ANALYST Q&A: OPENING THE RAN MARKET

David Martin

Associate Senior Analyst STL Partners

Gareth Owen

Associate Research Director Counterpoint Research

Stefan Pongratz

Vice President Dell’Oro Group

Eric Hanselman

Principal Research Analyst 451 Research

The Open RAN market is one of the most hyped sectors in mobile – so what’s driving growth and what are the biggest challenges ahead? Four top analysts offer their insight.

Q: Why is open RAN being hyped as the next big technology in the mobile infrastructure market?

David Martin (DM) There is a desperate need to expand the ecosystem of RAN suppliers, as the RAN market is dominated by a handful of major vendors at present. Disaggregating and virtualising the RAN provides a means to open it up to a broader range of suppliers, which will create more competition and opportunities for innovation in the space. This should in turn also reduce the cost to deploy and operate the RAN. This factor is all the more important in the light of political and regulatory pressures for operators to reduce the presence of Chinese suppliers in their networks.

Open RAN makes it easier and cheaper to put mobile network and radio capacity of different types into a variety of environments where this is not cost-effective using legacy hardware-based, single-vendor platforms. This includes rural and sparsely populated areas (including in developing markets) as well as densely populated areas in developed markets, where open RAN offers a cost-effective means to fill in and densify coverage, including outdoor and indoor small cells. It also provides a cost-effective means for both MNOs and non-MNO players, such as neutral host providers and others, to deploy and operate private LTE and IoT networks for specific enterprises and industries. I explore some of these use cases in STL Partners' recent report: Open RAN: What should telcos do?

Open RAN is part of a broader trend towards cloud-native network functions (CNFs), including containers and container orchestration frameworks (e.g. Kubernetes). This is seen by many as crucial to realising the potential of 5G. In particular, open RAN should allow 5G NR capacity and spectrum resources to be deployed, controlled and scaled in a highly dynamic manner into many more locations, including different types of edge site, where they can support use cases dependent on ultra-low latency or massive IoT capabilities.

Gareth Owen (GO)

I think it is inevitable that what happened in the IT industry a few years back is now happening in the telecoms industry, including the mobile industry, i.e. the disaggregation of software and hardware which allows new players to enter the market driving up innovation and (potentially) reducing infrastructure costs for operators.

Networks have to become virtualised and cloud native to benefit from the potential offered by 5G such as MEC, network slicing, etc. This does not necessarily need to involve open interfaces. It could still essentially be a proprietary solution as offered by Ericsson, Huawei, etc. However, Nokia recently announced its vRAN 2.0 product portfolio with open fronthaul interfaces, so I think open interfaces will become an integral part of 5G RAN going forward, although it may take some time for the other incumbents to follow suit.

The other angle to this is the MNOs. Some of the very biggest operators (Telefonica, Deutsche Telekom, Vodafone) are unhappy with the limited choice of products available from the big three vendors and so have been instrumental in trying to break down the stranglehold of the biggest vendors. They are doing this by developing open versions of the various RAN interfaces with the idea that they can “mix and match” software and hardware elements of the RAN thereby allowing new players to enter the market and drive down infrastructure costs.

I think the drive towards virtualisation is the really important trend here, more so than open interfaces at the moment, although the two go together. The priority for most operators is to get their 5G networks rolled out, particularly the new 5G NR stuff. I think open RAN is less of a priority for them right now but something they will definitely be interested a bit further down the line.

Stefan Pongratz (SP)

Open RAN has many of the right ingredients to address both supply and demand related challenges that continue to characterise the mobile infrastructure market. The momentum is picking up pace, partly a result of the realisation that the technology is working and partly because the timing is favourable, with the escalation in the geopolitical uncertainty acting as a catalyst for operators to rethink their supplier strategies.

With few signs that the revenue and market share concentration trends are about to reverse anytime soon, regulators, governments, and operators are increasingly concerned about the competitive dynamics in the mobile infrastructure market. More specifically, there is a growing concern that further consolidation among the top suppliers is increasingly likely. So while cost savings might have been the initial driver, operators are now considering open RAN because it can broaden the supplier diversity, simplify swaps, and improve the supply chain visibility.

And while few will argue that the leading RAN suppliers have done an excellent job introducing 5G at a much faster pace than expected, open RAN proponents believe that the shift toward more open, software-driven, virtualised, and intelligent multi-vendor RAN solutions could spur more competition and innovation, accelerating the shift toward 5G beyond MBB.

But the acceleration we have seen over just the past couple of months is likely more driven by the supply side than the demand side benefits with open RAN.

Eric Hanselman (EH)

The RAN is the last link in the wireless value chain to be fully virtualised and it presents a new opportunity for operators. Because there’s a lot more edge infrastructure than core, it also has a much greater potential impact on both operational flexibility and cost management. Opening the RAN environment enables innovation in an area that’s long been locked into physical implementations.

