5G, so current stories say, is floundering. In particular, they say, the 5G Standalone or Core implementation is floundering. I think that statement is false in that 5G Core never had a chance at all, from the very first. Wishing won’t make it so, and telcos need to understand that there’s been a fundamental shift in the market. Consumerism in data services, in the form of broadband Internet access, has killed the traditional settlement-based service model. It also killed the entire paradigm on which telco planning has been based, and it’s that impact that we’re now starting to recognize. Even telcos tell me that they’ve only begun to try to accommodate a troublesome future.
Back in the early 1980s, I was doing a fair amount of consulting to telcos. The Verizon CTO dropped into my office to chat at one point, in fact. This was the time when deregulation and unbundling was coming in, and one thing telcos really worried about was the fact that fiber technology was already creating capacity that was much cheaper, and phone calls used 56 kbps of bandwidth. How, they wondered, could they make consumers spend more on non-voice services? Things like ISDN came along to offer “data” capability, but ISDN was a point-to-point technology and users were hardly into planning for pairwise data connectivity. It took the Internet in general, and the web in particular, to change that.
Yes, we all know that broadband Internet has been transformational. Many know that the average household can buy, for at the most a couple hundred dollars a month, an access connection that faster than any business could have obtained in the 1990s, and the best they could get would have cost about $8,000 per month. At one point in the 1990s, I saw telco statistics that showed that the fastest generally available access line, a 45Mbps “T3” trunk, was used at fewer than 3,000 sites. Today, we have more gigabit consumers than that. But this isn’t the big transformation. That honor goes to the fact that consumer broadband was opportunity driven and not offering driven. We built it not because “they will come” but because they were already coming. Demand had reared its (for-the-telcos) ugly head.
Up to this point, we were supply-constrained. People wanted to talk to each other without having to climb fences between their properties. The Bell System revolutionized (well, slowly) the way we communicate, but not by teaching us something new. Instead it facilitated our doing what comes naturally, by supplying something more practical than tin cans and strings. Supply didn’t develop the demand; demand encouraged the supply. That was the revolution that should have mattered, the one that said that the Field of Dreams age of telecom was over. Telcos didn’t get it. In fact, they viewed broadband Internet as something their supply plans drove, missing the fact that the demand was already there and dial-up Internet had exploded.
So what does this have to do with 5G? Well, natural-think supply-side thinking would lead telcos to believe that applications would develop when in truth nobody had any real idea how that would happen. In fact, I think it should be clear that broadband Internet worked against any chance of 5G Core, meaning network slicing, creating significant revenue. Network slicing lets you divide the network into virtual subnets, each independent of the other, and each with specific QoS guarantees. On one hand, this sounds like a great strategy. On the other, it poses three questions.
The first question is how this intersects with net neutrality. The Internet is not only the primary consumer communications service, it’s also the primary element in corporate networking. Online shopping and sales, access to company applications, cloud computing, and pretty much everything but fixed-site VPN connectivity is really an Internet service on one side at least. Could you offer QoS on the Internet? Sure, if you didn’t charge for it. Otherwise you have to contend with the shifting sands of regulatory posture with regard to the Internet.
Second point; new applications for slices are totally unproven. Could IoT use an independent network slice? Sure, if it weren’t for the fact that IoT connectivity is looking just as much like a financial black hole as 5G Core. I’ve blogged about the challenge that operators face in trying to view IoT as a connection service opportunity, and there have been plenty of media stories that cast doubts on that course of action too. In fact, IoT is probably the poster child for the failure of supply-side thinking. Connectivity doesn’t constrain the pace of IoT, applications and business cases do.
Final point; the only undisputed values of slicing are security and QoS. QoS commitments are useful when best-efforts isn’t good enough, and we’ve got best-effort delivery of HD and even 4K video content. As far as security goes, we use the Internet for exceptionally sensitive missions, by adding on security/encryption. How many breaches of security do you think emerge from the network core? Inside the core, all the traffic confounds malefactors’ efforts to even finding something valuable. We’ve made the Internet good enough, at the very least, to prevent anyone charging much of a premium for slice services. Thus, there’s little chance of decent ROI.
I think that the important point here is that new telco services succeed when they fit into the market at a point where there’s what we always hear called “pent-up demand”. It’s not a Field-of-Dreams picture, because users are already driving up that dirt road through the corn. All that’s required is to open the gate and let them in. 5G is no such thing, it’s an attempt to justify something that should always have been considered nothing more than an orderly evolution of mobile services. The 3GPP made a mistake by trying to pull in features that were justified only by suppositional missions.
Telco intransigence on the issues relating to exploiting real demand opportunities isn’t the only problem here, though. It’s fair to ask whether there are any “real demand opportunities” to exploit. At the minimum, I think that we’ve picked most or maybe even all of the low apples. I’ve been a fan of the “digital-twin metaverse” as a means of framing real-time applications, including social media and gaming, in a common technology base. Even if telcos believed in that, and even if I’m right in the opportunity it presents, the problem is the sheer number of technologies that would need to be advanced to make it work. Could telcos advance them all? Almost surely they could not, but I remain convinced that what telcos now need to do is to get in touch with their demand side.
Industry forums are potentially powerful tools in building technology consensus when the community of players needed to support something is large and diffuse. Right now it would be possible to build such a forum around a credible demand source like my proposed digital-twin metaverse, but over time we can expect that the space will be filled by random development. Yes, that will eventually create the fusion of technologies we need for telcos to exploit, but it will take more time than telcos likely have. Continued profit pressures will force things like subsidies, and as I’ve pointed out in private blogs, that sort of solution hasn’t played well in the market so far. It would be better to do the right thing while it’s still on the table. 5G wasn’t it. 6G won’t be. It’s now a matter of demand recognition.
