Telcos see API future after “code red” warning

Telcos go in on APIs and AI at DTW23.

In the morning on the first day at Ignite DTW23 there was a large event  map on a wall near the main entrance, offering a navigation guide to the event’s many stages and areas. By mid morning it had been removed. An example, perhaps, of telco failing fast and iterating quickly. Or was the map just wrong? Either way – not the best of omens for an industry body that is intent on guiding telcos out of the pickle they have found themselves in.

Well hold on. Are telcos in a pickle? If you accept the analysis, as laid out by TM Forum CEO Nik Willetts, it’s this: telcos invest billions and billions and they don’t get anything like enough money in return. Less than 1% return on capital, according to Willetts. Accordingly, nobody wants to invest in telcos, meaning telcos stagnate and creating a vicious circle of under-investment and being out-performed.

This isn’t a new problem – telcos have been told they’re “missing out” on revenues since the first ring tone was downloaded.

The problem is, the assumed native telco stategic advantages – network connectivity, knowing who their customers are and what they’re doing, in many cases literally giving you and branding the device that’s in your hands – are the same advantages have had since 2G.

It wasn’t enough to stop them, by their own portrayal, missing out on opportunities that could have and should have been theirs. Take for example, said Scott Petty, Vodafone Group CTO, the way Twilio had taken control of enterprise messaging by aggregating network APIs and offering that to its customers, “decimating SMS” for operators. If we don’t build that abstraction layer for the network, Petty said, someone else will.

So what’s different this time, and in fact what is particularly “this time” about this time? Well, according to TM Forum Nik Willetts, the “this time” is that we’re now at a juncture where Generative AI, automation and cloud can come together to give operators a path through to a future where they do more than cart packets around for their Silicon Valley and hyperscaler betters.

Declaring a “Code Red” for telcos, Willetts said that operators have a three year window of opportunity to develop the capabilities to win out in this AI-led, automated and API-d world, carving a brighter future for themselves.

The TM Forum’s ODA (Open Digital Architecture) gives them the, well, architecture. The TM Forum’s APIs give them the ability to offer and charge for network assets and capabilities, turning their networks into platforms, their data into dollars. The will to do this must come from the telcos.

BT’s 100% in on public cloud (for data)

Many of the telcos at the event expressed the requisite will. BT’s CDIO Harmeen Mehta said that even though the company had promised it would generate half a billion in value from AI, actually it is sure it will get closer to a further billion. GenAI, combined with diversity of thought and hiring, will unlock opportunities from internal operations to external offers and customer experiences.

Mehta said that BT manages 30 Petabytes of data, around two billion malicious attacks a month on the network, and sees a peak of almost 30 terrabytes per second on the network. The operator processes 10 million data points just to manage its location information.

“We took a stance and we want to move 100% to the public cloud. We’re not going to be only dabbling. We’ve now managed 75% of our data onto the public cloud with Google.”

Mehta said that it is AI that has enabled BT to begin to leverage the power of its data.

“When we made the decision that we would forget about all this fragmented data across 2,500 systems that we had and consolidate everything in place with the partnership with Google, little did we know those were decisions that would unlock the next decade for our organisation.

As one small example Mehta said a new app called Sweeper is bringing £20 million of value by allowing BT Openreach engineers to managed how they lay fibre more efficiently. It took six weeks to build. Intelligent pairing at call centres can use AI to match the customer to the best possible agent for that customer.

Telefonica’s in CAMARA

Telefonica’s CTIO Enrique Blanco said that Telefonica has been learning from the IT world how to go to cloud, and is committed to that journey. Secondly, it needs to move with velocity. Thirdly is must assess how it can monetise its network, exposing it using APIs. Finally, it must guarantee how it can expose TM Forum APIs through CAMARA APIs.

Blanco said that Telefonica designed an autonomous network journey three years ago, to manage and orchestrate the evolution of the network. To do that it needed new capabilities in OSS.

“The traditional OSS doesn’t work. We need to build massively new OSS that are amaging these new software instances. And the answer to this is AI. We are building use cases for OSS using AI capabilities and trying to ensure we are doing that in some common way, it is extremely relevant that the TM Forum helps the industry to decide what it means to build the new paradigm.

“We are building the next three year plan in the budget and we are comfortable how we are doing this. We need to guarantee that the APIs managing the OSS are fully compatible with the GSMA CAMARA. We are building the pillar of the next services generation.”

This time is different. Now we think we have the assets in interactions with developers

Vodafone’s abstraction abstract

Vodafone’s CTIO Scott Petty also backed the TM Forum API and the GSMA CAMARA programmes.

“The most value that we can unlock is in out networks. The new services that we can build is based on our networks. And if we unlock that capability we can truly transform our organisations,” he said.

Petty said that Vodafone knows it must optimise its operations internally.

