As we look a bit more at the commoditization arguments and realize that a pricing shift is afoot, it becomes interesting to see how the industry will fare if there are wholesale pricing changes. Ultimately, it is business reasons more than technical reasons that everyone cannot be a winner in the networking space, so let’s start there.
Now we should be clear about a couple of basic facts. In the race to size the SDN TAM, there is a lot of excitement about the creation of new markets and how that translates to new opportunities. But let us not forget that whatever new category or market or accounting rules get created, the money is still a subset of IT spend. And there is nothing about SDN that is going to generate net new IT dollars. That spend is typically fixed as a percentage of overall spend (or revenue). Across thousands of companies in aggregate, we are not likely to see a sudden surge in networking spend because of SDN.
Put differently, if we believe that overall corporate expenses are set, then for every net new dollar spent on SDN, companies have to spend one dollar less on something else. If you ask anyone in the IT infrastructure space, they will tell you that new demand is driving growth. So where are the SDN dollars coming? They are coming from existing network spend. Sure, there will be some OpEx savings, but not initially as there are transition costs when moving to a new architecture with a new set of tools.
A lot of companies are banking on a new market creating new spend. The best examples? Any of the slew of SDN software companies. I’m not talking about companies who are replacing what has typically been a hardware platform with a licensed software product running in virtual machines (think Boundary or Embrane). I am talking about others, like the group of SDN controller guys who believe that their orchestration advantages will yield lucrative returns.
Think about it this way: imagine there is a networking budget of 100 bucks. Even with a controller, companies have to build out a capable underlay. Let’s say the underlay costs 90 dollars. That leaves only 10 dollars for the controller. Which of these do you buy first? In other words, the pure-play controller guys will be at the mercy of the underlay hardware prices (which, in theory, come down because of commoditization). Their business model is predicated on the hardware guys commoditizing their stuff, and quickly enough to make room for their products. It’s no wonder why there is so much energy around the commoditization arguments – companies will live and die based on it.
I am not suggesting that software won’t be a major part here, but rather that those who have both hardware and software elements of their solution will be poised to take advantage of the pricing mix and technology advances that come with SDN. Pure software companies will struggle with how to price and sell as the hardware companies add their own controllers to the mix, and pure hardware companies will be left to compete with an increasingly difficult-to-differentiate product.
So as you hear the pitches and place your bets, it is probably worth thinking through a bit who will be around in a couple of years. As part of that, you have to ask whether their business model is likely to yield a sustainable company, or whether the actual goal is to be bought out quickly at high multiples and before ever having to prove out the business. Put simply, is the company banking on your success or theirs?