An article published this week referenced a recent Hype Cycle diagram (pictured above) from the oracle of IT – Gartner. While the lede talked a lot about the apparent “death” of Fibre Channel over Ethernet (FCoE), there was also a lot of time devoted to discussing SDN’s arrival at the Trough of Disillusionment. Quoting directly from the oracle:
Interest wanes as experiments and implementations fail to deliver. Producers of the technology shake out or fail. Investments continue only if the surviving providers improve their products to the satisfaction of early adopters.
As SDN approaches this dip in the Hype Cycle it would seem that the steam is finally being let out of the Software Defined Bubble. The Register article mentions how people are going to leave SDN by the wayside and jump on the next hype-filled networking idea, likely SD-WAN given the amount of discussion it has been getting recently. Do you know what this means for SDN? Nothing but good things.
Software Defined Hammers
Engineers have a chronic case of Software Defined Overload. SD-anything ranks right up there with Fat Free and New And Improved as the Most Overused Marketing Terms. Every solution release in the last two years has been software defined somehow. Why? Because that’s what marketing people think engineers want. Put Software Defined in the product and people will buy it hand over fist. Guess what Little Tommy Callahan has to say about that?
There isn’t any disillusionment in this little bump in the road. Quite the contrary. This is where the rubber meets the road, so to speak. This is where all the pretenders to the SDN crown find out that their solutions aren’t suited for mass production. Or that their much-vaunted hammer doesn’t have any nails to drive. Or that their hammer can’t drive a customer’s screws or rivets. And those pretenders will move on to the next hype bubble, leaving the real work to companies that have working solutions and real products that customers want.
This is no different than every other “hammer and nail” problem from the past few decades of networking. Whether it be ATM, MPLS, or any one of a dozen “game changing” technologies, the reality is that each of these solutions went from being the answer to every problem to being a specific solution for specific problems. Hopefully we’ve gotten SDN to this point before someone develops the software defined equivalent of LANE.
The Software Defined Road Ahead
Where does SD-technology go from here? Well, without marketing whipping everyone into a Software Defined Frenzy, the future is whatever developers want to make of it. Developers that come up with solutions. Developers that integrate SDN ideas into products and quietly sell them for specific needs. People that play the long game rather than hope that they can take over the world in a day.
Look at IPv6. It solves so many problems we have with today’s Internet. Not just IP exhaustion issues either. It solves issues with security, availability, and reachability. Yet we are just now starting to deploy it widely thanks to the panic of the IPocalypse. IPv6 did get a fair amount of hype twenty years ago when it was unveiled as the solution to every IP problem. After years of mediocrity and being derided as unnecessary, IPv6 is poised to finally assume its role.
SDN isn’t going to take nearly as long as IPv6 to come into play. What is going to happen is a transition away from Software Defined as the selling point. Even today we’re starting to see companies move away from SD labeling and instead use more specific terms to help customers understand what’s important about the solution and how it will help customers. That’s what is needed to clarify the confusion and reduce fatigue.
The Networking Nerd 
Hey Tom,
Great article. The other thing I would add to this is the list of products that were “labeled” as software-defined but really equated to just lipstick, and sometimes mascara and blush and a quart of perfume on a very old stinky pig.
The ONF tried to give us a clear definition early on, which I still subscribe to, of what SDN actually was. ( Centralized control plane decoupled from forwarding plane with centralized knowledge of state ) but this was co-opted by the marketers and shoved down our throats with a extra scope of cloud on the side. ( Another meaningless term ).
I agree 100% that now is the point where we’re going to start to see the pretenders and the companies that customers perceive to offer limited value fall off the map and we’ll start getting down to the real players here who are doing/offering something that makes sense, offers value, is affordable and doesn’t represent a complete paradigm shift from traditional networking.
Personally, I’m a huge believer in SDN ( the orthodox definition ) as well as Network programmability. I think that applying software in intelligent ways to our current networking issues can really make a huge difference, but that’s got to be more than powerpoint and maketing fairy dust.
I sincerely hope we’re at the point where we stop talking about meaningless marketing terms and start focusing on the value that these products bring to customers, ie the problems that it’s going to solve for the engineers.
@netmanchris
P.S. For those who don’t know I work for HP and who has products in this space. I am the Lorax. I speak for the trees, not HP.
Pingback: Worth Reading: Software Defined Overload - 'net work