Framework For a Discussion of Cloud Computing (In Less Than One Page)

For a little over a year Sameer Patel (a good friend, except for this instance) has been bugging me about defining the cloud in a one-pager.

I have to say, it’s been challenging.  I have written before pages and pages defining the cloud (including my cloud purist e-book – check it out) in many ways.  I have done charts and slides, entire presentations, reams for electronic paper, and more – but defining it in one page was proving nearly impossible.

Until yesterday.

I finally figured out what the problem has been.  It is not a problem of definitions (there are plenty of those going around) but a problem of confusion.

There is a general confusion as to what the cloud is because we use the same word to define three very different things.

  1. The cloud as the Internet.  In the old TV commercials that Microsoft sponsored five years ago so support the launch of Windows 7 we used to hear the battle cry “to the cloud” as if it was a place where everything magic happens and exists.  They, of course, meant the Internet (actually, to be more precise the WWW – the subset of the Internet that we operate via browsers).  Movies, data, applications – everything was “in the cloud” – meaning available via a browser from anywhere you could connect to the world wide web.  Needless to say, this is an improper use of the term (I covered a myriad of definitions in my e-book, including the “standard” NIST definition that most people use).
  2. The cloud as a delivery model.  This improper definition led to the start of the confusion since it led to the wrong use of the term “cloud” by vendors when offered hosted and SaaS-like solutions.  Ever since the days of ASP (if you remember that far back) and NetScape Application Server (NAS – again, if you remember that far back) we had applications that could be delivered via a browser.  RightNow Technologies and were pioneers in this movement allowing customers to use applications they ran in a data center in a timeshare model (let’s call it what it is – even if we don’t have mainframes).  This was the way out of the dark ages of client-server architectures and the beginning of embracing the internet in organizations.  The main problem this model has (still today) is that it shares a monolithic server running in a data center instead of leveraging the power of cloud computing architecture (distributed computing, more on this later).
  3. The cloud as a computing architecture.  Research on distributed work (later computing) models began in 1939 with earlier models of what became Cloud Computing Architecture.  In this model, each job to be done is divided into smaller possible computational pieces (operational before computers) and distributed to the best resources able to complete them timely.  This was done all at the same time – versus the serial manufacturing model prevalent at the time (that was spawned by the industrial revolution).  Granted, it was impossible to implement in the early days but we have come a long-way since the start of computers and the growth of computing power to the point that we can have a three-tier architecture that supports distributed computing well today.

The above definitions are where vendors (especially Enterprise Software) vendors are stuck (a combination of one and two above) versus where customers want to go (three above).

And the source of the confusion.

I will build the above into a one page, downloadable infographic in the next couple of weeks (be careful what you ask for Sameer :)).  But I wanted to get the conversation going in the right direction.

No definitions, just a way to recognize what each vendor is doing – and a post I will point to as I begin to evaluate where each major vendor is, and where they are going, in the next few weeks.

What do you think? Am I off?

disclaimer: even though I don’t mention any vendors here, some of the major vendors are my clients – some are not.  I will disclose those more as I move forward – but I want to use this little piece of writing to thank Mike Fauscette, Paul Greenberg, and Denis Pombriant for helping me solidify my thinking.  Any interesting or good insight above came from them, any major errors are totally mine and I did not let them talk me out of them.  and for the record, the definitions above are 467 words without good editing done – under a page… :).  more on this will be coming soon, be patient… but this is critical to frame the conversation of what the cloud is and is not.

7 thoughts on “Framework For a Discussion of Cloud Computing (In Less Than One Page)”

    1. those two are implementations of grid computing, an earlier model of timeshare and cloud computing focused on leveraging clients (SaaS) as computational power. there are many, many issues with this (mostly around security) that are unsolvable at scale (why SaaS is not a fast evolving world – and PaaS and IaaS are).

      have not heard of fog computing in a while and as far as i remember it was a marketing term used by Cisco or HP or one of those HW providers. as with everything else that takes an open model (cloud) and makes it closed – not cloud computing architecture.

Comments are closed.