Let's stop pretending open RAN is in good healthLet's stop pretending open RAN is in good health

Open radio access network (RAN) fanatics are starting to sound like erstwhile Iraqi propagandist Comical Ali, vehement about the regime's good health even as the rumble of the opposition's tanks grows louder. The concept was supposed to produce alternatives to the dominant kit vendors of Ericsson, Huawei and Nokia. Instead, the trio looks as dominant as ever. Stalwarts would rather insist all is well than admit the battle is being lost.

They shall remain nameless in this story, whose purpose is not schadenfreude. Despite being called a "serial open RAN critic" – where critic is obviously used in the sense of detractor – this telecom journalist has no reason to celebrate open RAN's failure. In contrast, if it leads to better networks, lower costs and a more competitive market that is sustainable, it should clearly be encouraged.

What helps no one, though, is the pretense that open RAN is succeeding when it has so visibly crashed or gone rogue. More than six years since the creation of the O-RAN Alliance, the telco-led group developing specifications, no small vendor or specialist has a meaningful role in one of the recognized "brownfield" networks that make up nearly all of today's telecom market. Some would argue that open RAN has been co-opted by the big vendors whose power it was supposed to contain.

The most important changes since 2018, fueled by geopolitics and economic realities, could and would have happened without the O-RAN Alliance. They include the dimming of Huawei's star above European skies as governments moved to suppress China and its interests and the emergence of Samsung and what one analyst calls "Team Japan" (NEC and Fujitsu) as alternatives.

But the big winners of the European backlash against Huawei so far have been Ericsson and Nokia, rather than any of open RAN's new faces. Outside China, where Huawei and ZTE now land nearly all network contracts, Ericsson's RAN market share grew from 33% in 2017 to 39% in 2022, according to its own estimates. In 5G RAN, Nokia's share outside China has risen from 23% to 29% in two years, it has claimed (although this does not seem to acknowledge a recent loss of work it did for AT&T).

Useless adornment

Samsung, moreover, could have made do without open RAN and still become a third option for telcos denied the use of Huawei, and its recent gains are no open RAN endorsement. Samsung's main deficiency was its lack of a 2G product in countries where the aging network technology has had to be maintained, and open RAN has never dabbled in 2G. Its 4G and 5G products may comply with O-RAN specs, but operators are usually not taking advantage.

Hence what the industry has taken to calling "single-vendor open RAN," meaning all parts come from the same vendor's system. When Light Reading first reached for this term about 18 months ago, it was in mockery or as a pejorative. Open RAN specs, it seemed, would be a useless adornment, like the male nipple. These days, however, analysts are straight-faced when they include single-vendor open RAN in their market research and forecasts. The oxymoron is not acknowledged.

What some telcos have done with Samsung is run its software on general-purpose compute, carving out business for the likes of Dell (which makes servers), Intel (which makes chips) and Wind River (which develops cloud-computing platforms). Strictly speaking, however, this falls under the category of virtual or cloud RAN, not open RAN. Compliance with O-RAN specs is no prerequisite.

Vodafone is a rare exception, having paired Samsung's virtual RAN software with another company's massive MIMO (advanced) radio unit. But that other company is Japan's NEC rather than a product specialist. It made revenues of more than $21 billion for its last fiscal year and its 5G business was unprofitable.

According to one analyst, the combo is unlikely to be as good as a single-vendor product because the two companies' algorithms would have to be matched, and they would probably revert to a more basic formula to accomplish this (Ericsson and Nokia have made similar arguments). In any case, most of the Vodafone rollout so far has been Samsung. NEC's role might turn out to be very limited (Vodafone will not say).

Pairing small vendors seems even less likely due to the need for systems integration expertise. Ericsson and Nokia together invest almost $1.7 billion in this annually, according to an estimate shared at the FYUZ event in Madrid last year by Robert Soni, an AT&T executive. If the product vendors cannot afford it, a telco would have to supply it directly or hire a third-party systems integrator. Either would be expensive.

A democracy with a single candidate

Anyone who still believes in the original mission of open RAN should sound the alarm. And a few significant people have expressed concerns. "If there's a democracy with only one candidate, is it really a democracy?" said Bruno Zerbib, Orange's chief technology officer, when asked to comment at Mobile World Congress on the $14 billion "open RAN" deal between AT&T and Ericsson.

Others – including Zerbib's colleague Laurent Leboucher, the group chief technology officer – acknowledge the struggle many challengers face. "I think you are right that today there is a risk that smaller vendors – it could be quite difficult for them to join this initiative," he said. "Is there a way for smaller vendors like Mavenir? How can they be part of this initiative? It is not clear."

The names Altiostar, Mavenir and Parallel Wireless – all of which began as software companies headquartered in the US – were synonymous with open RAN several years ago. Altiostar was subsequently acquired by the Symphony unit of Japan's Rakuten and rebranded SymRAN. While its business model now undergoes radical change, it powers no one significant besides the greenfield networks of Rakuten Mobile and Germany's 1&1. Within Dish Network, the other notable greenfield player, its role seems to be as understudy to Mavenir.

Parallel Wireless, meanwhile, cut hundreds of jobs in 2022. When Vodafone executives discuss it now, they use the language and tone of someone talking about an ill or deceased relative. "Emotionally, it is sad, and yes, it is true," said Santiago Tenorio, Vodafone's head of network strategy, on the subject of smaller vendors in late 2023. "We were working together with them and particularly the name that no one mentions now is Parallel Wireless. They were very prominent at the beginning and now they are not a reference anymore."

These various telco executives should be applauded for recognizing the malaise. Conversely, those who regard the highlighting of problems as an attack on open RAN must be asked if they are happy with the status quo and direction of travel. To be clear: New interfaces have not been a multivendor catalyst; the biggest open RAN contract is claimed by the biggest 5G vendor outside China; and the expectation of Dell'Oro, an analyst firm, is that single-vendor deals will account for most future business.

Perhaps none of this is fixable anyway, and open RAN's champions undoubtedly worry that acknowledging any of it is tantamount to admitting defeat. But if there are no complaints, the market is even less likely to change.