CenturionStudio.it - Fotolia
Ever since the term software-defined networking, or SDN, entered into the information technology lexicon as the latest paradigm by which we should measure our networks, the result in the marketplace has more often been confusion than clarity of strategy.
Many entrants into the fray of the burgeoning SDN marketplace came with their wares, but none cast such large shadows as Cisco's Application Centric Infrastructure (ACI) and VMware's NSX. The early SDN marketplace started, then, as a movie with two stars. It has more recently become an ensemble cast, which may be too bad, because with only two entrants into the market, it would have been a fairly easy game of Coke versus Pepsi; we could all decamp to our respective safe places, and the world would keep turning. Unfortunately, with network infrastructure upgrades, it's not that simple.
Nothing ever is.
SDN is a tough concept to define
One of the challenges is SDN itself is not a product, a market or something you can go out and buy from the corner-store vendor, despite protestations to the contrary. It is an architectural model -- an idea more than anything -- for how a network should operate, and it represents one of the biggest changes in network design since we were all fighting over network and transport protocols.
Naturally, then, every vendor of networking products, no matter how niche or established, rushed to adapt their offerings to this new model. But because there weren't established standards, a la TCP/IP, which are agreed on by everyone, the SDN model quickly turned into the Wild West.
On its face, you might be tempted to say, "So what? Choice is good for the market," and move on. That would be true with a certain number of products with some agreed-upon standards. But, in this case, we have neither. We have way too many products clothing themselves in the SDN moniker, and we have no standards other than some loosely agreed-upon ideas around behavior.
Add in the increasing fragmentation that comes from many vendors turning everything from the access layer, backbone, applications, virtualization and storage, to name a few, into a "software-defined" something or other, and you have a perfect recipe to where we are now: confusion.
Understanding the next step for network infrastructure upgrades
I have spoken with many clients over the past few years, trying to understand their needs and help them sort through the competing vendors and ideas in a vendor-agnostic manner, and I can tell you the only thing that is a constant is bewilderment.
While some of those clients have come to a conclusion favoring Cisco or VMware for various reasons, and some have gone with niche players, the majority are still sitting on the sidelines trying to figure out when, or if, to buy in to someone's vision. Sociologists have coined the phrase "fear of missing out," or FOMO, to describe what many people feel when they are constantly bombarded by updates on social media platforms and run themselves ragged just trying to keep up with the proverbial Joneses.
The same thing is happening to IT departments and for much the same reason. But instead of leading to a rush to participate, it's leading to decision paralysis.
And there's this: With the exception of speeds and feeds, most enterprise networks are working just fine. They may be a bit inflexible -- particularly in light of today's standards -- a bit hard to move this way and that, but the design and function are sound. This makes it easy for IT departments to sit on the sidelines while they attempt to make sense of the new paradigms, to pull signals from the noise.
But how long will this strategy of sitting on the sidelines remain tenable? Expectations of flexibility, speed of deployment, reporting and troubleshooting are coming from all sides. Taking the time to develop a cogent network strategy is understandable, given the current state of the market.
Making the next move in SDN won't be easy
Buying the biggest products from the biggest vendors, simply on spec, isn't going to work any longer -- if it really ever did. At some point, a decision will have to be made. The question, then, is at what state does the market coalesce enough for the average organization to feel safe in placing a multiyear, multimillion-dollar bet on core infrastructure?
Given all of this, ceding risky decision-making to the status quo begins to look a lot like good jurisprudence. Buy the top of the SDN marketplace not because it meets all of your needs, but because it will likely still be around for the foreseeable future, and it probably won't get anyone fired. That isn't necessarily a bad thing -- not always: Stability and long-term support are as critical as features and functionality.
But remember that wholesale network upgrades, the kind that come with this new architectural paradigm, are neither cheap nor simple to deploy or operate. Make sure before you buy that you know what you are getting, and that you can live with it for a long time.