Network automation? It’s all about the data

• Network operators are desperately trying to automate all manner of workflows to become more efficient
• But it’s a massive challenge because any automation relies on data flows that need significant management and testing
• Telefónica’s Diego Lopez identified some of the key challenges and ongoing automation projects during the recent DSP Leaders World Forum

WINDSOR, UK – DSP Leaders World Forum 2024 – Network automation is a goal that has been high on telco agendas for way more than a decade, having gained a new lease of life when virtualisation and software-defined networking (SDN) became a serious consideration for telcos, but meaningful automation is very tough to achieve and relies on a thorough data flow and management strategy that the operator community is still getting to grips with, noted Telefónica’s senior technology expert, Diego Lopez, during the recent DSP Leaders World Forum event in Windsor, UK. 

Lopez was the co-host for the Improving network optimisation through automation session, which he kickstarted with an overview of how the giant Spanish telco is moving towards the automation of its network operations. He stressed that the key consideration for telcos is “how we guarantee sustained and trustworthy data flows that can be properly interpreted by the control elements” that aggregate the data, then how the resulting actions safely reach the network elements in a way that can be accurately interpreted and used. 

Essentially, the secret sauce to automation is all about data flows and how exchangeable that data is between systems, he explained, noting that it’s always critical to consider the bi-directional nature of these flows as well as figuring out how to make sense of all of the data in all its different forms. “It is not all raw telemetry data… it could be an exchange of knowledge, an exchange of models themselves, of parameters,” noted Lopez. 

Of course, no discussion about automation is complete without considering the impact of AI. In regard to what he termed as “the current trend towards the use of smart controllers, AI controllers [and] generative AI,” Lopez noted that he and his team that are very focused on “how to control the essential properties related to authenticating the data and the sources of the data, the authorisation of who can use the data and for what.” Just as important is knowing “who used the data, what the data was used for… [and then] being able to analyse why something happened, why a decision was taken and tracing it back” to the original sources and rules that had been set. 

All of this requires extensive validation, evaluation and debugging, something that the Telefónica team has been looking at since the early days of 5G developments, noted Lopez. His team are currently engaged in a project focused on data provenance and access control and another called Horse (holistic, omnipresent, resilient services) that is defining ‘what if’ scenarios “so we can evaluate in advance what could be the impact of a particular automated decision on network performance and characteristics,” he explained.

Many might call that a digital twin project, but Lopez prefers the term “synthetic environment”.

Lopez went on to discuss various aspects of automation and optimisation with his fellow session speakers from the NGMN Alliance, NTT Docomo, Rakuten Symphony, the Telecom Infra Project (TIP) and Wind River. Those discussions included a focus on various use cases (network congestion and energy consumption), how to measure the effectiveness of automation – “effectiveness data is a very important flow” noted Lopez – and the applicability of automation and data management models across different geographies that are governed by different legislation and regulations. 

You can view the whole session on demand here. 

- Ray Le Maistre, Editorial Director, TelecomTV