Revisiting Thin Provisioning's 'Firsts'

It's been a couple of months since George Teixeira, CEO of DataCore, unleashed his rabid PR team on the press, to "stop the insanity" of our insistence that 3PAR "invented" thin provisioning, or at least developed it first.

The memo I got from George's hired help back in May argued that news articles about HDS's introduction of thin provisioning was including an awful lot of press references to 3PAR, making it seem that either HDS or 3PAR had invented the technique. Before 3PAR's product introduction, DataCore maintained, it was humbly selling the very same technology without calling it "thin provisioning."

The DataCore missive included the following statement: "Where we take exception is in the 'giddiness' with which this was reported and how no small dose of accuracy and fairness has been thrown out the window in the midst of the marketing tsunami."

At the time, we paid no attention. After all, why was DataCore raising a fuss, when for five years it never mentioned "thin provisioning" in any of its briefings to us? Flaks, begone!

But this week, I had a chance to reconsider. I caught up by phone with George, and separately with David Scott, the CEO of 3PAR, and asked them to explain. What was all the fuss about?Well, when someone's right, they're right. It turns out that the term "thin provisioning" was first coined by the marketing team at 3PAR, but the principle of using capacity on an overbooked or "as needed" basis started way back before present storage time began. StorageTek used an early form of the technique in its Shared Virtual Array (SVA) system for mainframes.

Following that prehistoric beginning, both Scott and Teixeira agree that DataCore released thin provisioning in January 2002. (See DataCore Ships SAN Management.) 3PAR announced plans for thin provisioning in June 2002, followed by implementations in the first half of 2003 and an official shipment announcement in June 2003. (See 3PAR Debuts 'Thin Provisioning'.)

Still, it can be argued that both companies were "first," since -- in Scott's view, at least -- DataCore was the first to deploy thin provisioning as software for an appliance, while 3PAR deployed it in a proprietary array.

George Teixeira isn't conceding any ground on that one, though he credits 3PAR with popularizing the term that educated the world about what both vendors were doing.

So, after considering it all, we've made a few changes. We've corrected our online record, wherever we mentioned or even hinted that 3PAR may have been first to bat with thin provisioning. (See Pillar Pushes Provisioning, Capacity, 3PAR, NetApp Join Ranks, The Skinny on Thin Provisioning, Thin Is Definitely In, Hitachi Bulks Up, and EqualLogic: Thin Is In.)We didn't do this for anyone but ourselves and our readers. After all, both vendors seem to be trapped in their own spin about this, and spin shouldn't be a consideration when it comes to talking about the realities of storage products.

On the upside, the war of words here was over proven products and technologies that have stood the test of time -- so far, anyway. But we somehow got caught up in it, and for that, we stand corrected.

Mary Jander, Site Editor, Byte and Switch