I went back over some of my own writing a decade or two ago, and it made me wonder how much we could hope to uncover about the future of network infrastructure for service providers a decade or more from now. Everyone loves transformations; they generate interest for us and clicks for publications and advertisers. Can we expect one?
One thing we can say confidently is that there’s still more provider interest in mobile services than in wireline. That’s probably a good place to start, so will mobile drive the future network? I think it depends on what part of the network we’re talking about.
The recent interest in satellite-to-4G-phone service illustrates that one force in mobile networks is the desire to be connected everywhere. There are two prongs to this particular force; rural broadband and remote in-touch services. The former relates to the desire to project reasonable broadband quality to locations that are difficult to serve with traditional wireline or even FWA technology. The latter relates to both emergency calling in the wild, and simply keeping in touch when you’re kind of off the grid.
Another force in mobile networking is the presumption that broadband capacity limits the ability of the Internet and OTT ecosystems to provide new services and features. Thus, having more mobile bits to play with is inherently a good thing, which is why so much of 5G hype centered on the greater bandwidth it could provide. Pushing more bits requires fundamental changes to not only mobile infrastructure but also mobile devices, and this of course enriches vendors.
Wireline, as I’ve pointed out, is generally less interesting to operators than mobile services, but there is a third force in this space, created by the combination of a gradual gain in the use of the Internet as a transport media for VPNs (via SD-WAN or SASE) and the uncertain regulatory status of premium handling features on the Internet itself.
The future of service provider networking likely depends on how these three forces work for and against each other in the coming decades.
Let’s assume that the “everywhere” force dominates. This clearly further accentuates the interest in and profit from wireless, and likely focuses a lot of attention on 6G. Given the almost-universal (and rightful) discontent operators had with the ROI of 5G, we can expect this attention to take the form of explicit ROI pressure by the operators, pressure that will demand that either a convincing and significant new service revenue stream can be identified, or that the cost of the 6G transformation is minimized. Operators are skeptical about the former (once 5G-bitten, twice 6G-shy) and so are already saying they want 6G to be a software upgrade wherever possible. That means that any significant new equipment revenue would likely be limited to the radio itself, perhaps not even including other elements of the RAN subsystem.
I would expect that this would tend to direct 6G to “capacity” in some missions (like FWA) but perhaps more to the integration of satellite services and mobile services, improving on the symbiosis with 4G already coming in satellite broadband. There is potential for this to create a value-add (meaning incrementally profitable) service, which would help 6G’s credibility considerably.
Anything that makes FWA potentially better, particularly things that might make the number of customers per unit area and the per-customer capacity higher, would shift deployment focus more toward FWA. The limitations in mobile devices to meaningfully display content with resolutions above full HD (1920×1080) make it hard to justify faster device-mobile connectivity, but in FWA missions even 8k video is a reasonable longer-term target.
Universal connectivity might also come to mean IoT connectivity. For IoT to step up, it seems essential that it be both inexpensive enough to adopt for basic sensor missions, and widespread enough to accommodate the sort of mobile IoT applications already emerging in transportation, even personal vehicles. Might a universal 6G mobile-and-satellite combo be a way to go there? Perhaps, if it could be structured to be inexpensive for the real missions it targets, and have limited or no risk of being arbitraged for missions served already by higher-priced mobile services.
The big risks in this evolution overall are the proven inadequacy of standards initiatives to address real-market conditions, and the risk of vengeful vendors pushing back on a process that leaves them limited opportunity. One of the largely unrecognized impacts of a software-6G framework is that there is virtually no opportunity for any vendor other than one of the wireless infrastructure giants of today (Ericsson, Huawei, Nokia) from gaining much traction. This is the group that has dominated 5G despite operators expressing a desire for an “open” solution, and so whatever 5G technology has deployed and is expected to be software-upgraded to 6G (with perhaps some new radio stuff injected) comes from these vendors. Operators admit that they don’t see much chance of their buying 6G software from a new vendor to go into 5G hardware from one of these three. “We’d be multiplying integration and finger-pointing when we want to limit it,” one told me.
All of this tumult tends to collide with the goal of operators to converge all their services on a single infrastructure. This collision is created in no small part by the nature of the Internet’s bill-and-keep, net-neutral approach to services. Mobile services, from Release 5 of IMS two decades ago, have allowed for special handling of sessions to mobile devices. Might this encourage operators to think of their access networks as service-neutral, their core as an extension of “the Internet”, and try to push services out into the access side to avoid colliding with neutrality rules, which blow in the political wind? Could IMS, network slicing, or both somehow extend into wireline access? Could 6G help with that?
Another possibility some operators cite is repurposing the current MPLS VPN model. MPLS VPNs are created with substantially the same resources as operator Internet backbone services, and they don’t get neutrality push-back. Suppose that MPLS was used for what we might call a “SD-WAN-net”, a backbone where business services run, separated from each other by the SD-WAN level of technology and from the Internet by MPLS? Would that still pass the regulatory sniff test? If it did, might it meet up with an IMS-ish structure in the access network of both mobile and wireline? Might something like that also support IoT? I think this is the question we need to be looking at if we cast out eyes two decades forward to look at service provider evolution, so I’ll touch on it later this week.
