In a previous post, I made the potentially controversial comment that SD-WAN will ultimately be a table stakes requirement for all (or at least most) connectivity solutions. And when it becomes an add-on to base connectivity, it will be seen less as a premium service and more as a component of a commodity service.
It is worth pointing out that this doesn’t imply that SD-WAN is not important. To the contrary, it elevates SD-WAN beyond a niche feature to a nearly ubiquitous element of all distributed enterprise connectivity solutions. That said, when SD-WAN becomes a dominant feature in all offerings, it changes the criteria for building out branch connectivity.
SD-WAN value accrual
When the service is the thing that is driving the purchase, then the value will accrue at the service. In vendor speak, this means that if SD-WAN is the hot thing, then SD-WAN is the thing that attracts the value. During this phase, the industry will react by building SD-WAN, bundling SD-WAN, marketing SD-WAN, and ultimately selling SD-WAN.
But when the SD-WAN functionality becomes a foundational element, it is no longer worth calling out separately. Using a somewhat tired networking analogy, imagine a car. Every car has airbags. Where airbags were a meaningful differentiator when only a few cars had them, today they are just a line item that everyone can claim. This doesn’t mean airbags are not important, but it does mean that anyone trying to charge extra for airbags is going to have a hard time. Indeed, airbags are just part of the car—like tires or a steering wheel or 7 cupholders.
Similarly, pricing will start high on SD-WAN but eventually drop. And this is true for carriers who offer SD-WAN as a connectivity service as well. As an early draw, it has power. As a later requirement, that power shifts from the ability to attract users to the ability to shed them (if it is not there).
When pricing drops, where will value accrue?
When airbag functionality becomes part of the car, then the value accrues at the car. When SD-WAN functionality becomes part of the underlying branch connectivity platform, the value will accrue at the platform. For services, when SD-WAN functionality is just part of a connectivity service, the value accrues at the service.
In fact, in the networking instance, this is even more true. Where cars are mostly static (you get fixed functionality for most buyers), network devices are adaptive. Vendors can breathe more life into them by releasing additional software features. This means that the leading service will actually change over time. Today, it is SD-WAN. In a year, it might be something different (location services in the branch, or dynamic remediation for application SLAs, or improved security functionality).
The question then shifts from: do you have a particular service, to can you support any number of future services? This is a platform question. Basically, is the platform extensible, and how scalable is it to address future needs that might not be well articulated at present?
Extensibility and longevity
I would argue that there are two dominant traits in a solid platform. First, can the platform do a lot of stuff? Put differently, how extensible is it? Does it support third-party innovation? Does it work at different performance points so as to be useful in different environments? How well does it integrate with required surrounding tools (like provisioning or billing systems)?
Second, will the platform endure? Does it have enough resources (CPU and memory) to expand over time without needing to be replaced? Does it have support for key technologies that are likely to play a role in the future?
If the platform is not extensible, it will be limited in what it can do. If it is not designed for longevity, the lifespan will be limited, and the underlying platform investment will be short-lived.
Connectivity services and the platform
The networking industry has been looking for the elusive killer app for a while now. I personally won’t profess to know what that app is going to be. And I would venture that most people are in a similar position. In the absence of perfect information, the best we can do is build out a platform that can support as wide a variety of options as is financially and operationally practical.
So while the discussion today is SD-WAN, the decision-making criteria need to extend beyond just the service. Decision-making needs to consider the underlying platform, because that is likely to be the lasting indicator of long-term value.
The bottom line
SD-WAN is absolutely crucial. While I think there are business model challenges based on how services are consumed and monetized, I have very little doubt about the role of the technology. This is a big reason why Juniper launched our SD-WAN solution last year. Where I think much of the industry is focused on the service, I suspect the long-term branch connectivity value will be a combination of both the individual services and the underlying platform (using platform as more than hardware here).
To drag out the car analogy again, if you invest in a superior airbag but have a shoddy car, you won't make it far. And having a great car without working airbags is equally damning. You need both. And buyers need to examine both.