Tag Archives: aas

Simon Wardley responds to my “Cloud Relationship Model” article

Simon Wardley, Software Services Manager at Canonical UK and noted Cloud Computing expert and public speaker, responds to my article “Cloud Relationship Model“, with specific mention of my paraphrasing of a talk I saw him give; and thereby explains the history behind the “*aaS” double entendre.

Seriously though, the number of heads on the “*aaS” Hydra continues to grow, and Simon’s comment soon focuses upon the genuine need for a stable and standardised taxonomy, something I agree with wholeheartedly, along with a little bit of temperance and cool-headedness when it comes to thinking up and announcing more “*aaS” classifications…


Alas I cannot claim that joke to be my own.

Back in early 2007 when I gave a talk at ETECH, I described the changes in the industry as a continued shift of the computing stack from a product to a service based economy.

At that time I often categorised the computing stack into three layers – software, software platform and hardware (back from 2006). Whilst I had made comments that the software layer was really about applications and therefore SaaS should have had a more unfortunate acronym, this was not my true crime nor the origin of the joke.

The problem was that whilst SaaS and HaaS had been in common usage, the mid layer was known as SaaS Platform. This neither fitted neatly into the naming convention nor was it correct, as this layer of the stack contained many framework elements. So I started to describe this as the framework layer and FaaS seemed to be the obvious choice.

Hence at OSCON in Jul’07 I described the stack as a trio of SaaS, FaaS and HaaS.

Robert Lefkowitz (in a later talk at OSCON’07) warned us that this trio of “oh so wrong” nomenclature would lead to a whole lot of “aaS” and hence the joke was born.

Well Robert, as always, was spot on. In the last few years we’ve seen a plethora of different “aaS” terms (at last count it was about 16, including multiple versions of DaaS). The last few years has seen a constant exercise in revisionism.

Whilst the distinction between the layers of the computing stack is valid and meaningful, especially in context of the shift from products to services, what is not meaningful is the constant creation and recreation of terms.

Fortunately we now seem to be settling down to a three layer stack of application, platform and infrastructure – though I’m sure there is going to be more arguments.

This is why I argue the one thing we need in cloud computing today is a stable taxonomy.

Good post by the way.

Simon W