January 2021 News
Access networks are where the action is.
Contributing editor Annie Turner rounds up AI and automation highlights from January.
Deutsche Telekom (DT) and Türk Telekom have begun roll out of disaggregated broadband in their access networks, based on the SDN-Enabled Broadband Access (SEBA) reference architecture developed by the Open Networking Foundation (ONF).
The two have been heavily involved in SEBA’s development, working in the ONF Engineering Cluster in Berlin, which was set up last year. DT’s deployment was announced first and is smaller scale, rolling out disaggregated, microservices-based, fixed broadband architecture– what it calls Access 4.0 – in Stuttgart.
Access 4.0 has been years in the making for DT, but finally, as Abdurazak Mudesir, DT’s Head of Services & Platforms and Access Disaggregation, notes, “Disaggregation is now a reality. For the first time we’re producing a BNG [Broadband Network Gateway] on Whitebox hardware and are using software-defined networking technology to control that gateway.
“That’s a hugely important step toward our broadband network’s future structure. With the software-defined approach of Access 4.0 we’re driving forward automation and can implement lean processes ourselves in combination with our OSS platforms.”
That BNG functionality in question was developed by RtBrick: the company told TelecomTV that its software is now part of DT’s IP core production network
Türk Telekom’s R&D arm, Argela, working with a US-based affiliate firm, Netsia, developed the SEBA technology that it has deployed in two provinces,. The operator said it wants to reduce its reliance on overseas technology – a recurring theme as the telecom equipment supply chain fragments along geopolitical fault lines.
State support for Open RAN?
Still with access technologies, Deutsche Telekom, Orange, Telefonica and Vodafone started the year as they mean to go on regarding Open RAN. To demonstrate their individual and collective commitment, Europe’s four largest operator groups signed a memorandum of understanding (MoU).
In a joint statement, they said, “The introduction of Open RAN, virtualization and automation will enable a fundamental change in the way operators manage networks and deliver services. Operators will be able to add or shift capacity more quickly for end users, automatically resolve network incidents or provide enterprise level services on-demand for industry 4.0.”
They appealed to the European Commission and the national governments to support their endeavours. The operators want them to foster the development of the Open RAN ecosystem by funding early deployments, R&D, open test lab facilities and incentivising diversity in the supply chain.
The four argue that greater diversity will come about by lowering barriers to entry for small suppliers and startups that can use these labs to validate their open and interoperable solutions.
AI-powered RAN automation
Nokia and China Mobile announced that they have successfully completed live trials of an AI-powered radio access network (RAN) over the operator’s 4G and 5G networks.
Utilizing the Chinese carrier’s 4G and 5G networks, the companies completed an AI-based, real-time trial of user equipment to forecast bandwidth needs for network traffic in Shanghai.
They also ran a trial to detect network anomalies in Taiyuan, the capital city of Shanxi province. It explored automating network operations across more than 10,000 cells in China Mobile’s 4G/5G infrastructure. The AI/ML technology assistant automatically detected network problems more accurately and quickly than the usual methods. Nokia reckons this could reduce the need for human resources to deal with anomalies by more than 70%
During the trial, China Mobile debuted its i-wireless-intelligent and simplicity 5G network concept – a collection of technologies designed to support a greener, smarter and more efficient 5G network.
A RAN Intelligent controller (RIC) was included in the edge cloud using Nokia’s AirFrame Edge server platform which according to the vendor, “enables increased network optimisation capabilities through policy-guided, closed-loop automation.
AI pioneers feeling the benefit
A report from STL partners suggests that AI, coupled with a data-centric approach and automation, is starting to pay for operators that led in this field. For these more advanced operators, STL reckons the challenge is no longer in setting up data management platforms and systems, and identifying promising use cases for AI and automation.
The next big steps are overcoming the organisational and cultural barriers to becoming data-centric in mindset, processes and operations. In particular, a significant part of this challenge includes disseminating AI adoption and expertise in these technologies and associated skills across wider organisation, beyond a centralised AI team.
STL says the benchmark for success is not other operators, but digital- and cloud-native companies that were set up around data-centricity and practices. They include the likes of Microsoft, Google and Amazon, which are increasingly looking to telcos as partners as well as likely future competitors, plus the greenfield operators – Rakuten Mobile in Japan, Reliance Jio in India and DISH in the US.
As STL note, as well as more modern infrastructure from the start, they did not have to battle ingrained legacy processes and cultures.