Your browser is not supported. Please update it.

20 July 2020

O-RAN may be both too early and too late to achieve a diverse vendor base

The rising pressures on Huawei (see separate item) provide an opportunity for the emerging open RAN architectures, and their suppliers, that previous attempts at a new open mobile platform did not enjoy. When a vendor is incumbent and doing a decent job, the disruption and risk of swapping equipment and trusting to a less familiar partner often outweighs the temptations. But when a vendor has to be compulsorily removed – and at a time that coincides with the adoption of new technologies in 5G and virtualized networks – there is far more incentive for operators to try something new.

In the parliamentary hearings that led up to the UK government’s decision to ban Huawei from the country’s 5G networks, two of its major Huawei-using MNOs, BT/EE and Vodafone, said they would consider O-RAN platforms as part of their replacement strategy. The government itself is said to have held recent talks with Samsung and NEC, two major Asian vendors which hope to boost their tiny RAN businesses by riding on the O-RAN wave, while offering operators deeper pockets and longer track record in the mobile ecosystem, than the O-RAN challenger suppliers.

Of course, O-RAN has become very politicized, having effectively been adopted by the US government as the poster child for a campaign to build a homegrown mobile networks industry again. That favors the group of US-based challengers, whether well-known midrange players like Airspan, or start-ups like Mavenir, Parallel Wireless and Altiostar.

But the O-RAN suppliers with the bulk and international presence are Samsung, NEC and, of course, Nokia, which recently unveiled a portfolio of O-RAN and vRAN products designed to assert its own leadership of the nascent open platform. This will, it hopes, enable it to regain operator trust as a strong replacement candidate for Huawei, after recent 5G setbacks; to score points against Ericsson, which has not made any definitive O-RAN commitments; and to fend off the challengers by citing its own expertise and experience in cellular networks.

The tactic may still work – and if Nokia does increase its market share thanks to O-RAN, the full vision of open networks supplied by a host of interoperable vendors will be severely compromised.

The risk for that O-RAN vision lies in the timing. The new platform is coming to the commercial market too late in one sense. As noted above, as swapping base stations and mounting antennas becomes easier thanks to virtualization and new site processes, one of the key arguments for O-RAN – that it allows operators to change their suppliers more easily – is being weakened. This will be especially true if large vendors support the basic O-RAN interfaces as a tick-box item – they can then rely on operators being far keener to have the option to change suppliers whenever they wish, than actually to go through the process, however simplified it may have become.

In another respect, O-RAN may be too early. It has made significant progress, as have other open RAN initiatives such as Small Cell Forum’s nFAPI, in defining a new, interoperable architecture, highly disaggregated and software-driven, with a path to cloud-native support. These organizations’ efforts will almost certainly dominate the future in private, industrial and small cell networks. They have also succeeded in changing the conversation about how RANs are built and establishing openness as a credible future.

But as we have often argued, a fully mature O-RAN platform, which operators are confident to deploy in their primary networks, is years away, which means it will miss the second wave of 5G build-out – the migration to Standalone 5G – in many developed mobile markets. O-RAN, or some successor to it, has a strong chance of leading in the next wave of macro 5G, especially when supported by efforts to make it readily deployable and scalable, and to fit within an end-to-end open network (key contributions by the Telecom Infra Project to this ongoing process). But all these developments will take time to mature, and some operators want to run RFPs and even roll-outs right now.

In the UK, BT’s and Vodafone’s discussion of open RAN in the parliamentary Science and Technology committee was tempered by caution, even about Samsung, one of the biggest O-RAN supporters and a company which is involved in large-scale commercial 5G deployments (though not O-RAN ones as yet) in north America and elsewhere.

Howard Watson, BT’s CTIO, testified that his priority for now was single RAN (a platform that supports 2G, 3G, 4G and 5G in a software-defined way on the same infrastructure) rather than full openness – and the single RAN trials have been conducted with Ericsson and Nokia. He was asked – before the Huawei ban was announced – what impact the previous policy of a 35% cap on the Chinese vendor’s kit was having on the business. He said: “We are going through a commercial and technical evaluation of, do we use more of our second-source vendor, which is Nokia, to shrink Huawei to 35%, or do we introduce a third vendor that is capable of single RAN?”

