We might be justified in calling the next four months the “Summer of 5G Disillusionment”. On the one hand, the 5G providers seem to be facing off in a big win-customers campaign. On the other hand, some recent research says that users are disappointed with the lack of new 5G applications. A good part of the problem arises from over-promotion of 5G, something that almost every player shares the blame for. Another part is that “new applications” really don’t have a lot to do with 5G per se, and the final blame point is lack of real progress in the thing that would likely generate new applications, which is carrier cloud.
There’s not much need to amplify the over-promotion point; what technology isn’t over-hyped these days? What makes 5G perhaps more problematic than most technologies is that it was promoted to cellular users and not to deep-in-the-organization technical geeks. Ordinary consumers like a good tale, but they also turn on the teller if it turns out that the story isn’t true.
The reason for the face-off among 5G operators is a good example of hype, in fact. The truth is that operators aren’t looking for 5G customers as much as for customers, period. 5G is just another competitive point that savvy marketers in the mobile operator space can leverage to try to gain market share on competitors. The notion that there’s a race to exploit 5G opportunity surely contributes to a consumer notion that there’s some special 5G opportunity to exploit, and that contributes to their expectations that some new applications will burst on the scene.
New applications for network service “burst on the scene” when a network service change bursts through a bottleneck that had previously constrained them. Most of the applications that consumers say they’re waiting for (in vain, so far) are applications that either already exist or could exist under most developed 4G LTE services. 5G offers some improvements in speed, some reductions in latency, but in the applications the article cites, it’s hard to point to specific 5G capabilities that burst through barriers. If we don’t have those applications, it’s less because of network constraints than because of application constraints.
Let’s look at the example of gaming. 5G’s capability to deliver a low-latency connection is cited as a specific reason why 5G gaming would be better than 4G gaming, but does that stand up to examination? Yes, 5G could reduce latency, but one big contributor to gaming latency is the message paths between various players and a “compositor” application that builds the visual experience for everyone who’s avatar is in the same local space. Some argue that 5G drives edge computing, which then reduces the latency of that path, but that begs two questions; does 5G actually drive edge computing, and where is “the edge” when players are scattered all over the world?
What users really want when they say “new applications” is “new experiences”, and experiences aren’t created by the network, only delivered by it. Hosting creates experiences, which means that the connection between gaming and edge computing isn’t specious, only improperly developed.
In order for 5G to promote “hosted experiences”, we have to presume edge hosting resources would deploy and would be exploited for higher-level services. That requires two specific things; an architecture to define how edge services would be presented to applications, and regulatory and business accommodations to encourage investment. Right now, we have neither.
I’ve noted in the past that one problem with 5G as a driver of edge computing overall is that 5G hews to the NFV model, which is at odds with how cloud applications overall are evolving. If we expect 5G to create broadly useful edge facilities, we have to target 5G implementations to the features that would be broadly useful, which we are not doing.
The regulatory side is even more frustrating. Every major market area has their own network regulators, and their own vision of what constitutes good public policy. The current state of all of this is a swirling mess, and in some areas (including the US) it changes as the party in power changes. The general position is that telcos, as former regulated monopolies, can’t exploit the fruits of that protected relationship by expanding the service set of the core telecom entity. Instead, they have to form a separate subsidiary. The problem is that it’s the core entity that owns the real estate in which edge hosting would have to be installed. Forming a separate subsidiary would typically mean telcos would have to share the resources with not only their separate subsidiary, but competitors to it. Good luck making a business case for that.
Right now, I see the “cloud-providers-host-5G” story line as sitting on the fault line of these two points. On the one hand, if public cloud providers did the 5G hosting deployment, whatever they did would be instantly available for broader exploitation and the creation of new experiences. On the other hand, as I pointed out in an earlier blog, the Bell-heads and Cloud-heads don’t seem to be converging on an architecture that would fulfill both 5G and experience-hosting missions. There’s also the fact that the telcos, in the long run, would have to cede a lot of infrastructure ownership to cloud providers, and pay their profit margins.
The source of 5G disillusionment isn’t 5G as much as “carrier cloud”. Users are implying that they want 5G experiences rather than just 5G connectivity. Somebody has to build them, but if a third party is going to try to build an edge computing strategy, why link it to 5G? If the telcos themselves are to build that strategy, the “carrier cloud” concept, then how do they resolve the technical and business/regulatory problems I just noted?
The problem with hype waves in general is superficiality. You can’t create a hype wave using two-thousand-word articles, but you can’t raise and resolve complex technical, business, and regulatory issues in 300-word sound bites. Companies are focusing more and more on simple promotion, which means simple hype, rather than on addressing complicated problems. Solving those problems is essential if we’re to avoid disillusionment in 5G or any other technology.
How might this all shake out? It’s hard to decide even what the “ideal” outcome would be, but if we assume that user satisfaction has to replace disillusionment, then it would appear that having cloud providers host 5G is the best path to the optimum state. Cloud providers may struggle to get their telecom-cloud-native act together, but they already have the technology needed to host experiences that would create the applications of 5G users say they crave. The question is whether cloud providers could present a compelling story to telcos and induce them to commit to the approach.
Over-promising is an all-too-common marketing strategy, but it can lead to long-term problems, and both cloud providers and network operators are at risk if they can’t come up with a strategy to deliver on 5G expectations.