Gap between IT and network skills remains challenge for autonomous networks

During a recent TM Forum autonomous network (AN) workshop, hosted by Telefónica, communications service providers (CSPs) outlined how they are overcoming obstacles as they progress towards Level 4 AN.

One of the chief hurdles remains managing people through large scale change.

“Our main challenge is embracing new ways of working, which require continuous process adaptation,” said Nilmar Seccomandi, Director of Autonomous Network and Infrastructure, Telefónica.

“We are talking about a huge network transformation … touching all areas … of the networks,” he added. “We are talking about technology transformation, the data, our processes and the people to organize the program.”.

After all, as several of the 70 participants in the three-day event in Madrid pointed out, autonomous networks alter how IT, data engineering, data science and network teams collaborate with each other. And it isn’t always easy to align them efficiently and identify who has the right expertise for a given task.

Network know-how

Mojdeh Amani, Mobile OSS Strategy Specialist, BT, pointed out that there are challenges to closing the gap between IT and network skillset. Software developers, data scientists, data engineers are all critical skills, and telcos need to know how and where to deploy them to achieve the best outcome.

For example, software developers don’t necessarily know what handover is, Amani explained. Yet network know-how is critical because “we are a heavily regulated industry, and it’s different from IT and from the cloud,” she added.

BT’s AN ambition aims to deliver capex and opex reduction; enable revenue generation; and meet regulatory requirements. One aspect of regulation BT hopes to address with AN is compliance with the UK’s Telecoms Security Act. The act includes a requirement to report manual network configuration, which would benefit from automation, explained Amani.

Like other operators, BT has weighed up different approaches to optimizing internal collaboration. After considerable trial and error with external partners, the telco ended up breaking down programs into small, manageable elements so that it could focus on identifying and executing what is achievable from a human and technology perspective.

One of BT’s notable successes to date has been advancing automation across its CI/CD pipeline.

“This is our highest priority, and we are achieving a very good result,” said Amani, adding that the company expects to complete the process by the end of 2025.

Pick your battles

Telefónica, meanwhile, aims to achieve an average of AN Level 4 across all domains and processes by 2030.

The operator runs its global AN program from an AI center of excellence, which develops use cases and guidelines for network teams to use across Telefónica’s’ international footprint. The AN teams have board-level support, with autonomy levels presented as “a KPI that the board … follows in the same level of revenues,” said Seccomandi.

Nonetheless, Telefónica is not pursuing AN Level 4 for its own sake. Its choices have to make business sense, which means it will not achieve Level 4 across the board.

“We know that in some areas we will not achieve Level 4 in 2030; we will achieve 3.8, 3.7 but in other areas we will achieve 4.2, or 4.3,” according to Seccomandi.

In some instances, for example, cost represents a barrier. Seccomandi believes industry collaboration on creating use cases could help lower implementation costs.

“It takes money to implement new use cases. I mentioned we have 400 use cases implemented. We are not counting how many use cases we didn’t implement because of cost,” explained Seccomandi. “We see cases where the payback is 20 years, 30 years. We have to move as an industry to deploy some kind of use case agents or other techniques … to reduce the price of implementation.”

Afnan Ahmed, Director Network Strategy & Architecture, Telenor, also emphasized the importance of achieving measurable returns for the business.

“We have a value-based approach. For this reason, “We started with RAN and assurance. It's not that it's the only thing,” added Ahmed. “But it's where the most OpEx and Capex goes for most operators.

In addition, “we feel all operators to ask the question on why they want to implement autonomy in the network or what problem they are trying to solve,” said Ahmed. “Our value based approach is cross dimensional - there is an efficiency angle to it but equally important is how we can use it to improve customer experience of our products, how we can generate more revenue through new services as well as how we can impove the ARPU of existing services.”

The potential for new revenue generation is another factor in the business calculation. Orange’s SVP Operations and Networks Economics, Emmanuel Chautard, described several ways in which network automation supports the company’s ongoing build out of a horizontal, disaggregated telco cloud network, including when it comes to orchestration. But he also noted that creating more autonomous networks could help operators monetize the 5G SA network core. “We see some interesting perspectives, looking at Asian markets, with quality on demand, with more dynamic management of connectivity characteristics, … ensur [ing] SLAs,” said Chautard, adding that: “To monetize this is a way to get out of the commodity market.”

Standardizing an open ecosystem

Several attendees also highlighted the importance of standards to simplify how CSPs work with partners within the AN ecosystem, including systems integrators and vendors.

CSPs pointed out it would be useful if vendors could demonstrate what level of network autonomy different products and solutions would help them achieve. They would also like to build on the openness established by TM Forum’s ODA.

“One of the key assets that TM Forum brought to the ecosystem is the Open Digital Architecture,” said BT’s Amani. “We need that openness across the ecosystem. And now we have better foundation to move along to our autonomous journey.”