I’d bet you that if technology buyers were asked for a single word to describe what they’re looking for in tech these days, the word would be “open”. Why, then, do we keep hearing that network operators are disappointed in the progress of things like Open RAN? Why do enterprises, in contrast, seem to adopt open technology so easily? There are a lot of factors here we need to explore.
The most obvious factor to consider is the credibility of the product source and the source risk tolerance of the buyer. In general operators tend to keep infrastructure longer than enterprises, so they have to worry more about whether their vendor will stay in business and keep supporting their stuff. Of the 88 operators who have commented on this topic, all say that they require a financial/business-stable source. Of the 411 enterprises who offered comments, only 69 say that, and 104 say that it’s a “secondary” factor for them.
The “why” of this is important, though. Operators say that one of the goals of open technology is to multiply the number of sources, where enterprises presume that their source for open technology will be a vendor they’re already familiar with. Thus, enterprises’ presumption is that there will be no new source for open technology and operators think developing one is a major goal for openness. They measure open offerings to conform to that.
Even that begs a “why?” and here again we find a difference in operator/enterprise view. Operators want new sources because they have believed for years that it’s the only way for them to reduce capex, and every operator says they need to manage capex. In contrast, only a bit more than a third of enterprises even mention that, and so while enterprises may also believe that vendors overcharge them, they’re comfortable with competition between their vendors, or with the notion of beating them up over pricing.
The final “why?” comes in here. Enterprises are in general looking for an open technology as a point product. A server, platform software, whatever. Operators are looking for an open solution, usually to a new standard or regulatory requirement, that will mandate multiple product elements. If all these elements come from a pool of sources, there’s a massive integration requirement that operators simply don’t want to deal with. Enterprises rarely even mention this issue, again perhaps because they presume that their current vendors will be the source of open technology. Think Red Hat.
Another factor in the openness challenge that some operators point out is that a solution that doesn’t require integration, or that has less integration risk, is necessarily based on some widely recognized model, like a standard. Operators have generally faced a standards-driven upgrade requirement every five to ten years, and some (as I indicated in yesterday’s blog) have begun to take an interest in long-cycle planning for infrastructure changes even absent a driving standard. In contrast, very few enterprises talk about any preemptive modernization of their IT systems. The need for that has been largely erased by decades of stable client/server and virtualization thinking. The architecture of applications these days still has questions regarding the hybrid use of the cloud, but the framework is evolving in maturity even there.
There are also some supply-side considerations to think about here. For several decades, almost all network operator infrastructure change has been driven by mobile standards, given that mobile services have for decades been the bright spot in profits and the focus of competition. This space has been dominated by a small number of vendors (Ericsson, Huawei, and Nokia, primarily, with Huawei being replaced in many areas due to government pressure). Enterprise IT, in contrast, has a dozen major vendors or more. The chances of a “new” vendor entering the enterprise IT space is far better than the chances of a new vendor entering the mobile infrastructure space. In most cases, Open RAN players not one of the big vendors are partial-solution players, getting buyers back to the integration problem.
The most significant new factor in the picture differs for the enterprise and operator communities, too. For enterprises, its the cloud and AI, perhaps more generally the role of as-a-service versus in-house. For operators, it’s the determination to push 6G standardization as a software-only strategy, which would tend to make it easier for new players to enter the space.
What operators are asking for is an infrastructure model that, in the end, is more like that of the enterprise. Give us a stable, long-term, equipment model and then augment it with the necessary feature advances using software and perhaps the introduction of feature-as-a-service. In all, it could create a revolution in operator infrastructure, which is why perhaps the whole Open RAN or open-model 5G should have focused on that goal from the first. Which is also why 6G needs to follow operator demands here. Which it may, or may not, and we can see some hints of why that is by looking at Open RAN and 5G.
One contaminating issue was that the 3GPP 5G specifications didn’t really think about creating a universal hardware framework, one that could even support 5G much less endure beyond 5G. There was an unhealthy mix of trying to preserve the general 4G-LTE model and trying to converge with wireline infrastructure. We all know the old saw about serving two masters, and making it three might well have stalled 5G for years more.
That was the second contaminating issue. The whole 5G standardization process was simply too lengthy, and that’s typical of these international standards initiatives. Another old saw is “The IQ of any group of people is equal to the IQ of the dumbest, divided by the number of people.” Consensus is not only time-consuming, but it’s an enemy of innovation.
The third issue was Wall Street and the media. The Street values stocks by quarterly earnings, and something that’s aiming at 10-year relevance is hardly going to come to fruition in three months. Something whose value is boring financial credibility isn’t going to generate any clicks on stories either, and those clicks not only help ad-sponsored media but also the stock price of vendors and operators.
All of these issues will impact 6G evolution as much as they did 5G, and you also have to wonder whether the vendors, who tend to dominate all standards processes by the simple mechanism of staffing them with their own people, would accede to a standard whose goal was to reduce operator spending on equipment. That’s something we should watch as 6G evolves, but we probably won’t know the final picture on 6G until around 2029.