I’m a fan, and an advocate, of open systems. Thus, I’m a fan of the concept of Open RAN. The problem is that what I hear from the mobile operators doesn’t support my convictions, and I think it’s important that we look at the Open RAN situation to see why that is. We’ll start by asking the usually-obvious, often-ignored, question of “Why are we doing this?” Then we’ll ask whether the issue here goes beyond Open RAN, to open-model networks in general.
Why indeed? Telcos have said for decades that they wanted to avoid vendor lock-in, get innovative best-of-breed technology, reduce costs by undermining big-vendor pseudo-monopolies…you’ve heard it all before. The problem is that all of that is under threat, and maybe even more than that. Some of the cherished open-model benefits that both telcos and enterprises have touted have been repudiated in 2023, and more seem likely to fall. Things like Open RAN, things that would have been automatically accepted as essential even two or three years ago, are now being rejected by a growing community of buyers. They’re embracing the proprietary instead, and it’s really important that we understand why.
From the first, it was clear that there was a collision developing between past, cherished, notions of an open market for network technology and a growing fear that multiple sources meant multiplying integration challenges for network operators. A big part of the problem stems from the increased difficulty that buyers of technology report in acquiring and retaining skilled network professionals. The number of “skilled network professionals” in the market has more than doubled since the early 1990s but the number of jobs that buyers say require them has risen 500%. In 1995, over two-thirds of users and 100% of telcos said that they were fully staffed with those professionals, but in the last two months only 13% of users and 37% of telcos said they had even an “adequate” level of professionals on staff. Among telcos of all sizes, the average percentage of operations jobs filled with skilled professionals this year comes in at 22%.
While fears on integration difficulties’ impact on operations is higher among companies with a shortage of professionals, even companies who feel they have a strong staff are twice as likely to be concerned about integration as they were even five years ago. This is due to the exploding number of error-driven outages, which almost all network operators (telco or enterprise) say is expanding because of expanding complexity. When users are asked about the relative complexity of multi-source versus single-source networks, 88% say that the latter are less complex.
If integration is the big negative driver issue on open-model networking, including Open RAN, what’s the big positive driver. When we asked that question three years ago, the answer was “innovation” in general, and specifically the notion that competition would drive a larger number of useful features that could then be exploited to gain business benefits and profits. Open-model networking in general, and Open RAN in particular, depend on this positive balancing the integration negative. For that, there are two main barriers.
The first barrier is the general problem associated with 5G, which is that nobody has really devised an incremental revenue benefit for the technology. Whether 5G is “better” than 4G might be an easy question, but so is whether the “betterness” is something that users will pay more for is also easy—no they won’t. Every step in 5G, from New Radio through Core, has been accompanied by ballyhooed stories about new applications, benefits, and revenues. None have come true to any significant degree, nor IMHO are they likely to. Could there be 5G application revenues? Sure, but they’d depend on applications whose core technologies themselves are still waiting in the wings. You can’t connect what cannot yet be deployed.
That raises the second barrier, which is that 5G RAN isn’t where the new features are likely to be needed. When you look at the ROI of a feature, the “I” depends largely on how efficiently you can support the feature’s hosting. I’ve noted many times that the sweet spot (geographically) for feature hosting is the metro concentration points, and these points are really the on-ramps to 5G Core and not a part of the RAN at all. Opening the RAN is therefore unlikely to spawn competition that leads to revenue gains, though it might lead to cost reductions.
Telcos, meaning in this case mobile operators, are privately willing to tell me that they no longer believe that there is significant incremental revenue opportunity in 5G deployment for at least three years, and many thoughtful telco planners say that they cannot say how long it might take for any new revenues, given that this depends on things that are out of telco control, as well as out of the scope of network standards in general.
This all illustrates the weakness of the “Field of Dreams” or “build it and they will come” mindset. Even if it’s true, how long will it take before they do come? What we’re seeing in the world of telcos, and in Open RAN, is that all the risks of open-model networking have to be faced immediately, and those risks relate primarily to complexity and integration challenges. The benefits? Well, we don’t even know for sure when they might roll in, nor do we know what benefit levels we could expect. Given this, telcos are finding themselves in exactly the same place enterprises have found themselves, which is that they can’t justify the risks through credible benefits.
Enterprises are pulling back, not only from open-model networking but from multi-vendor networking. Telcos aren’t quite as far along as enterprises in this shift, but they’re pretty clearly on the same path. I don’t see any force on the horizon that’s likely to change this, and so Nokia’s approach to Open RAN, which is to support it while emphasizing their single-source solution, is the best one.
How did it come to this? It seems that a combination of strategic and tactical forces are to blame. On the strategic side, the big problem is that we’re at a point where new network spending needs to be tied to new benefits, and those benefits have to be delivered through applications and experiences that need more than network connectivity to be realized. On the tactical side, we’ve built more hype than reality in open-model benefits. In particular, a “benefit” to be credible has to not only be sufficient in quantity and timeliness, but also within the scope of the investment we’re proposing. If a RAN can’t deliver the benefit, then the benefit can’t overcome incremental costs and risks associated with the investment.
Are we headed for a future where only the major network vendors can win in telco networks? I think that may well be the case based on what telcos themselves are telling me. The only thing that could reverse this trend is the development of real new revenue opportunities based on feature advances that a competitive market could deliver best. That may take some time.
