The big drawback of not attending MWC in person is that we cannot have that private conversation with an old friend after his VP of Marketing gives a grand keynote promising the world. Over a glass of sangria and a plate of tapas, we can get a more nuanced view of what is actually happening in the market. That being said, we can still read a lot between the lines of the various press releases and articles coming out of the show. Nowhere is that more interesting than what is happening around the Open RAN (O-RAN) project.
We are going to have to dive into the acronyms. Apologies, but this is telecom. O-RAN stands for Open Radio Access Network. The Radio Access Network is the layer of the network with which mobile phones make their first point of contact with the wider telecom network and through that the Internet – think base stations and the system that coordinates among all the local base stations. This is really important real estate, it is the main point of access control that the wireless operators have for their networks and it directly touches their most valuable asset – the radio frequency bands they control. Changes here have a big impact on the operators’ financials.
For years, the equipment for the RAN was entirely controlled by the big equipment makers – Ericsson, Nokia and previously Huawei. The communication between a customer’s phone and the base station is determined by the wireless standard (aka 4G and 5G), but the software that runs on the base station was largely controlled by those equipment makers. This is a crucial point of differentiation for them. The rest of the network could largely be run over standard communications protocols like IP/TCP (aka the Internet), but their position in the RAN allowed the equipment makers to resist the shift to Internet protocols in the rest of the network – and thus sell more proprietary equipment.
However, with time, this position has grown steadily less tenable. A big part of the 5G standard is focused on moving the rest of the operators’ networks to shift to more “open” hardware – decoupling the software and hardware that runs on core switches and billing and management software. Operators like those changes because it allows them to pick and choose their hardware and software separately. They could, in theory, buy inexpensive commodity servers to run whatever software they wanted, rather than buy expensive, proprietary hardware and software bundles from Ericsson. For the most part, 5G does not extend those changes to the RAN, which was a deliberate and sizable gap.
So a few years ago, a group of companies got together and proposed an alternative approach to building the RAN, extending the separation of hardware and software down to this level. This is not the first time someone proposed something like this, but the big changes in 5G gave the project considerable momentum.
We mention the history of past initiatives because the equipment operators held those off with what is now a fairly well established playbook. First, they deny the utility of O-RAN, then they shift to cautious skepticism continuing to raise doubts. Finally, they embrace the proposal, but do so in such a way as to shift its implementation in their favor. We are now in that final stage.
Case in point, we were prompted to write this post by an article in Light Reading entitled “Nokia mobile boss lashes out at open RAN fakes“. The quick summary of that article is that a Nokia executive is calling out unspecified others for not adhering closely enough to the O-RAN ‘standard’. He is not talking about arch-competitor Ericsson. Instead, we read this as Nokia having fully “embraced” O-RAN and are now in the process of shifting it to align with their interests. It is an amazingly effective position. Nokia can wrap itself around all the goodness of the O-RAN marketing image, and then call out the apostates who want to take a different approach. Left out is the fact that the unspecified others have been promoting a much more flexible O-RAN for years before Nokia jumped on the bandwagon.
As with most such standards disputes, there appear to be two camps emerging in this fight, those who want a more flexible, “open” O-RAN, and those who want something that is not that. We view this as a spectrum, there is a lot of squishiness in these definitions, with different vendors partnering with opposing camps for various operator deployments. On one end of the spectrum are the incumbents, along with their various proxies among the operators. The opposite end has new entrants to the market like Mavenir, Fujitsu and Rakuten. A big variable are the chip vendors. This is shaping up to become a conflict areas between Intel, who has been working on this for years, and Qualcomm who seems to have gotten religion fairly recently on this, but is losing no time in jumping into the fray. Both companies are investing heavily in support for telecom software. Intel is much better at this game than Qualcomm, but Qualcomm has 30 years of operator relationships on their side.
Perhaps the biggest wild card in this is Microsoft – and to a lesser degree Amazon and Google – who is pushing its telco cloud solution for Azure very hard at this year’s show. Our best guess is that Microsoft is largely ambivalent, they do not care whose software they run, so long as it is run on Azure. As a software company, they will tend to prefer a more open O-RAN implementation, and they have built a lot of software to support this, but they have partnerships with both Nokia and Ericsson as well – insert picks and shovels in a gold rush metaphor.
This is a high stakes fight, there are considerable amounts of money at stake here, as well as long term implications for the operators’ future. Which is why we tend to err on the side of O-RAN ending up with a more conservative approach. The operators do not like to make radical changes quickly, and O-RAN still has a lot of rough edges. For years, the operators have relied on Ericsson and Nokia heavily, and it is unclear that they have sufficient motivation to move away from that influence right now.
Ultimately, we think the operators are going to be dis-articulated, O-RAN may be a step down that path. So it is important to watch this, especially at a time like this when the big cloud service providers are making increasingly attractive offerings to entice the operators onto their respective clouds. It is too simplistic to say the operators are being forced to pick their poison with a choice between equipment vendors on one side and the hyperscalers on the other, but all of this O-RAN positioning and the history of networking should keep the operators cognizant of the risk that they end up someday being consumed by those hyperscalers.