Q: Do you seen open RAN players being able to pose a challenge to the traditional network vendors, like Ericsson, Nokia and Huawei? How will these vendors counter this challenge?

DM Open RAN undoubtedly already is posing such a challenge and is forcing the traditional vendors to engage with it in forums such as the O-RAN Alliance, the TIP OpenRAN group and the Open RAN Policy Coalition. One way in which the established vendors can respond is by developing their own open RAN platforms. And indeed, Nokia did announce its own 'open RAN' solution based on O-RAN Alliance-compliant interfaces, albeit using its own hardware. This will be rolled out commercially next year.

The traditional vendors also have a lot of ace cards in their hands, including their vast installed base of proven RAN hardware at operators worldwide. These vendors are adept at offering a smooth migration path from one mobile generation to the next, for example with multi-G radios that can be upgraded to 5G via a remote software update. They provide pre-integrated, turnkey, carrier-grade solutions along with packages of operational services, monitoring and analytic software, etc., all of which are designed to lower MNOs' operating costs (although there is a trade-off with the higher capex involved in purchasing and deploying their platforms).

It is far from a straightforward proposition to rip out integrated base stations (radio and baseband) and replace them with a multi-vendor, open RAN solution. And the larger vendors' radios have proven performance and energy efficiency characteristics, while open RAN solutions are as yet largely unproven at scale.

With Nokia now starting to offer O-RAN compliant interfaces with its vRAN 2.0 product, it is clear that the open RAN movement is having an impact. However, I suspect that the two biggest vendors, Huawei and Ericsson will resist adopting O-RAN interfaces as long as they possibly can as they have potentially more to lose.

However, I believe that the traditional vendors will play an important role in how open RAN unfolds. The fact that Nokia now offers the option of an O-RAN fronthaul interface legitimises the open RAN approach. However, it could also be viewed as a threat to the new RAN players.

Nokia could market its vRAN 2.0 product to MNOs as a single vendor product, thus dispensing with the potential integration and reliability problems of a multi-vendor open vRAN product, while at the same time dangling the prospect of enabling MNOs to switch to alternative vendors a few years down the line, should they wish, for example, when a market in whitebox O-RAN radios has developed.

Nokia’s vRAN 2.0 solution is still a 100% Nokia product, although that may change if other equipment and software vendors link up with Nokia to diversify its portfolio. Nokia could develop a preferred vendor ecosystem consisting of selected alternative software and hardware providers to offer MNOs “tried and tested” open RAN/vRAN integrated solutions. This would provide “peace of mind” for customers while at the same time minimising market share loss and maintaining some control of the market. This would make it more difficult for the smaller O-RAN players to enter the market.

Having said that, I think there are opportunities for the open RAN players. In the short to medium term, the best opportunities are probably in greenfield, rural and private networks.

One of the reasons the RAN supplier market concentration index has trended upward rather than downward over time is that the R&D intensity required to maintain a competitive mobile infrastructure portfolio is significant. This in combination with the inherent vendor lock-in has complicated the entry point and resulted in a rather poor track record for new entrants.

Open RAN should be able to lower the R&D threshold and reduce the barrier to entry by spreading out the resources over more suppliers. The geopolitical uncertainty should also help to mitigate some of the lock-in related concerns and provide an improved entry point.

At the same time, even if 5G investments are accelerating rapidly, it is still early days in the broader 4G to 5G transition with 5G now initially being just another G. But the long-term vision still holds, and we firmly believe that we have just gotten started on this journey and the opportunities ahead are simply massive as connectivity continues to proliferate through all aspects of society.

Open RAN players now have a ticket to the table and hopefully they will use it wisely. But no one should be under the illusion that opening up one interface is all that it takes to succeed. Open RAN or not open RAN, everyone is going to have to transform, accelerate innovation, and continuously adapt to position themselves for success beyond this first phase in the 5G journey.

We’re already seeing the market effects of greater openness in RAN. Traditional vendors are offering decomposed virtual functions and enhancing platforms to support externally sourced functions. Newer vendors are offering both architectural and functional innovations that are competitive. The shift to containerisation has created openings in the market that are toe holds for innovators.

Q: What are the biggest hurdles/challenges that open RAN faces?

DM The biggest hurdle is the integration challenge: the difficulty, time and cost of integrating a multi-vendor solution, with all the disaggregated components of the RAN potentially each being supplied by different vendors. It is known, for example, that the open RAN and cloud-native champion Rakuten went through considerable pain, extra time and cost integrating its multi-vendor platform; and it is behind schedule with its 5G open RAN launch. Rakuten is now marketing that platform as a pre-integrated solution for other operators. And members of the TIP OpenRAN group are also assembling and testing standardised, multi-vendor deployment templates that can be reused and adapted by multiple MNOs.

