Jarich: SDN - a rose by any other name… still wouldn't be NFV

Current Analysis Peter Jarich

     Peter Jarich

We've talked about Software Defined Networking (SDN) and Network Functions Virtualization (NFV) multiple times this year. While still relatively new concepts, there's general consensus that both have the potential to revolutionize the way in which we think about mobile networks and services: SDN thanks to driving service velocity and manageability, and NFV thanks to rethinking the platforms on which networks are built and how network capacity is (cost-effectively) turned up or torn down. To be sure, there are still plenty of questions around them. What use cases will drive early adoption? Are the barriers to adoption skewed towards operational or technical issues? Will SDN lead to the commoditization of routing platforms? What are operators looking for from their vendors as they investigate their options and plan their launches? It's these questions that have driven us to plan an operator survey around SDN and NFV, getting input from operators around the world in order to gauge use case expectations, buying preferences, vendor perceptions, etc. Later in the year, I hope to share some insights from the survey. In the meantime, if you want any details, don't hesitate to reach out.

Yet, if there's one thing around which there seems to be wide agreement, it's the notion that SDN too often seems to get conflated with NFV--marketing around the (for some reason) sexier SDN subsuming any message around NFV. I've fought the "what is 4G" battles in the past, eventually siding with the marketers trying to convey a simple message. Here, however, confusion between the two is a real problem. Why?

SDN is not NFV. At a basic level, SDN is about separating out the network control plane from the data plane, with "open" interfaces to the data plane supporting service programmability and dynamic service set-up. NFV, in turn, is about building telecom applications--think IMS, policy, mobile packet core--on industry standard platforms which, thanks to virtualization software, allows for multiple applications to be built on common hardware and the capacity of those applications dynamically created or torn down. Yep, these are very different things. Sure, both simplify the way services and networks are supported, but in very different ways. Going back to the squabbles over "what is 4G," I sided with the marketers because nobody ever defined 4G. There are definitions for NFV and SDN.

The Two Are Linked.  If the story ended with "SDN and NFV are different," a critical link between the two would be completely missed. Return to the idea of NFV supporting capacity allocation across telecom applications while leveraging common platforms to do so efficiently; easily creating more VoLTE user capacity, for example, as user demands ramp. It's a nice idea, but this added capacity could be a moot point if network resources can't be somewhat dynamically allocated out to the platform. This is where SDN comes in. NFV allows an operator to quickly scale the capacity of their network functions. SDN allows them to get necessary network resources out to them. 

Do Operators Care? NFV and SDN may support one another, but that doesn't mean they require one another. SDN solutions can be deployed on their own in order to support a number of different use cases--none of which require NFV. NFV solutions can be deployed on their own to support a number of different use cases--none of which require SDN. We've seen this in talking with operators actively engaged in deploying one or the other. "SDN has nothing to do with NFV--they serve different purposes." The second part of that statement is true, true enough to mirror our first bullet. The first part is an oversimplification.

The question of whether or not operators care about the linkages between SDN and NFV is more than a rhetorical one. The answer to that question ultimately determines how vendors need to market their solutions. It impacts how they plan to sell their SDN and NFV solutions, and the people within an organization they sell to. Most importantly, the answer will help determine whether or not the full value of SDN and NFV are realized.

The good news is that we're still early in the evolution--and understanding--of these technologies. Last year this time, the only people talking about SDN and NFV were the super-techies among us. It shouldn't be surprising, then, that a deep understanding of what SDN and NFV are--and how they're linked--isn't as common as it might be. By all accounts, however, operators are moving quickly to investigate their SDN/NFV options and build out their solutions. It's just that, as always, doing so driven by a faulty or partial understanding of the technologies--potentially born of expedient marketing--wouldn't be good for them or the vendors selling into them.

Peter Jarich is the VP of Consumer and Infrastructure at Current Analysis. Follow him on Twitter: @pnjarich.