Today in Cloud

The blog, Confused of Calcutta, has been a fixture in the ‘Must Read’ folder of my RSS reader for a very long time. Typically insightful, unreservedly thought-provoking, and often beautifully written, it’s everything that a good blog should be. Its author, JP Rangaswami, also happens to be Chief Scientist at Salesforce. Over the weekend, JP posted on the distinction between public and private clouds, characterising them broadly as “shared” or “not shared.” I find myself more or less in agreement with this distinction and some of the ways in which JP illustrates it, and also tend to believe that most (but not all) private cloud implementations are a point-in-time solution that will eventually be absorbed by the public cloud. The problem I have with most of those attacking today’s private cloud implementations — and something that JP’s post does not really address — is concerned with scale. At scale (the BBC, Johnson & Johnson, Ford, the U.S. Federal Government) a “private cloud” may very well be cost effective. At scale, a “private cloud” may very well make perfect sense. At scale, a “private cloud” is most definitely shared; by tens or hundreds of thousands of people. The cloud makes sense when the resource is shared. The cloud makes more sense (economically, at least) as more people share. But that doesn’t mean that a compute utility has to be open to all comers before it can be called a cloud. Our shared public utilities in the physical world are certainly both shared and public, but we don’t assume that the inhabitants of neighbouring countries have to be allowed to hop across the border and help themselves before we can call our municipal water supply “public,” or “shared.” The rhetoric of false cloud simply doesn’t add up. It’s time it went away.