The UK's open RAN plan is a joke

Two leadership changes since the blond buffoon assumed power haven't sapped the UK government's bizarre enthusiasm for open radio access network (RAN) technology. By 2030, 35% of UK network traffic will pass through a mobile site built by multiple suppliers instead of one, it hopes. It's like having an ambition for British consumers to buy mobile separately from broadband rather than in a bundle offered by the same provider.

Rishi Sunak, the latest prime minister, is evidently not prepared to bet any serious money in this area, though. A government that spends more than a trillion pounds a year, according to its own figures, and that is flinging tens of billions at a new railway line, has dug into its back pocket and – amid all the fluff and detritus – found a loose £88 million ($110 million) for open RAN research and development. Ericsson and Nokia, which collectively spent $9.1 billion last year, must be quaking in their Nordic boots.

From a government perspective, open RAN is about broadening the supplier options beyond a few big players. Chinese vendors are banned, or will be a few years from now. The idea of a Nordic duopoly seemingly holds little appeal. The idea of cultivating homegrown alternatives holds lots. Open RAN is supposed to facilitate joinery between different suppliers. Without it, operators have had to buy all the parts from one vendor. With it, specialists could have a role.

But the effort so far looks pathetic, whatever supporters say. It's been more than three years since the government said it would ban Huawei and began preaching about open RAN. Today, the percentage of UK mobile traffic on open RAN technology is roughly zero. The only UK telco half-serious about open RAN is Vodafone, and Vodafone's current expectation is that only about 14% of its mobile sites will be open RAN ones by the start of 2028. Most of its Huawei estate, it turns out, is being shifted to Ericsson.

Local suppliers are nowhere to be seen – at least, there is none that could be a realistic substitute for Ericsson and Nokia in helping the UK arrive at the government's 35% traffic ambition. Vodafone's planned deployment of open RAN at about 2,500 sites involves not a single British firm. Its latest published list of vendors names Samsung (South Korea), Intel (US), Keysight (US), Dell (US), Capgemini (France) and Wind River (US). Samsung, the most important, is an electronics giant that was active in 4G and 5G long before open RAN became a popular cry.

No Marconi, plenty of baloney

In the absence of a modern-day Marconi, authorities have actively cultivated links with Asian firms such as NEC and Rakuten Symphony. Yet neither of those Japanese companies has much to show for it. NEC featured on Vodafone's original list of suppliers more than two years ago but had the ignominy of being the only one omitted from the latest update. Rakuten has no deals with major UK telcos – at least, none it has publicly announced, and it is not typically reticent about its commercial progress.

Unless competition eventually pushes prices down, open RAN per se (as opposed to relatives such as virtual and cloud RAN) will cost no less than traditional RAN. And introducing additional vendors into the mix usually made things more expensive, said Neil McRae, chief network strategist for Juniper Networks, when he was previously chief architect for BT. It shouldn't come as a surprise. It's the reason consumers spend less when they buy mobile and broadband from the same company instead of taking each from a separate supplier.

If specializing were such a good idea, why are the specialists evolving? Having started off as a software company, Mavenir now pitches itself as hardware vendor and systems integrator, too. Fujitsu and NEC originally focused on radio units (RUs) but have since added RAN software to their portfolios. NEC has branched into systems integration. Rakuten has done the same. It gives rise to the oxymoronic phenomenon of single-vendor open RAN, where an operator invests in interoperable products that all come from the same supplier.

Vodafone is essentially doing this with Samsung, which is to provide both RUs and server-based software, the most important RAN components. One of those could theoretically be swapped out for another vendor in future, assuming open interfaces are baked in. But this would be expensive unless it happened when the products were due for replacement. When that day comes, Vodafone will naturally look for the best deal around. Vendors offering RUs and server-based software as a package will be hard to beat.

MIMO jumbo

Technical difficulties are also a factor. Open RAN interfaces don't alleviate the pain of integration. What's more, the current specification for connecting RUs to distributed units (DUs) hosting server software is not ideal for massive MIMO, an antenna-rich 5G technology now being widely deployed. That, at least, is the assessment of numerous vendors along with telcos such as AT&T and Orange.

Approved optional fixes come in two flavors. One means doing channel estimation, an important function, in both the RU and the DU. In the original spec, it was limited to the DU. But replication, while addressing performance issues, could lead to interoperability problems if the RU vendor is different from the DU vendor and their algorithms do not align, according to various critics.

The alternative would shift another function called the equalizer – there to address interference – into the RU. But an awkward compromise means DU makers must include equalizers in their products, even if they are not in use, to comply with the specs. This guarantees the DUs will also work with the RUs that do channel estimation but don't include equalizers. Yet it seems bound to make this option more expensive than a traditional deployment.

Moody's blues

Thanks to the single-vendor open RAN phenomenon, the UK might just realize its 35% traffic ambition by 2030, but not in the way it intended. If specialists disappear, and operators decide to buy most products from one company, the challengers to Ericsson and Nokia will lose the competitive leg-up open RAN was supposed to give them, especially if the Nordic vendors also promise to be open RAN-compliant.

In a stagnant sector prone to consolidation, it is hard to see how smaller companies can succeed on these terms. After portfolio expansion, Mavenir is putting huge pressure on cash flow by investing more than 50% of its sales in research and development, according to Moody's, a ratings agency. That figure is up from 21% in 2019, according to a previous Mavenir filing with the US Securities and Exchange Commission. Mavenir's 2022 sales were more than $500 million, it claimed early this year. Ericsson's were $24.3 billion.

The likeliest outcome is that telcos continue to buy products from the Nordic giants or switch partly to big Asian vendors such as Samsung. It's notable that Ericsson's RAN market share was up from 33% in 2017 to 39% at the end of last year despite the professed telco interest in open RAN alternatives. Nokia, after earlier 5G stumbles, also now claims to be gaining share. In Germany, it was recently described as "our partner" on open RAN by Timotheus Höttges, the boss of Deutsche Telekom, Europe's biggest telco. So much for open RAN spurring diversity.