While the Open RAN offers obvious opportunities for non-traditional or challenger vendors in the mobile network ecosystem, it presents a dilemma for established network equipment suppliers (NEPs).
So far, Huawei and ZTE have remained distant, even though many contributions to Open RAN code and plugfests have come from the Chinese operators. Ericsson is being cautious. On the one hand, Open RAN appears to challenge its business model just at a time when it is well-placed to dominate 5G RAN. The different recent misfortunes of Nokia and Huawei, plus the strong platform Ericsson built in the 5G-ready phase, have combined to put the Swedish giant in pole position to win the next stage of 4G/5G roll-out.
That explains hostility to the concept in some quarters of Ericsson, but in others, the company is making active contributions to the O-RAN Alliance and is positive about some specifications such as the non-real time RAN Intelligent Controller (RIC). There is an argument that, if Ericsson is sufficiently strong in the RAN, it can afford to support open specifications and still win a large number of contracts on an end-to-end basis – or as the prime integrator in multivendor deals.
This is the gamble taken by Nokia and to some extent Samsung (though Samsung’s small presence in the RAN, outside the USA, India and South Korea, makes it effectively a challenger vendor that can leverage Open RAN to gain market share). Nokia, on the back foot after missteps with its 5G processors, is certainly taking the line that it can afford to back Open RAN reasonably wholeheartedly, and then rely on the reassurance factor of its long history of radio expertise to win the largest deals over less tried and tested rivals.
It remains to be seen whether the large operators stick to their avowed desire to enable a broad RAN ecosystem, to the extent of selecting new players for large-scale macro RAN deals, not just for smaller networks. That will require some big early movers to take some risk and expense on helping their chosen suppliers to scale up, so they become viable contenders alongside the NEPs. But such projects will be important to allay the anxieties of other operators about taking the plunge into a new supplier or a multivendor RAN.
Samsung could be the best of both worlds – major technology powerhouse but in a challenger role in the RAN – or it could fall between two stools and prove too cumbersome for the most fervent open ecosystem supporters and too risky for the conservatives (lack of support for legacy 2G/3G networks is a big issue in Europe, for instance).
But it is hard to argue against the Korean vendor’s engineering smarts, especially in millimeter wave spectrum and fixed wireless access, where its expertise was important in winning it significant roles in AT&T’s and Verizon’s 5G roll-outs. Now it has announced a 5G RAN contract with another extremely demanding operator, in technology terms – NTT Docomo of Japan, which aims to build an Open RAN ecosystem, rather like Rakuten’s, that it can share with other operators and so drive a global platform.
Samsung will provide Open RAN-compliant radios as part of Docomo’s latest procurement. The operator has already deployed a multivendor Open RAN as a commercial proof of concept in parts of Tokyo (it was planned to support the Olympic Games). Much of its early equipment came from vendors in its long-established inner circle, many of them local players like NEC and Fujitsu. But the Japanese MNOs, as newcomer Rakuten Mobile, and also Softbank, have shown, are broadening their supply chain reach, and Samsung is already a long-standing supplier to KDDI. It provides the second largest Japanese MNO with 4G and 5G equipment and recently extended their contract to include new 5G radios for mmWave and sub-6 GHz bands (not necessarily Open RAN-compliant however).
“We are excited to collaborate with Samsung for the next phase of 5G Open RAN and accelerate the expansion of our ‘lightning speed 5G’ coverage in the nation,” said Docomo’s general manager of RAN development, Sadayuki Abeta.
The operator started 5G deployment, using a combination of virtualized and conventional RAN elements in Non-Standalone mode, last year. By mid-2022, it is targeting more than 20,000 live 5G base stations. From later this year, it will start the migration to 5G core and Standalone mode, and aims to introduce more Open RAN elements, especially for enterprise networks. A key focus of the first phase of this deal with Samsung is on private network platforms.
Meanwhile, at Nokia’s Capital Markets Day earlier this month, Open RAN was a key topic for the firm’s head of mobile networks, Tommi Uitto. However, while Nokia is happy to make itself into the self-anointed king of Open RAN in public strategic statements, Uitto’s approach is more nuanced behind the scenes – no doubt reflecting very varying points of view across the Nokia organization. Primarily, Uitto sees Open RAN as a relatively easy route into new customers because customers do not have to feel locked-in, and they can choose just one element or another from Nokia. Nokia’s former CEO, Rajeev Suri, highlighted the hidden benefit of Open RAN for NEPs – currently, if an operator rips out one vendor, the (often multimillion dollar) cost of the replacement often falls to the incoming supplier. The time and cost of rip-and-replace should be greatly reduced once Open RAN hits its stride.
Uitto told FierceWireless last week: “We can win business either in baseband or radio, or both, in some of the networks where we’re not present today by being a leader in O-RAN.”
He acknowledged that meant competing with a wider variety of companies, but he also hinted at what is probably the dream outcome for Nokia – Open RAN deployers that opt for a single-vendor network for simplicity’s sake, but with the reassurance that they can swap in an additional supplier in future if required (of course the NEPs will hope that customer apathy will set in and that most operators never make the effort to swap, in the same way that most consumers never change bank accounts). “For them it’s the flexibility and the ability, for instance in some part of the network to be mixing and matching different suppliers, in other parts of the network use radio and baseband from the same supplier as long as it’s O-RAN compliant so they have flexibility for the future,” Uitto said.
There are areas, he suggested, where operators will make no compromises even in the interests of openness, such as spectral efficiency in dense urban networks or in bands where the MNO has limited capacity. This will make the work that challengers such as Mavenir are doing in radios very important to winning large-scale 5G contracts, and is also one reason many eyes are on NEC, with its long history of radio expertise in a demanding market, but very limited RAN incumbency. “In terms of spectral efficiency, operators are not very willing to make compromises,” he said. “There could be some areas of the network where it’s not such a big deal and then a lesser product would do.”
In the same interview, Uitto highlighted some challenges of Open RAN, and the importance of achieving more consistent blueprints – a key objective of the Rakuten Communications Platform effort. Currently there are too many O-RAN profiles and sub-profiles, covering different RF configurations with different combinations of radios, basebands and spectrum. “It’s a bit concerning that there are so many of these profiles, and then sub-profiles, so then the permutations start to become quite many to manage,” Uitto told FierceWireless.
Another challenge is “reaggregation” of a multivendor disaggregated base station to “ensure interoperability between different suppliers, but also to optimize network performance, network operations, lifecycle management, and ensuring standards release upgrades and features roadmaps are in alignment.” This should lead to “operational standardization”, otherwise every RAN is a systems integration project that is semi-customized and not easily repeatable.
And his third challenge, which applies to all vRANs, relates to the cost and power inefficiencies of using general purpose cloud computing for Layer 1 processing.