History and other reasons why defining Cloud Computing is irrelevant

Despite the volumes written about cloud computing over the last couple of years, a ubiquituous topic is still defining and explaining the concept. The absence of an agreed upon definition is seen by some as a problem and an indication to not believe the hype. In late 2008  Andy Isherwood, VP of software services at HP in Europe, expressed his sceticism:

“A lot of people are jumping on the bandwagon of cloud, but I have not heard two people say the same thing about it”

Another known cloud critic is Larry Ellison (though he seems to have re-evaluated his opinion), who also uses the lack of a strict definition as an argument why there cannot be any substance in cloud computing.

“We’ve redefined ‘cloud computing’ to include everything we currently do. So it has already achieved dominance in the industry. I can’t think of anything that isn’t cloud computing. [...]

Then there is a definition: What is cloud computing? It is using a computer that is out there. That is one of the definitions:  ‘That is out there.’ These people who are writing this crap are out there.”

However, as Simon Wardley illustrated at the OSCON ’09 conference, a phenomenon does not require a consensus definition in order to have great impact. There is still no definition of the Industrial revolution, despite the centuries that has passed and its indisputable importance.

The roots of the term “cloud computing” is easily understood. Most people have drawn, or at least seen, a cloud-shaped object object on a whiteboard, symbolizing the internet and how it abstracts the actual intricate network paths between computers communicating over it. The essence of cloud computing is also easy to understand, almost intuitive: computing  that takes advantage of the cloud – abstracting away parts of the underlying processes, physical resources or complexity. However, efforts to specify exactly what that means have been many and they have most often fallen short.

Some have been focused on delimiting the concept with a short paragraph specifying what is cloud computing and what might seem to be cloud computing, but really is something else. One example of this is a study by McKinsey & Co that roughly concluded that cloud computing is similar to Infrastrucutre as a Service, but with additional criteria to be met.  Other efforts define the concept has taken a more inclusive approach. A good example of that is the US National Institute of Standards and technology (NIST) who have come up with a model illustrated below.

NIST Cloud Computing model composed of: five characteristics, four deployment models and three delivery models, based on this document at NIST.gov

Although a narrow and strict definition might be tempting, the drawbacks are obvious: what about everything that is excluded despite it being intuitively understood as cloud computing? Should there be another, new, concept to describe them? A strict excluding definition merely pushes the fuzziness onto another definition. A wider definition, e.g. NIST’s, is more useful and closer to getting the job done, but since it encompasses so many different phenomenon, it does not always have a meaning with enough precision. Should there instead be another difinition, stricter than the NIST alternative and more inclusive than what McKinsey suggested? Probably not. In fact, the question “What exactly is cloud computing?”, is probably irrelevant. References to “the cloud” are seldom perceived to be ambiguous and when used as an adjective “attached to another more specific terminology”, e.g. cloud infrastructure offerings or cloud storage security, the term provides meaning.

Cloud computing can function as comprehensive term roughly encompassing the concepts described in the figure above, and e.g. Gartner have adopted the practice of primarily using cloud as an adjective.

Different value propositions

Accepting the fact that it is not the definitive definition but rather the start of a description, let’s look into the NIST model (because it is a quite good description) to see what aspects that are important and how the various components are different. It is easy to see how for instance Software as a Service (SaaS) in a personal setting presents a very different value proposition from enterprise Infrastructure as a Service (IaaS). The first emphasizes e.g. collaboration, value derived from combining multiple users’ data  and device-independence while the second emphasize e.g. elasticity, and legal compliance aspects. When analysing cloud computing, it must therefore be important to sometimes look at the different delivery models, deployment models and use cases separately. Nevertheless, all of them have a common cloud component.

Another distinction that is often overlooked is how cloud computing is most often described both in terms of technology and terms of business model. This is true for attempts at strict definition as well as fuzzier descriptions. In the list of characteristics in the figure, the first three are simply abstracting the scale (“rapid scalability”), time (“on-demand self service”) and location (“location independent resource pooling”) of the computing resources while the “ubiquitous network access” is an enabling technology and an indication of the important of the Internet. Lastly, “measured services” is primarily a property of the business model. The choice of deployment model will entail different technical demands, but is essentially a question of ownership and as such, a business model consideration. IaaS, PaaS and SaaS are different in technological terms while also representing different business models.

In summary: The lack of a strict and clear definition is not a hindering issue. It is more important to understand how different technologies and business models are encompassed by the term, and how they should sometimes be analysed separately despite there being common aspects and implications of the cloud computing and IT as a Service paradigm.