“It’s too expensive to run our infrastructure and we can’t commercialise that into value. In networks, we’ve really focused on how can we leverage digital technologies, particularly our data ocean platform, machine learning and AI to transform the way we run our networks.”

One example he gave was a capability called Smart Capex.

“We used to do our network planning based on technical parameters. Probably the same way everyone else does it. Where’s the network congestion? How’s it running? That’s totally disconnected from the customer. You could build capacity, but you don’t know if you’re going to earn revenue for it. You don’t know who those customers are. You don’t know how they’re moving around. We built an AI ingestion platform working with Facebook analytics, working with retail operators to take all of that customer data and build it into our platform so we can make more intelligent decisions around how we built our sites. We were able to fund a lot of our 5G site deployment based on planning where we could do 4G offload faster.”

A second example was taking all of its OSS and network probes and integrating them into a data ocean in Google Cloud. Vodafone reduced 120 tools with a set of AI and ML tools sitting on top of the data, building its own presentation layers.

“Once you’ve got that aggregated layer you can start to manage your network much more efficiently as you can create the observability that you need for the upstream applications to leverage that.

“We’ve been able to improve customer service and reduce churn, which is just starting to fund what we really want to do which is build new sources of revenue.”

Petty gave IoT as an area has been able to drive revenues by operating as a platform.

The company has 170 million IoT SIMs connected to its platform, making a billion Euros of revenue.

“Our IoT platform is an aggregation layer. It hides the complexity of the underlying networks for people who want to embed IoT into their applications and have them work no matter where they turn up in the world.”

Petty said there were four steps telcos need to take to build that abstraction capability: a network platform (TAS) in the cloud.

“Why do you need that? Because you have a whole bunch of developers that are writing code, doing CI/CD pipelines producing infrastructure, and they’re putting it somewhere. It might be in our private cloud on VMware. It might be on AWS, it might be in GCP. It might be in Azure, and I want to use all four of those. But I want it to be consistent. I want it to be automated. I want all of my observability tools to be built into that. I want to make sure that anyone writing a single piece of code no matter where it is in the technology stack is leveraging that capability. And I can use that to expose services to the external world. Once I have those platforms in place, I can start to build tenants.

“It’s a platform that you can start to expose all of those capabilities in APIs that you can leverage in a consistent way because you built a single cloud platform that sets the consistency of how you deploy an app. That is going to be critical to decide whether you’re able to launch services across multiple markets or make multiple capabilities to enable the services that we talked about.”

The third part of the capability is what an operator does in the core, layering observability and orchestration into the upstream applications.

“Some piece of the lower level infrastructure is about to fail, do something about it, move your node somewhere else. We don’t do that in networks, it just fails. Or we tell an OSS tool that something’s going to happen. If we’re going to be successful with 5G SA applications, we need to think about observability and supporting applications that sit above that.”

Finally, Petty said that achieving this is “all about culture” – bringing together network and IT teams. He said that he “really believes” that integrating TM Forum ODA APIs into the CAMARA and other network APIs is the future.

“Taking an ODA API, using that to hold a payload for a network API, in this case a CAMRA API, and making those dual stream to manage and provision services. I’m really excited about the opportunity this will create for us to bring these two layers together to unlock new services.”

Petty concluded with a warning that if operators don’t do this the somebody else will.

“There are opportunities for us in 5G SA but only if we achieve this. If we do it country by country, and we don’t build the abstraction layer, then somebody else will. You can see it today, Twilio totally decimated our SMS messaging markets. The same thing will happen in SA unless we can step up and run some of these services.”

AT&T: “this time is different”

Igal Elbaz – SVP of Technology and Network Services, Network CTO AT&T, had a similar message. He described a network that sits after a decade of investment to be open, software defined and programmable, that already acts as a “platform for innovation.” 60% of its traffic is running on an open, disaggregated network platform, he claimed.

“We were so good at writing software for an automated cloud core that Microsoft bought it,” he said. That may not be, to be fair, exactly how everyone saw that sale. How is ONAP doing?

But scepticism aside, Igbal was locked in on the benefits of creating a programmable, software defined network, exposed by APIs to enable customer and partner access to security, connectivity and other capabilities. We have a software-based multi-layer machine learning controller, we have machine learning traffic engineering.

“You take that platform, you put a software layer on top of this an expose this though APIs you can start to introduce the concept of network as code and this time around we are confident that this time is going to be different.”

“This time is different. Now we think we have the assets in interactions with developers.

“APIs are not new for us. We have a couple of thousand APIs with a billion API calls every month. MVNOs are integrating to our network through APIs. However for the majority of these are system wide. I think the opportunity we have is to take the knowledge we have in APIs and put it on an open programmable network which can help us think about new business models that are beyond our traditional models.

“Think about network embedded X – security, healthcare, transportation. I think we are very uniquely positioned in terms of the assets, the scale, the network, the openness, the knowledge that we have in cloud and software to get us to the point that we can actually see the value of monetisation and how we can tap into that.”