Vodafone has been more upbeat about open RAN. In response to news of the ban, the company said in its official statement that it was disappointed by the decision “disappointed because this … will add delay to the roll out of 5G in the UK and will result in additional costs for the industry. We will work with the government to address the implications of this decision, including the cost and the need to increase vendor diversity through Open RAN.” That was taken as a veiled demand that the government provide some funding to support the replacement of Huawei, as the US administration is planning for rural operators – but also suggested that, as in the USA, that might be tied into support for an increasingly politicized open RAN platform.

But Vodafone is clear that open RAN technologies are not yet ready for prime time. Santiago Tenorio Sanz, head of group network architecture at the telco – and chair of TIP – said in an interview with Mobile World Live that these solutions are not yet ready for dense urban areas or 5G. “I think the technology is ready in 2G, 3G and 4G to serve some … not the most demanding areas. So, outside of dense, urban cities: suburban, rural areas, we’re ready for that, but I think it will take some more development time to deploy open RAN and to bring commercial service to dense urban, and also to 5G.”

He is encouraged by support from large vendors like Nokia, though some of the smaller firms may regard that ambivalently. He added: “When you reach a point at which incumbents in general and the industry stops telling you that you’re doing the wrong thing and starts jumping on the train, that is when, you know, the train is about to leave the station.” The question is whether the train will be ready, and operating at full speed and safety, in time for the deployments of the first wave of major 5G SA deployments.

Samsung is, understandably, positioning itself assertively to replace Huawei in markets such as the UK. Testifying to the Science and Technology Committee, Samsung EVP Woojune Kim claimed it had as much as 20% of 5G RAN share. But this does reflect its success in two major markets, South Korea and the USA, but not its limited role in existing networks.

On this point, Vodafone and BT highlighted another of the key issues of vendor replacement – they are not yet ready, unlike operators in Samsung’s largest markets, to turn off 2G or even, for a shorter time, 3G. Andrea Donà, head of networks at Vodafone UK, told the committee hearing: “There are devices which are 2G/3G only enabled. The transition to a fully blown 4G/5G has to go hand-in-glove with our commercial colleagues to replace the devices on the ground and ensure that only 4G/5G devices are present before we can completely move away from 3G and 2G technology.”

That makes it hard to work with a vendor which does not support the legacy networks in any meaningful way, though of course Samsung has devices for all the mobile generations, and may judge it worthwhile to place heavier focus on 2G to avoid being excluded from major replacement contracts. In other respects, it does look like the heavyweight among the O-RAN/TIP challenger suppliers. When Vodafone conducted a 5G request for information (RFI) under the auspices of TIP in 2019, Samsung – not one of the start-ups – emerged as the vendor which complied most fully. But its lack of 2G/3G/4G/5G Single RAN products puts it at a serious disadvantage in markets, notably in Europe, where most operators aim to hang onto 2G and even 3G for many years.

“The key point is the single RAN concept and the 2G and 3G legacy,” said BT’s Watson in the committee hearing. “Right now, 50% of my voice calls still use the 2G and 3G network. My view right now is that for a replacement technology today, 2G has to be a critical part of the solution.”

Woojune Kim argued that Samsung could offer a 5G overlay solution based on the X2 interface, which allows handover from 4G equipment from one vendor and 5G from another, but has not been fully supported by the large OEMs. X2, along with technologies like dynamic spectrum sharing (DSS), can make it easier to phase a new vendor into a network without ripping out the old equipment too rapidly, and Kim says that would be easier for Samsung to achieve than full O-RAN, as well as a more mature solution.

“In Korea, we showed you could put a 5G network on top of another vendor’s 4G network and it works perfectly well,” he told the committee. But Donà argued: “I need to build 5G coherently with the 4G stack I have to maintain the customer experience.” And Watson said an overlay takes far more spectrum than SIngleRAN.