Ray Le Maistre, TelecomTV (00:05):
So we're in the delightfully warm and sunny Copenhagen in Denmark for DT w Ignite 2024, the annual TM Forum event. And I'm here with Rahul Atri, who is the president of Rakuten Symphonies OSS Business Unit. Rahul, great to see you. Thanks so much for joining us.

Rahul Atri, Rakuten Symphony (00:25):
Happy to be here.

Ray Le Maistre, TelecomTV (00:27):
Now, obviously at this event, the big talk here is about how the industry can transform and transform mainly through the use of software systems as the telcos transform to become digital service providers. What are the technology and business drivers behind the recent transformations of OSS

Rahul Atri, Rakuten Symphony (00:51):
Ray? I think one, Rakuten always been an internet services company and that's how we landed up into telecom. Now, one thing which we were already clear about is in the end, the user case about the use cases and the services they get, and we carried the same DNA into OSS and automation platform. I think OSS is the automation platform, which enables the autonomous networks of what we've been talking about, especially in DTW where things are around monetization or autonomous networks. I think keeping that into mind OSS is equal to solving use cases via automation.

Ray Le Maistre, TelecomTV (01:27):
And what would you say are the main requirements for a next generation OSS?

Rahul Atri, Rakuten Symphony (01:34):
I think how we want to run the autonomous networks is how soon you are able to get the information. Is it real time, near real time? Are we able to pick it up from multiple southbound systems? How do you aggregate that and plot it across to generate insights out of it? Telco doesn't have a data problem. Telco has value out of that data. How soon you can depict those value because the sooner you get the data, the sooner you analyze the data and then you take the close to actions is what we're looking at. We've been able to do a lot of things internally, especially in Rakuten Mobile. And then taking this technology globally. We have created our own observability platform which collected data from every part of the network, whether it is infrastructure, cloud, application services, correlate them into one vertical integrator information available, then generate insights.

(02:25):
The insights could be dashboards, the insights could be alerts, the insights could be service assurance. And based on that you can even structure your teams. Do you want to still formulate that L one, L two, L three, L four triaging teams or you want to check out into more IT way of servicing your customers and say, yeah, I will look at our service assurance. Is my voltage running successfully? Is my handover successfully are my high speed train movement? Customers are feeling any challenges. All those things are possible once you have the insights available into the OSS systems.

Ray Le Maistre, TelecomTV (02:59):
And how do you think these changes in OSS as automation and AI have an impact, how that would change the customer experience management?

Rahul Atri, Rakuten Symphony (03:11):
There are multiple things to think about before customer experience as well. And it depends on for OSS system, actually everyone is customer, internal teams are customers, and then external customer of course. And it starts with what is the success? For us, the success is what are the customer using which services customers are using. So for example, in our OS systems, you can not only visualize the coverage holds, but also experience holds which kind of service is your customer experience, at what location, what amount of time, and then all the contributing systems can be correlated. So we have a platform where we correlate the coverage which was planned, the coverage which customers are feeling and the experience they're getting on every application because in the end, consumers are consumer are using applications. We just talked about before this interview started that the networks are so important and always cursed.

(04:07):
If a streaming platform is buffering, nobody will blame the streaming platform. They'll say, okay, my network is not good. And that's how we focus on the customer experience as well. And obviously internally automation is a journey never ends, but how do you define the automation? How do you define the success criteria is how we focus on, and automation is not a plug and play. It doesn't happen that I do automation in one network, take it in second network. Every OEM partner is different, every cloud is different, every application is different. But how do you create common tool set so that you can do it much more effectively, much more efficiently and much more agile is what we focus on our platforms.

Ray Le Maistre, TelecomTV (04:46):
Now, how should a network operator think about modernizing its support systems and sort of replacing its legacy systems and bringing in the new platforms it needs?

Rahul Atri, Rakuten Symphony (04:58):
I'll say something which is cliche. It's always a people platform technology kind of thing. I think we have reached to an age where technology is not the problem. AI can write the code now. So technology open source, there is so much opportunities and features and capabilities available. I think what is missing, what do we see the future mode of operations would be? What do we see the future of telecom industry would be? How do you see the customer interact with network? Would you still be selling data SIM cards or would you be selling use cases? Start from there. Define what the future customer would want and how are you going to achieve those services runtime? How can they buy drone as a service? How can they buy connectivity as a service going beyond? And that's what people do talk about when they say open network as api APIs network as a service.

(05:47):
But when you define that, then you come back and say, how do I enable technology to make this happen? Would this be a OSS plus BSS problem? Would this be orchestration problem? Would this be all problems at once? Solving together? I think what we want to achieve is what needs to be defined technology stack to enable that and also how do we change the culture and people to understand that this is where the industry is going, this is where the user is going, this is what they want. And that's why we are changing the technology shift. What we do with our customers is work on the use cases, work on understanding how the organizations are currently doing the job. If we convert that into digital workflows, for example, how would the same organization now adapt to the same tools? The successes, doing more work in less, doing less errors, doing let's say multiple rollouts at once, solving the meantime to resolution faster. I think it's very important that we define the use cases, the success criteria and how people would adapt that. I've been a product manager all my life. The basic success criteria my boss told me once is how many customers use your tool every day?

