With all the talk about commoditization on one side and a brand new market on the other side, there is an immense pressure building to be able to size the “SDN Market”. I put “SDN Market” in quotes because I think it is a bit of a misnomer, but that is for another time. I suspect things will start to reach a fever pitch around April when two forces will come together: the Open Networking Summit (ONS) and earnings season.
Now for those of you who don’t know, the ONS in April is when the SDN community comes together for a conference in Palo Alto. You can expect lots of new announcements, demonstrations, and proclamations. Many will make the pilgrimage, and it should dominate headlines for a few weeks.
Now the more interesting of the two forces for me is earnings season. The large legacy vendors have been doing a bit of a dance for the past few years. On the one hand, they need to embrace SDN because it is the new thing. They need to show that they are capable of evolving their products and their business models. They need to express excitement about technology advances that will challenge their positions because “SDN is a huge opportunity”.
The other delicate side of this Tango, though, is the financial analyst community. As you know, it is difficult to talk about SDN without hearing the word commoditization. And the financial analysts have been keenly aware of this word because it could mean increasing margin pressure in a space that appears destined to compete on price more and more. With merchant silicon narrowing the gap, even for higher-end systems, and new entrants creating more competition, the last thing the financial analysts need to hear is that SDN poses an existential threat to what has been ridiculously healthy margins for more than 15 years.
Now the dynamic that this will create is a need to size the SDN market. For the SDN purists, it is about sizing the opportunity (how much money will these startups make?). For the analysts, it is about assessing the risk and tracking market share and adoption in a newly created market.
So what exactly is the SDN market and how do you size it? I have seen more than a few analyst firms present methodologies, and they are remarkably similar. The basic idea is to size the “market” (for some definition of “market”) by the number of Ethernet ports sold. This works well on one hand because measuring ports and pricing those ports is easy. But what is SDN in this context?
Where people are going to wind up is measuring OpenFlow-enabled ports on the hardware side. This is wrong on so many levels. Forget the fact that SDN is applicable outside the datacenter (which makes traditional port counts like this difficult). But obviously SDN is more than OpenFlow. And SDN is certainly more than HW. As I suggested before, we are likely to see a shift in the pricing mix, which makes any market sizing based primarily on hardware neither illuminating nor meaningful.
Now the good news here is that the commercial controller market was just killed by the Daylight project. Offering that for free will reduce to noise the contribution tied to the SDN controller market, so any sizing will be more accurate than it would have been a few months ago. Brilliant move by Cisco, but it might be the nail in Big Switch’s coffin (regardless of how they spin it).
Where does this leave us? I think the industry does need to size the SDN market – the buzz is too strong to ignore it, and it will help sort out uncertainty about which companies and solutions are gaining traction. But I think that companies that talk about their business as SDN and non-SDN are missing the point. In the rush to SDN wash this product or that solution, we have forgotten a very basic tenet of business: what you solve is more important than how you solve it. If we as an industry were not so embarrassingly bad at evolving the network and meeting customer needs, maybe the conversation would still be where it should be: doing meaningful things for customers.