Alongside the integration challenge, there is also the challenge of new operational processes and organisational capabilities involved in continuously implementing and developing an agile, software-based platform such as open RAN. This also requires new partner ecosystems and cross-industry collaborations with a host of new vendors and actors, as well as with other MNOs.

In addition, open RAN needs to demonstrate that it can offer comparable and indeed superior performance to the traditional vendors' platforms (particularly in the radio) along with reliability at scale. Open RAN is still widely perceived as immature and unproven; but the many trials and tests that are currently ongoing are working hard to provide the required proof points.

I think the main concern for MNOs is the increased complexity and potential integration problems of dealing with a multi-vendor open RAN solution. Not so much for green build networks such as Rakuten Mobile in Japan or Dish in the US who do not have an installed user base to worry about, but it is a definite issue for legacy network operators.

Then there are also issues with regards to network performance, particularly with dense networks in city centre locations, TCO (both OPEX and CAPEX) which is not a “slam-dunk” as some of the open RAN players claim. There is also no whitebox market in O-RAN radios yet, and in any case, it will take time (i.e. volumes) before the cost of these radios come down.

As with most technology transitions, there will always be some speed bumps over the short-term/near-term – some of these include:

The inherent vendor lock-in with brownfield networks. The geopolitical uncertainty is helping, however the business case will likely remain challenged for a sizeable portion of the NSA market.
Open RAN readiness with more advanced systems. While the win-rate for Open RAN in NSA NR settings was always expected to be weak, Open RAN suppliers need to accelerate investments to ensure they can capitalise on any green-field networks and all the upcoming swap and overlay opportunities popping up as a result of the geopolitical environment. 3) Interface split - one or two interface options might not be enough to find the best balance for managing the agility and complexity required to support both basic and also more advanced radio deployments for all the usages scenarios. 4) System integration for smaller operators.

Open RAN is a new operational model for operators. They’ll have to be much more involved in the selection and qualification of functional elements than they’ve had to be in proprietary vendor relationships. That gives them a new level of control, but it will require resources and skills that they haven’t had to use in the past. The agility that a continuous integration/continuous deployment (CI/CD) operational pattern can deliver to operators could have huge benefits, but it will take new skills and operational procedures to capture them. Competitive pressures make this transition inevitable, but operators have to consider whether to undertake it on their terms or have it forced upon them.

Q: Why would open RAN technology be attractive to operators and even governments in the future?

DM For all of the reasons indicated above, and others: more choice and competition; more flexible, programmable and scalable RAN platforms to support different use cases; the general migration to cloud-native networking, of which open RAN is an instance; the potential for considerable cost savings and efficiency gains from using multi-vendor software platforms on general purpose hardware, along with cheaper, open source radios.

From a government point of view, there is also a clear interest in stimulating more competition and innovation in the mobile infrastructure market, as 5G in particular is viewed as a major driver of future economic growth and social benefits. With the pressure to reduce the influence of Chinese technology firms in Western infrastructure, the need to stimulate more competition and choice in the market is all the more urgent.

The promise of open RAN (indeed, any truly “open” technology) is that it provides a level playing field and introduces new players into the market which fosters innovation and costs reductions. Operators thus benefit from a greater choice and variety of products as well as lower costs.

Unfortunately, there is also now a political dimension to open RAN which the US government is using as part of its trade war with China. Although there is a security risk element behind the US government ban on Chinese vendors, I think an additional and equally important objective is to support US companies at the expense of Chinese vendors while at the same time curtailing the US’s dependence of Chinese technology.

This should be a good opportunity for open RAN. With Huawei and ZTE banned in the US, the choice of vendors is restricted to just two really big vendors (Ericsson & Nokia) plus Samsung. Open RAN can therefore become a 3rd (or 4th) choice for operators and although the technology is not ready for the big time in very dense city centre-type environments, it could be a suitable alternative for rural operators, many of whom may soon be faced with the prospect of having to replace their Huawei and ZTE kit with alternatives.

The near-term drivers for both the operators and governments are driven primarily by the supply side benefits including broadened supplier diversity, simplified swaps, and improved supply chain visibility. In addition, the technology could help to address concerns about market concentration. And over the long-term, the shift toward true Open RAN solutions could spur more competition and innovation, accelerating the shift toward 5G beyond MBB.

Openness has a range of benefits that have already become well established in operator networks – operational agility, the reduction of friction in adding new vendors and the ability to experiment and innovate more easily. In the RAN, these benefits are only just starting to be realised and can increase the competitiveness of operators and the agility of governments. Operators can deliver new services faster and governments can respond to emergencies and disasters in ways that are much simpler than proprietary systems offer.