It was not the iPhone.
It was Red Hat's purchase of Gluster.
Why should a $136 million corporate purchase mean more to the world than the latest and greatest phone, the first Apple product to be introduced in the post-Jobs era? Because Red Hat has just stuck a fork in the biggest obstacle to mass use of the cloud, the cost of data storage.
As Glusters CEO Ben Golub explained it to me in an interview after the announcement “Customer data storage needs are growing 40-50% per year, but the cost of proprietary data storage clusters are dropping only 15% per year. Meanwhile, at Fry's, the cost of a 1 terabyte drive is down 90%.”
While clouds let Google run on commodity PCs, in other words, most companies won't trust their data storage to commodity hardware. Instead they're buying specialized data clusters from EMC, from HP, or from IBM, and losing many of the cost-savings clouds were meant to provide. Clouds aren't doing it for them, in other words, as long as storage remains a cost bottleneck.
As a stand-alone company, meanwhile, Gluster lacked the sales and support infrastructure needed to reassure these customers. “With Gluster, and Red Hat, we can deliver many times the performance and availability of what they'd expect at a fraction of the price,” concluded Golub. And he's right.
What's it mean when you have an open source cloud infrastructure with open source data storage? First it means more clouds, and bigger clouds, faster than we might have expected. Second, it means the time when clouds have to compete with one another on cost comes closer.
We're still at what I call the “Oprah” stage of the cloud game. You know, “you get a cloud, and you get a cloud, everybody gets a cloud.” Everyone can claim a cloud strategy, and everyone does. All the big enterprise guys – Oracle, IBM, HP, EMC, etc. – claim to have a cloud strategy. They say they've got cloud handled. But they're not going to be called on that so long as clouds are something for the future.
Once clouds are ubiquitous – and virtualizing storage as well as servers makes them so – then clouds gain power, clouds gain market share, and clouds start to compete with one another on the basis of cost. People are going to start comparing clouds apples-to-apples, and the cheapest infrastructure is going to win.
This will be very bad for companies that claim to have cloud solutions, but are really just sticking “cloud” on their current products without passing along the savings. It's the best move Red Hat could have made against arch-rival VMWare, because VMWare's EMC parent suddenly turns into an anchor rather than a float. EMC, you see, is in the business of proprietary data storage.
It's also important for consumers, much more important than a new gadget you're going to throw away in a few years. Because now clouds will be able to fulfill their promises, even Apple's cloud, of unlimited free storage of all your stuff, unlimited processing on any idea, at a cost as close to free as possible.
That was the actual dream of open source. It's gotten a lot closer with this deal. And clouds will remain relevant long after the iPhone 5 has gone to the recycler.
