Private 5G is even more difficult to assess, in value-proposition terms, than 5G overall. My own chats with enterprises suggest that there’s really not all that much going on. Other sources suggest otherwise, and are looking at possible “private 5G models”. TelecomTV has a story on this, and we’ll use that today to look a bit harder at the private space. We’ll also roll in initiatives like those of the ONF with Ananki and Rakuten’s own Communications Platform.
5G is a 3GPP standard for next-gen mobile networking. Semantically speaking, anything that claims to be 5G would have to conform to that standard, but the standard would allow for two broad 5G subdivisions—public 5G service offered by mobile operators, and private 5G services deployed by users. If you think of 5G at this level as having a sort-of-cloud-like framework, you could also see “hybrid 5G”, which would be a combination of the two. That combination might hybridize with a public 5G service, or with a network slice using that feature of 5G.
There are also mission-based private 5G models. Again, at the highest level, there are “industrial” or “IoT” applications, and then applications designed to offer communications through traditional mobile devices. I’ve not seen very much interest in the traditional-device model so far, but then (as I’ve noted) I’m not seeing a lot of private 5G adoption either, just tire-kicking.
We could also divide private 5G according to the means of implementation, and there are four broad options. First, an enterprise could partner with one of the mobile infrastructure vendors (Ericsson or Nokia, for example), or they could acquire their infrastructure from one of the open-model 5G providers. Finally, they could acquire the technology from an integrator, or they could do their own 5G integration from available elements.
The TelecomTV article I cited references third-party research, and I’ve pointed out in the past that asking enterprises things like “do you plan to adopt private 5G” nearly always yields little in the way of accurate results. I think that’s the case here. The lead graphic says, for example, that the top industrial connectivity options are WiFi, fiber, Ethernet, private LTE, and SD-WAN, with each of the first four having over 60% adoption. That’s in no way consistent with my own findings, unless you broaden “industrial connectivity” to mean anything an industrial company uses, which would of course have little relevance to private 5G opportunity.
The assertion of the same chart is even more striking; over three-quarters of companies expect to adopt private 5G by 2024. Again, this is totally at odds with what I’ve heard from companies actually looking at industrial connectivity; over three-quarters of them said they had no idea when, or whether, they’d adopt private 5G.
What’s really going on with “connectivity in manufacturing” has to be divided to assess. Almost all enterprises have the same basic issues in their enterprise connectivity strategies. Manufacturing differences arise largely because of manufacturing, meaning industrial IoT. That topic divides twice—once by whether the elements are fixed or mobile, and again by whether the application is green- or brown-field. Each of these divisions impacts 5G value, whether public, private, or hybrid.
My (limited by the small number of actual adoptions/considerations of private 5G) contacts tell me that the biggest driver of private or hybrid 5G consideration is the need to support locally mobile IoT elements. If the application is also greenfield, then private 5G is very likely worth looking at. If some IoT elements could be mobile beyond the scope of a facility, then hybrid 5G is a viable option. Users offered the example of warehousing and transportation as an application space where this could happen.
Where IoT elements are fixed, the preference I get from my contacts remains for some form of wiring. True manufacturing/industrial IoT often bundles IoT elements and controllers with the tools involved in creating the product(s), and in most cases they say that the connections are made using wiring. Greenfield versions of these applications seem to follow the same wiring preference, but where something has to be added to a manufacturing/industrial process, wireless may be easier to integrate. There, 5G might be an option.
When does consideration of 5G turn into adoption? There’s some congruence between what the article suggests and what I hear, but not complete congruence. Security and latency are what I hear could be differentiators for private or hybrid 5G, but that’s true only where the alternative is either 4G/LTE or WiFi, since wired IoT is obviously the lowest latency of all. What’s interesting is that a slight majority of enterprises tell me that their applications wouldn’t demand 5G-level latency control, and a significant number didn’t even know what the latency difference would be.
WiFi 6 is another interesting factor. Only about a third of prospective 5G users distinguish between WiFi 6 and earlier versions, and only about a fifth know anything about the latency control available in WiFi 6. Even fewer know anything about WiFi 6 security. Given those points, it’s difficult to justify the assumption that three-quarters of enterprises with manufacturing connectivity needs would be adopting 5G; wouldn’t any serious movement in that direction compare obviously cheaper and easier-to-adopt alternatives?
This is where things like the Ananki/Rakuten announcements come in. Enterprises really want private 5G as a service or in a package. They’d love to get it from a vendor they’re familiar with, but an open approach is nearly as good according to what I’m hearing. Having easy access to a consumable private 5G offering could lower the barrier to adoption significantly, though I still believe that users will have to weigh the value proposition of private 5G for their applications before they move. Thus, I don’t think that even these recent private-5G offerings will create a private 5G avalanche.
Then why would users respond to a survey with what seems obviously incorrect or poorly thought-out comments? I’ve had many opportunities to assess survey responses, and it’s my experience that roughly a third of people surveyed will tell the agent what they believe they want to hear, or give the answer that makes them look smart and up-to-date. I’ve cited in past blogs the specific example of a third of enterprises telling a survey form they used 100 Mbps Ethernet when there was no commercial product available at the time. Survey firms also tend to go back to people they can get answers from, sometimes stretching the knowledge of the people if the new survey topic is out of their area.
The net is this. I see no indication of anywhere near the private or hybrid 5G interest that the article and survey suggests. OK, maybe my data is wrong, but I’ve gotten consistently right information from my contacts over years, often decades. I also see a lot of worrying inconsistencies in the data, places where it contradicts or conflicts with other information whose reliability is almost unquestionable, like sales.
There is a value to 5G, and to private 5G. That value, though, lies more in what applications it might enable than in what it’s going to do on its own. The difference is that things dependent on the enabling of other things are dependent on the pace that those dependent elements actually evolve. Many of those dependent elements are citing the success of 5G to pull them through. If A depends on B, which in turn depends on A, then we don’t have symbiosis, we have a circular argument.
That’s why we need to be realistic about what’s going to happen with 5G; the answer will be “not much” unless we take on the task of creating things that actually justify it.