“The tusks which clashed in mighty brawls
Of mastodons, are billiard balls.”
So the poet Arthur Guiterman tells us. Time marches on. The days when enterprises built their own networks using routers and digital trunks are, for all practical purposes, gone. Thirty years ago, more than three-quarters of enterprises said that was how their wide-area networks were built, and so the "service" they consumed was the digital pipe. Remember T1 and DDS? But time marched on, and today only 18 of 294 enterprise IT professionals who provided me with information on their WAN in the last six months said they used such digital-pipe services today, other than for access. It's all about the VPN, and 224 of those enterprises think it will be that way for at least five years.
The interesting thing is that five years before the great VPN transformation, only 9 of 174 enterprises thought they'd be abandoning digital pipes. What revolutionary stuff are we ignoring today that will come along to change our network future? Sure, SD-WAN is expanding, but it's not a revolution. Here are the three things that 70 enterprises - the ones that didn't think VPNs are their future - are watching and expecting will bring about major changes as soon as next year.
Cloud-first invisible network
We could call the top change on the list, cited by 68 of the enterprises, the "invisible network". For this group, the top new service concept is…(drum roll) no service at all. Envision an enterprise that uses the cloud as the front-end technology for customers, partners, and workers alike. Everyone uses the internet to get to the cloud, and the cloud connects to the data center. No remote office VPN connections, no SD-WAN, no nothing. Of course, there is a "something." Two, in fact. Part of the old VPN mission is served via Internet access, and part lives inside the cloud provider infrastructure.
Of the 224 enterprise IT leaders who see VPNs stretching out into the dim future, all but 111 think the big problem with this invisible network approach is the reliability and performance of the Internet. The remainder is split among those who say the transition to this model would be "too disruptive" (49), those who say that the underlying cloud-everywhere transformation isn't imminent (48), and those who have compliance or regulatory concerns (15). The 68 who believe in the invisible network think that internet QoS, fears of "disruption," and slow cloud transformation don't hold water, but they concede that new governance policies might be required. Interestingly, of the 224 who say that VPNs will live for at least five years, 201 concede that they won't live for ten.
The undernet relies on 5G network slicing
Obviously, the invisible network option doesn't do much for service provider revenues and profits, so it would be logical to expect a response from the providers. 51 of the enterprises think that providers' response will take the form of what we could call the "undernet" (well, why not; some non-poet already said that below "middleware" was "underware"). This notion combines a couple of concepts that are popular with providers and a couple of concepts users are interested in, to create what's effectively a network that's linked to but underneath the Internet.
In the undernet, 5G network slicing concepts would create a separation of traffic, a subnet where better QoS would be offered. All business traffic wouldn't necessarily ride on the Undernet; both operators and users think that it would be focused more on employee access and high-value applications. The service would look like network-as-a-service (NaaS) in that you could target Undernet handling at specific applications, users, or application/user sessions. What was handled that way would be charged for premium handling, and the NaaS steering mechanism would also provide built-in security.
Given that the biggest objection to the invisible network is internet QoS, the undernet option could not only be an independent service but also could facilitate a shift away from explicit IP VPNs. Some service providers who tell me they're looking at the idea see this as "subducting an MPLS VPN". However, operators face real risks with the approach. One is that it could foster abandonment of MPLS VPNs in favor of a service model that, because it's new, might have to be priced lower. Another is that net neutrality rules in various markets could bar any attempts to offer QoS for a price. Even where they don't, we all know how quickly these rules can change with shifts in the political winds.
Stranger network via alternative fiber access providers
The third potentially revolutionary service option for 2024 is the "stranger network." Enterprises know that network services can be divided into "access" and "transport", and that in the age of the Internet the transport piece of almost every network service is an IP core network that carries Internet traffic. The access network, for business services, is usually provided by a telco or cableco, but over the last decade there have been a number of new access providers, primarily offering fiber connections. Think Google. All of the enterprises I've talked with are aware of these "strangers," and 44 said they were looking at them in 2024, primarily because they were on the average just over half the cost of telco or cableco fiber.
The biggest problem with these players, according to enterprises, is their relatively narrow geographic scope. Most can cover only a single city or county, so enterprises would have to build a network from as many as a dozen different ones, and even then would likely have to rely on their traditional access provider for half of their connectivity, or more. A possible solution to this problem is the use of a managed service provider (MSP) as the access integrator. While most MSPs have focused on internet-and-SD-WAN services to enterprises, eleven of the enterprises told me that at least one MSP had offered them alternative fiber access providers, and four said that an MSP had identified over fifty such providers they could draw on.
MSPs have their own competitive challenges, too, which could force them to think about taking a bigger stake in the Stranger Network opportunity. In the past, SD-WAN services were offered primarily through MSPs, but now most telcos now offer SD-WAN services, and they represent the fastest-growing segment of the SD-WAN market. MSPs need a revenue kicker, and promoting Stranger Networks could be the one.
The strangest thing about stranger networking is that it would shift the notion of "service provider" to be more about service guarantor than provider. Break up networking into a bunch of little federated pieces and there is no big provider to lock you in. You have local facilities, joined to be your network by a support umbrella. Could this encourage municipal fiber, alternative players like Google or Microsoft or Amazon? We have perhaps five thousand utilities in the US, providing gas, electricity, and water. Might they all deploy their own fiber? And remember that despite all the independent electrical utilities, we still have a grid. Access and transport, stranger setwork and the Internet? Stranger things have happened.
You can probably see common threads here. All these off-the-wall concepts are off-the-wall because they break up, in some way, the core of the "service provider" concept, not by eliminating service but by disconnecting it from traditional mechanisms and business models. They're off-the-wall in a business sense, delivering connectivity as always, but through different sources and at a lower price.
So, do enterprises believe that the service provider businesses of the past are under pressure? Of those 294 enterprises, 227 said they believed that in five years, the business of providing network services would be radically different. Of that group, 201 said they believed the signs were already visible, and 218 said that they're either visible now or would be visible in 2024. Maybe off-the-wall isn't that far off after all.