Swisscom makes business sense of network disaggregation

Swisscom is a relatively small telco, but when it comes to network innovation it can punch above its weight. In 2022, for example, it was among the first operators in the world to successfully test 50G-PON fiber-optic tech­nol­ogy.

It has also been investing in network automation, which underpins Swisscom’s plans to develop a network-as-a-service (NaaS) business and deliver capabilities such as basic connectivity or IP VPNs via API interfaces.

Examples of how Swisscom has gained value from automation include enabling cus­tomers to set up their own network configurations, helped by APIs, and the development of its own service management platform, which Swisscom claims can make a 5G network available in just 43 minutes.

Yet the road to automation is arduous, and for Swisscom one of the necessary, if potentially complex, parts of the route is disaggregation.

“We strongly believe that it's easier to automate and digitize your entire IT processes if you have disaggregation. So, it's a key element of our strategy,” according to Rudolf Strijkers, Enterprise & Security Architect at Swisscom, speaking at FutureNet in London in May, and separately in an interview with TM Forum’s Inform.

Breaking up what you have and reconfiguring it to increase value, however, can unearth surprises.

Swisscom started disaggregating its network systems back in 2018. It quickly notched up success with its mobile systems. But wireline was a different matter. By 2020 the company realized when it came to fixed network operations “introducing agile in wireline has provided valuable learning opportunities that showed us where we could do better,” says Strijkers.

The company then dedicated “over a year of very thorough discussions with our network development and operations team to understand the challenges encountered during our agile reorganization.”

It turned out the way IT was structured hampered agile working.

“Both teams said if we work on one tool chain, we have one CI/CD pipeline in which we both work, then we can create new features and use the same tools to continuously improve the network,” explains Strijkers.

This led to the adoption of NetDevOps, which combines network development and operations in a domain driven IT concept, in which teams gain full control and autonomy over the IT stack in their domain

Then they had to answer the question of “How do you cut up your current IT landscape into domains and then combine the parts so that something useful comes out?,” explains Strijkers.

Ruler of your own domain

Exchanging experience with TM Forum members, including Deutsche Telekom, confirmed to Strijkers that network “domains need to make business sense and the teams must be in control. It's their own little company. They have to decide what they do with their money, how they operate. Within your domain, you control your IT architecture, you determine what's available, what to prioritize, you have your own backlog.”

The domains, meanwhile, communicate with each other via standardized APIs.

Creating domains that operate as small businesses has empowered the combined network development and operations teams, but like any organization they still operate within parameters.

IT choices, for example, are still governed by the need to control costs and prevent the proliferation of IT silos.

“There's a high degree of autonomy but still, if you want to introduce a tool … it cannot be completely free, that has to be managed to avoid duplications and redundancies,” says Strijkers. Ideally, we shouldn’t force IT solutions; it must be easier and more convenient to consume IT as a service, than to build it yourself in your domain. Here the developer experience of IT as a Service is a key element to provide a real alternative."

And when it comes to sharing data with the rest of the company “we want to make sure that the data that goes out and is published in the company has a certain quality,” according to Strijkers.

Other areas where boundaries are set include service modeling, where “it needs to be explicit what what we offer and how resources, services, and products are interrelated.”

Architecting automation

All of which requires a strong enterprise architecture. Here Swisscom was helped by TM Forum assets.

“Our biggest challenge is architecture consistency and there we really aligned with the industry standards; we follow the TM Forum’s NaaS and ODA (Open Digital Architecture) and adapt it to our specific context. We call it Swisscom Digital Architecture and it greatly simplifies the orientation, planning, and execution of our transformation initiatives.” says Strijkers.

In NaaS, “There's a lot of work that has been done by Telstra and also other operators that have been defining this architecture framework [through TM Forum],” he explains. “We want to align to industry standards; it is key if you want to do business model unbundling.”

Swisscom is benefiting from changes to its working practices; for example, it is now running a fully automated CI/CD pipeline for its IP networks.

Crucially for employees and customers alike, Strijkers believes that standardizing the entire infrastructure “will give a dramatic increase in quality – so much that we claim that we will have 80% of the changes done during working hours.” The company also expects the mean time to repair to fall by 50%.

And whereas in the past operations teams may have had to wait six months for an OSS feature to be implemented, their goal is to make it happen within a two-week sprint.

“It is costly to build these stages. But on the other hand when you really need it, you're glad that you have it. So, we are really investing in disaggregation,” explains Strijkers. “If you think about the future [and the] need to integrate with other companies … then you need to do this stuff. And it is better to prepare for it now because you can only spend the money once. Each life cycle is an opportunity to move into the new world.“