There are lots of technology “advances” that ride the edge between forward-looking and hype. Private wireless is one of them, most conspicuously in 5G and in discussions about 6G. Sadly, much of the dialog has really been driven by vendors trying to keep the 5G hype alive. Nokia, who’s been perhaps the most visible vendor proponent of private 5G, recently tried to debunk rumors it would exit the space.
Telcos may now be looking at playing some role in private 5G, according to THIS article, but do they have a chance, or even a plan? If they do, I think they’re following the same road that vendors like Nokia have, and almost every wireless vendor admits that private 5G has been very difficult to promote successfully. Many tell me that they’re really not even sure how to promote it. Well, I think they’re all missing the point.
If you look at private 5G (or 6G) as a kind of WiFi alternative, my enterprise contacts say that the value is limited to complexes of facilities that cover blocks or even square miles. Based on this, you’d expect the focus of opportunity to be in transportation and utilities, which is how most private 5G initiatives have been targeting. I think that may be the wrong place.
OK, indulge me here. We need to have more benefits driving changes in services, networks, and IT spending. Absent that, everything will focus on orderly modernization that does more for less, meaning total commoditization. Talking about 5G or 6G or AI or quantum computing as a driver is skipping a step; what makes the business case for these technologies? It’s the missions, the applications, that they open up. If we want revolutionary results from any of these things, what we need is revolutionary business cases. WiFi replacement isn’t that, so we need to think deeper.
Hence, indulgence. I’ve blogged before, often, about real-time services. We’ve only managed to empower about half the workforce with technology, because the other half can’t sit at desks and use computers. The barrier to doing that is the complexity of the mission, the size of the ecosystem of products and services that we’d need.
What we need for real-time services is first and foremost an application, and second a viable evolution from business-case low apples to something that fully realizes their potential across the market. We have real-time applications today; enterprises in many verticals have “local” process control applications that are based on self-hosted-close-to-the-process edge computers. These are often specialized devices, using custom hardware and even a custom operating system and middleware toolkit to manage event-based rather than transaction-based missions. What would be nice would be to figure out a way of expanding these to create an evolutionary bridge that full exploitation of real-time computing could cross.
Private 5G/6G is not the entire bridge, but it is a critical element in it. Real-time means real-world, and so expansion of current applications means introducing a larger physical footprint for the activity. Private 5G/6G can easily evolve to, and even mesh with, public 5G/6G services. This would allow enterprises to first expand the geographic scope of current real-time applications, and then build new ones on the same network framework. The private 5G/6G service could cover a large multi-building complex, and when augmented with public services, embrace related complexes as well.
Most real-time systems are hierarchical, meaning that there are layers of intelligence. If you use a smart thermostat, you’re using such a system. The thermostats have a local intelligence that allows the devices to be controlled physically, and also supports a schedule of operations to run things for you. The remote intelligence, usually linked to your phone, lets you control and schedule remotely. We can expect that the evolution of current process control missions would also be hierarchical; local intelligence would handle the most time-sensitive and high-availability pieces, and subsequent layers would handle things that aren’t as demanding. For evolving enterprise real-time applications, we could expect “local expansion” to be supported by application elements hosted in a user-owned complex that’s part of the expanding system, but eventually we’d see an opportunity to cloud-edge host some elements. Edge computing could evolve much as cloud computing has evolved, to offer better resource economy for applications where average offered load can’t fully utilize what’s needed during peak periods.
There’s both good and bad news for many of the prospective IT/network stakeholders in all of this. The most obvious bad news is that technology advances can only drive massive spending if the business cases can be made throughout the ecosystem needed for the advances to deploy. AI can’t change things unless AI can change business cases. So also for 5G/6G. The most obvious good news is that this paradigm, while perhaps not as newsworthy and satisfying of desires for instant gratification, is a proven model. We’ve done this before, so we know it can be done.
The challenge here is that evolution isn’t revolution, so instant results cannot be expected. For vendors, and telcos, facing a combination of higher costs and slower revenue growth, there’s no prospect of immediate relief. In fact, there is a real risk that some more aggressive player will take a risk you’re unwilling or unable to take, and so gain control of more of the real-time pie than they might have had, than you will have.
This all comes to focus when you consider the role that the hyperscalers, the cloud providers, might play. Real-time today is not a cloud mission. Real-time today isn’t a telco mission either; all the telcos have done is propose service features to connect still-hypothetical real-time edge hosting. The cloud providers, via tools like Amazon’s Greengrass (Google and Microsoft have similar tools), which is a middleware toolkit to unify cloud and self-hosted edge features. Obviously, these could become feature conduits to create real-time processing capabilities that will eventually be offered in the cloud. All three providers already position “IoT” versions of their tools, and this could mean that enterprises build applications on a platform that self-integrates with cloud provider edge services, where they are financially viable. Google and Microsoft are particularly aggressive in linking their edge IoT stuff with enterprise IT tools and practices, making it easier for enterprises to bridge to a highly distributed deployment via these tools.
This isn’t automatically a big new lose for telcos, but it’s a big new risk. Telco complaints about OTT disintermediation reflect the truth that any network-connected application tends to first value the requirements for the application and only then for the network. The real-time edge, controlled by cloud providers, might still need some real-time-capable service features, but cloud providers would likely work hard to ensure that the layer of processing hierarchy above on-premises was in-cloud, meaning that they’d likely work to reduce the need for an edge hosting service. Telcos would then either lose an opportunity to exploit their geographic spread to host resource pools, or find the value proposition more challenging.
Focusing on 5G/6G should include focusing on how to maximize the telcos’ role in what will justify those services, and private 5G will make this “should” into a “must”. Will telcos be ready for that, or will they get caught up in the “G’s” again? We’ll look at this further tomorrow.
