Wednesday, January 28, 2009

A stable taxonomy ... please

Cloud computing is a consequence of the transition of IT from a product to a service based economy. The concepts behind it are simple enough, and it's not even something new having first been discussed in the 1960s. However, a major problem with the cloud today is confusion.

There exists a plethora of different terms and mental constructs for the same problem. Hence I've been pushing the case for a clear taxonomy, James Urquhart agrees and so does Reuven. This industry desperately needs people to be talking from the same page, I've personally had enough of the *aaS and *OA wars.

Back in 2007, I gave a keynote at OSCON where I talked about the three layers of the computing stack (which we called software, framework & hardware) and how this affected utility computing. This was based upon much older ideas. Within twelve months utility changed to cloud, each of the layers were renamed and the stack was now a cloud triangle.

Now there is nothing wrong with this triangle, it's the same concept of componentisation. My concern is what is this going to degenerate into in a further twelve months? An n-dimensional bubblegum chart?

Whilst Randy argues against a one size fits all taxonomy, I'm not sure who he is arguing against. Any taxonomy must consider componentisation and how the shift from products to services effects all the layers of the computing stack / triangle / paisley coloured bubblegum graph or whatever we end up with. In the last year, I've seen plenty of execution in cloud computing but very few new concepts. However during this time there's been an awful lot of renaming and revisions.

We need a common language.