Its only the second day of what is the biggest IT conference out there these days, but I think I can already see the direction that things will be going for the next couple of years. We all like to believe that we are heading onto a post-virtual "cloud" era for computing infrastructure. Certainly, if you listened to the panel participants, we are all going to be devoid of any physical computing infrastructure someday soon. It will all magically migrate to vast data centers in the desert, leaving us free to interact with it through processors in our phones, cars, watches, underwear, etc.
So then I went to lunch with one of my former colleagues from Dell. We talked through the usual pleasantries and then we got into the happenings at the show. "You know," he says, "it's still all about the box." "Yes," I replied. "Everyone is fighting about who owns the box. Nothing has changed."
The box, in this case, is your server, or your storage, or your network. By virtualizing them, we really haven't changed the discussion much. We only changed the venue. The new NSX (nee Nicira) product is an obvious play to redefine the edge of the network, to the exclusion of Cisco. Nowhere was the tension more evident than on the stage, where Cisco distinguished itself simply by it's absence from the list of partners. I've made my opinions on the networking space known elsewhere on this blog, but kudos to VMware if they can actually succeed in shaking up the space a little.
Then there's storage. If you haven't noticed, it's nothing but boxes there. Physical, ever-expanding, power sapping boxes. There, you see a whole host of players trying to collapse your storage into your server, or replace your old storage arrays with new ones that contain flash, or maybe throwing some flash into your server and then collapsing. Of course there are the products that push you to just use servers as storage arrays - VMware's distributed storage software, for example. No part of the box is more hotly contested than this one. Here, it's not clear that there is going to be a single winner, for the simple reason that there is just too much data to manage in too many different ways. (Hence, a reason to gravitate to this market for new ideas)
When you juxtapose the battle for the box against the lofty cloud talk that Andreesen and company were spouting up on the stage, its hard to draw a linear transition from the real world to the vision. Perhaps vision is just that: an aspirational goal that is not intended to be achieved. I'm not sure that I even want cloud-connected boxers. Perhaps for consumer IT, the vision is easily achievable. In enterprise IT, the legal, security, and psychological hurdles just seem to generate far too much inertia to get us there. If you are a strategist, it seems foolhardy to bet that IT will be physically gone in 5 years. Hence, we have to place our bets on some fuzzy, semi-cloudy, hybrid middle ground.
Lots to think about.