MWC, like all trade shows, generates an enormous amount of ink; more ink than insight in my own view, which is why I no longer go to shows (unless someone pays me to be there). One problem this generates is the predictable shifting of focus from tech evaluation to tech entertainment; people seem to adopt the same attitudes that must have prevailed in the Roman Colosseum. Another is that sensationalism can trump actual value and impact, hiding things that could be really important in the dazzle of hype-lights. Let’s cover our eyes and look at things coming out.
The most prevalent theme is “mobile traffic growth is slowing”. Well, what did you expect? There’s only so many ears you can hold a phone to, eyes that can focus on a screen, wallets that can pay the bills, and hours people can spare from things like eating, sleeping, and working, to fiddle with their phones.
What will fix it? AI will sweep away all problems, apparently. The notion that AI may be the answer is facile, of course. There is no credible indication that AI impacts traffic, based on my research and on no-attribution comments I get from operators themselves. But you need to say something, offer an alternative to “this is as good as business gets”.
Connectivity, and connection capacity, have been commoditizing for decades because there is a limit to the communications needs of consumers and enterprises, and if there are no really meaningful differentiators in a market, cost becomes the only thing you can compete on. The Internet’s price/speed curve demonstrates that people won’t pay proportionally for more bandwidth anyway, so it’s questionable whether having new traffic sources would help operators, but MWC is (like most shows) really a vendor show anyway, and more bits means more bit handling, more device spending.
The deeper truth here is that we have two issues. One is that, at the top, we’re running out of total addressable market and average revenue per user insofar as connection services are concerned. That means operators see revenue growth problems that can be countered only by lowering costs. The second issue is that new applications that could justify greater spending because they present conspicuous user/enterprise value aren’t evolving. While it’s not certain that such applications could drive incremental operator revenue, it’s pretty darn certain that they’re a necessary, if not sufficient, condition of operator revenue growth. They are likely the main potential drivers of new equipment/device spending in any case.
As I’ve said in prior blogs, you can’t push the notion that a technology shift (to 5G, to AI, whatever) will be a driver of other technology spending. Every step in a technology ecosystem value chain has to make a business case. 5G can’t drive anything, but some new mission could drive 5G and also drive whatever a 5G evolution would then generate.
Private 5G, another of the big stories of MWC, is an example here. So what if you can identify a use case for it. A “use case” is proof that something can be used in a given mission. What’s needed for private 5G isn’t that, it’s a business case, and no enterprise I’ve ever chatted with thinks that any significant number of enterprises can make a business case for private 5G. Yes, there are some that perhaps can, but not the great majority, which means that this is another story that’s intended to improve a vendor/market image but isn’t likely to do much for its bottom line.
Of course, there’s always 6G. According to Fierce Networks, “Like a lot of folks, the Wireless Broadband Alliance (WBA) would like to avoid making the same old mistakes when it comes to 6G.” Ironic, at a place and time where most of those mistakes get made. The important thing about the coverage of 6G is that it really is focusing on not repeating 5G mistakes, which isn’t exactly a pathway to a specific positive business case. The story seems to suggest that the problem with 5G was “very closed environment that creates a very huge burden” for operators to upgrade and adopt it. But what makes a new broadband generation less a burden? The only possible answer is that it grows revenues significantly more than it grows costs. Why would 6G do that? Only by enabling new missions, new applications. We’re back to business case.
What is the really biggest news of the show? My candidate is VeloSky from Broadcom. The company describes it as “a converged networking solution that enables Communications Service Providers (CSPs) to offer integrated fiber, cellular, and satellite connectivity through a single appliance”, and putting it another way, it’s a device that lets multiple kinds of connectivity combine into a single virtual service whose properties and availability are the sum of the connectivity choices.
OK, it’s clear that this sort of box has a value for operators who have multiple service options available, and that it presents a nice upsell to a remote site whose wireline and even wireless broadband service might not be fully reliable/available. It might also be a nice tool for enterprises to purchase directly, to offer a way of generating seamless backup for sites (Broadcom doesn’t suggest direct marketing to the enterprise at this point, though), and for managed service providers (MSPs) to use to beef up their value to enterprises. But does this justify it being the biggest news, at least on Tuesday? No, but there’s more to it.
One thing is that this is a platform that includes an appliance, not a chip. There’s a platform specification for it. Broadcom is providing direct operator equipment here, and introducing it in a way that links it to a specific target market. Chips are the foundation of network devices, but if network connectivity is commoditizing (if you don’t believe it is, you’re probably reading the wrong blog for you here) then having the premier communications chip player offer a platform that can be installed in a network is not only smart, it’s inevitable. It even introduces security features.
The second thing is that VeloSky plays on the VMware SD-Access and SD-WAN tools, so much so that if you search Broadcom’s site for “SD-WAN”, VeloSky comes up. The Internet is surely an example of connection network option, so could you see VeloSky as a way of using the Internet to back up an MPLS VPN? And the features of VeloSky include the ability to route traffic to a connection network based on QoS (and, presumably, other requirements), so you could see this as a way to sneak Internet into a position as a service option for branch sites. For operators, it also offers a way to offer security-as-a-service. Juniper, who had what I’ve already said was the best of the SD-WAN implementations with SSR, didn’t play the capability as strong as they should have. Did they open things for Broadcom?
The final thing is that VeloSky separates the notion of “service” from the specific functionality of “networks”. A service is virtual, constructed from features than are drawn from whatever connection network options are available, and whatever features are included in the platform. In short, it’s a provider of a virtual network, and virtual networks are way more feature-agile and flexible than real ones.
Broadcom’s announcement deserves the best-so-far award for MWC because it not only reflects the real conditions and opportunities in the mobile market, it takes a step toward defining the virtual-network framework that I think is key to the future of network services overall. It also establishes Broadcom as a player, maybe even the player in service evolution.
