SONIC is a unique program of activity aimed at accelerating open-based innovation to diversify the UK telecom supply chain. How’s it doing? 

in March 2021, we covered the launch of SONIC, the UK’s government-backed initiative to grease the wheels of the Open RAN supply chain. Today, SONIC provided an update on its progress.

To recap (and that’s important, because judging by the questions on the live chat, expectations of SONIC are already running far beyond its well-articulated remit – more on that later):

  1. Establish a facility that accelerates learning about the practical challenges of re-aggregating a disaggregated RAN.
  2. Help pre-commercial products/vendors gain early sight of downstream challenges in integration and end-to-end (and live) operation – thereby mitigating later integration risk.
  3. Feed the learnings back into the industry to help standards refinement.
  4. Repeat the process using alternate vendors.

The June 2021 update shared the news that item 1 is pretty much complete: in just three months, SONIC has built seven sites around the UK, constituting three representative – but fully operational – RANs developed integrated based on the available open standards.

It’s a sophisticated setup. The three radio access networks (“chains”) can use one of four separate core 5G networks (though for now it’s the Druid + Radisys 5G Core combo that’s in use). There are three different configurations of RU, DU and CU vendors:

  • DUs from Mavenir, Phluido and Effnet, and Radisys.
  • CUs from Accelleran, Mavenir (with added RIC – for good measure) and Radisys.
  • RUs from Foxconn and Benetel.

There are indoor and outdoor setups. They’re using standard 5G handsets for at least some the initial tests. And they have created some cool augmented reality apps to provide some real-world context.

SONIC’s technical lead, Dritan Kaleshi, highlighted the key (technical) challenges in Open RAN:

Broadly speaking, SONIC is working from left to right on this chart (there’s little point setting out to benchmark performance if the component pieces won’t fit together). To be more precise, SONIC is as much about testing the standards, as about testing component pieces against the standards.

It’s clear there are some initial learning points. Integration of Open RAN components was not quite “plug and play” (hardly a surprise – but no kind of show-stopper either – SONIC Labs is first to integrate these particular combinations of components); some aspects of standards that made sense on paper were trickier to realise in practice; general purpose compute works “to a point”. On the plus side, CU integration with the 5G Core was unproblematic. SONIC says that more detailed findings will be released in October.

Reasons to be Cheerful

Panelists from Mavenir, Radisys and Accelleran had plenty to be cheery about. For one thing, SONIC Labs is proving what they all already know to be true: that the future is open. It works, as a technology but also as an approach. There’s enough proof already that from a standing start, RAN components based on the available open RAN standards can be integrated into an operational solution relatively quickly.

The next target is swappability – a further test of the O-RAN standards, as well as of their interpretation by different vendors.

There was more than a hint that the SONIC may well be spending more time on enterprises and private networks, which is certainly of interest to all of the vendors involved. Despite a question on it, SONIC does not really need CSPs’ involvement, it’s not “that kind” of testbed. The value is not so much for CSPs, as for vendors looking to accelerate the path of innovation, from in-development to market-ready. The ultimate benefit that SONIC is looking for is growth in the pool of technically viable vendors for UK telecom.

And that’s what makes it different. It’s not about supporting a “Made in Britain” agenda. It’s not about picking winners. It’s not about proving blue-sky examples of 5G applications (there are other groups doing that). It’s not about pushing the envelope of functionality, in a way that vendor labs more typically are. Sure, SONIC is government-backed, but it isn’t responsible for government policy. And it’s not responsible for telling operators what to buy. It has no plans to get into the open RAN certification business. Its remit is to add to the pool of knowledge of possible options – for vendors, for potential collaborators, and yes for CSPs, regulators and governments too. And that can’t be a bad thing.

Is there work still to do? Of course there is! Just as there is in 5G, and 6G and application development, and cloud infrastructure, and handsets and devices and radios and a hundred other never-ending fields. But in just a few months, SONIC and its first cohort of Open RAN vendors have taken a huge step forward. The telecom industry as a whole will benefit from its experiences and findings.

More on SONIC here.

For further analysis of Open RAN and how it is reshaping the telecom market see Appledore Research on Open RAN.