Ray Le Maistre, TelecomTV (06:52):
Now one of the things when people have really get into the conversations here at this event these days, they're often talking about intent driven end-to-end orchestration and other processes. Can you give us some examples of real world use cases of that kind of setup?

Rahul Atri, Rakuten Symphony (07:16):
Telco industry, my God is so good at creating new ves every day. I think we talk about the same thing again and again, but in different ways. I keep it simple things very simple because I said I'm a product manager. I work very close to development and engineering teams. They need to understand things so that they can code. So intent for me is what are we looking at as end results? For example, things have been moving. So earlier we used to see KPIs and say, my handover success rate is dipping or my accessibility is dipping, or my downlink rate is dipping, or my customer experience is dipping. We used to see that from KPIs. We used to see them from dashboards. Then we started creating alerts on top of it to say, give me an alert if my KPI DI by 1% or 0.5% and something.

(08:02):
And then we started creating tools like SAW and other application which will keep on monitoring this and sometimes maybe predict and tell us that this is what the problem is happening, probably this is the closed loop you have to achieve. I think if we combine all of that and say there will be a genie, there will be a tool where you can say, I want to improve my customer downlink throughput rate by two MVPs or five MVPs or 10 MVPs. Please reconfigure the network accordingly. That is what actually the intent driven network means. Do we achieve that? Is it simple? No, because the moment you say improve down link, it might impact your footprint or the coverage and other things. So it's journey towards there. I keep it very simple and think about it as a virus scanner every day if I'm a cluster head, I'm a knock head and say, please tell me my network is doing. The scanner should go around and say, okay, everything looks usual, or this is where your golden parameter template or your golden parameter audit has failed. Can you correct it? Yes, I'm correcting it right now. And the tool should be able to do that. Do I want to do it with the Gen I? Of course. Do I want it to be doing by dashboards? Doesn't matter.

Ray Le Maistre, TelecomTV (09:17):
Another thing, and some of the operators here at this event are talking very proudly about how they are going from having thousands of systems to just hundreds or even dozens in some instances, and they're bringing together what used to be thousands of disparate O-S-S-B-S-S and other IT systems into a converged platform. And that's seen as a real enabler of how they're going to behave in the future. Is there going to be a unified layer that can support all of that functionality in the future that's going to enable the cloud native digital service provider?

Rahul Atri, Rakuten Symphony (09:56):
Actually it should have been always like this. So when we started in 2018, we took a call, very simple calls to say we want to minimize the number of layers. First time I landed in Tokyo, I saw the Etsy man stack. There were like seven layers from infrastructure from VM and vm, also guest and host images, then applications, then, oh, sorry, VMs, then vfs, then applications, then VNFM, then NFVO and then os, my god, how many management systems. It's like having less people and more managers, right? So we thought, okay, how do we make it simplify how many layers we can converge? The first step was, okay, we're going to have one OSS and NFVO converge them. They can be one, the orchestrator can be the same thing. Then we went and said, why do we need VNFM? Thankfully the job was done easier by Kubernetes from OpenStax.

(10:51):
So they merged the Kubernetes architecture into the cloud platform CAS as a service itself. Then we thought about do we need EMS? Can we not replace that and make it a part of OS S itself and move FNPM up and have only telemetry agents coming in? And we did that for entire stack horizontally from ran transport core, anything application, 200 plus integrations which are running right now. So we did that unified layer and then now I think the world is shifting towards more and more. Sometimes I laugh about it. I said, we used to say that in 2018, 2019, people have started to realize, which is great for us, great for industry. It's not unification. It's actually the right thing to do for the operation because now you need less people. You don't need 200 different kind of skillset to manage each of the EMS and have the capabilities.

(11:41):
The tools are smart enough to say, okay, I'm going to converge all 200 together and give you alerts in case something goes on. You can even define services together to say this is voice as a service, this is data as a service. Also, you can have less tools and less implications because somewhere or the other you will reach out to the point of scaling as well. Resources, cloud, everything is cost. We also believe that it also led to simplification. You can spend more energy, more time. The operations would be much more faster because if something breaks, you need to have 20 people on the call. 20 people will do their regular stuff to optimize and see, okay, this is what is breaking, this is not, it becomes a lot more simpler on day two if not day zero and day one because when you're building day zero, you have inertia.

(12:25):
People would not like to move what they're doing. People would still like to see EMS screens. People would still like to see multiple screens and they want to have more number of people reporting to them. But that's how you will, and that is the only possibility where you can support the future use cases like AI because the data needs to be, knowledge needs to be collected. That's the only way you can even tomorrow do slice assurance because the slice is end to end. It won't require rancor transport to work silo only then is end-to-end slice possible. So great if this is happening and thankfully we were doing it always and we are ready for that.

Ray Le Maistre, TelecomTV (12:59):
Well, it's good that as an example for the industry to look to and it's clear that there are other operas, operators moving in that direction and starting to see the benefits of that as well. So Rahul, thanks so much for joining us today and giving us your insights from the Rakuten perspective. Thanks very much.

Rahul Atri, Rakuten Symphony (13:17):
Thank you.