and in that case, vendors like us, we help them with tools, with programmable platforms, highly capable software, but in the end, they perform the customizations. For the other end of the spectrum, lot of our customers don't have those skills and they come to us and say, "Hey, can you help us achieve these outcomes?" And then for them, we basically, they have taken approach of kind of use case catalog, where they pick from a catalog of use cases, what do they want to automate? And as they pick those use cases, we provide software, but also customizations that they need in the form of highly skilled engineers who perform those customizations for them. Right, that's the approach we taken and it seems to be working quite well.
Kevin Gray: Okay, great. So last kind of question for you here, some final thoughts I'd love to hear from you. We talked about why network automation is important, and we talked about some of the stats that back up why operators should pursue those network automation journeys, and even some of your experience from the real world there. In the end, it's all just about simplifying an in infrastructure that's becoming kind of more and more complex, is that fair to say?
Sasa Nijemcevic: It is, and when you think about it, I think it was Leonardo da Vinci who said that simplicity is the ultimate form of sophistication, right? We're talking about very sophisticated and complex environments, right? When you look at where our industry has gone, we are moving into the cloud, we are decomposing network functions, we are automating everything using open-source tools, and very, very disaggregated environments. So to put it all together, that quite a complex problem. So simplifying that for our operators is a critical part of what we do as vendors who provide automation frameworks, right? And the vision there is to have an ability to provide intent-based networking, right? To provide that network easily consumable, either by a piece of code in IT departments, and basically provide infrastructure as a code paradigm, or simple APIs that can be consumed from traditional OSS applications or homegrown tools, right?
That their intents are defined and expressed at the higher level, and then realization of those intents in the network is taken care of by the controllers like ours, right? And ultimately, even in some cases, they require simple goodies and dashboards, and point-and-click applications that can help them achieve that. But in all of those cases, we want to get away from this kind of imperative writing code for it, or CLI commands for every single configuration, more into templated intent-driven world, which can be spending multiple vendors, multiple networking technologies or topologies, and obstructing those for whoever is consuming the network. Right, that's the ultimate goal, and of course, as a vendor or network automation team, that's our mission to help our customers along that journey, really.
Kevin Gray: Okay, and that is a great way to wrap things up. We are out of time. Sasa, thanks so much for joining us.
Sasa Nijemcevic: Thanks for having me, Kevin. Thanks lot.