How cloudy are you?

Can a data center be called cloud computing? Well, I would not say “no” a first.  Finally it boils down to defining the terms properly. What is a data center? What is an in house IT infrastructure? What is the cloud and does all relate? In our daily work with cloud computing economics we discovered that the concept seems to be simple at first, but the confusion on the market place could not be bigger. A mail from a data center provider in northern Europe illustrated this confusion:

…Our continuous services department offers server, database, Microsoft, network, application server and also PC management services. We are in that area the biggest in XXX. We have 4 datacenters and manage about 1300 servers. We have installed and manage a number of virtual-server environments (I guess we could attach the buzzword “cloud” to them J) and use virtual servers for hosting.  We are using both VMWare and Microsoft platforms….

Not everything that has written cloud on it,  is really cloud. This raises the question: How cloudy are you anyway and how do we measure your cloudness?

We have identified the following two parameters as indicators for cloudness: the granularity of the service offer and the speed of resource scaling. The granularity refers to how small is the smallest bit of computational resource that you can buy.  This spans from an on IT center size to seconds of computational power.  The speed of resource scaling on the other hand indicates how long it takes to increase the size of available resources.

During the last century utility companies have gone through the same process as IT providers will be going now. About 100 years ago the electricity market was much more fragmented than nowadays. Almost every company needed to produce its own electricity. As we see oligopolies and  monopolies of electricity companies  nowadays, the same consolidation will happen to IT service providers.

It all started in the 80’s of the last century with every company cooking its own soup. You would have to plan at least 5 years in advance when setting up you IT environment. It was the times when a company had to invest millions and millions in its own IT infrastructure that were only rarely fully utilized. You had to buy 5 year capacity at a time. It was also times, when increasing the computational resources would take you months along with  big number of highly skilled staff to be hired. You would then be married to those professionals for a lifetime even if the demand weakened or you simply did not like them any more. If you wanted to increase the capacity, this could easily take you 3-5 months. You would order a couple of new servers.  It would take you a couple of weeks till the servers arrived.  Unfortunately the specification was not right and you got the wrong model shipped.   After yet another couple of weeks you would finally have the right hardware in place.

Then came the data centers. Outsourcing was the buzzword.  Suddenly you could fire many of your unwanted IT managers and outsource your infrastructure to a data center. OK, it could still take you weeks to increase resources when needed.  You could outsource maintenance, but still had to buy computing power in big chunks with little elasticity. Now you could buy 1 server year at one, with 1-2 weeks notice. Only the electricity bill came to you.

Finally cloud computing appeared on the horizon around in 2007. Internet giants like Amazon, Google, Microsoft and Salesforce.com had own big data centers in place.  They looked for ways to capitalize on unused resources. You could buy an infinitely small amount of computational power with just in time scaling of resource. As in utilities, cloud computing provider have the size to be a buffer for demand fluctuations.

Cloud computing is a new paradigm that data centers can not just put on like a jacket.  Cloud computing services must offer computational resource at an infinitesimally small granularity and must have the size to offer a just in time speed of resource scaling.  If, and only if, those two factors are given, it can called cloud computing. Everything else is old wine in new bottles.

Advertisements

About Georg Singer

Georg Singer is an analyst at the University of Tartu in Estonia. He works for the institute for computer science on cloud computing economics.

3 Responses to “How cloudy are you?”

  1. Certainly one of the best explanation that I have read about cloud computing. Well- structured and very simple but definite. Even though that the cloud computing technology had started last 2007 but I think today its still premature for most companies grasp the convenience and the benefits that the technology can offer. Maybe in the next 3-5 years cloud computing will be implemented by most companies on their IT infrastructure and it will gain its stability on the IT World.

  2. Thanks Sarah! Yes, you are maybe right with your projection.

Trackbacks/Pingbacks

  1. Tuesday data center tidbits. « The Server Room - February 9, 2010

    […] comes from an article about “cloudness”. Grading a cloud computing provider based on the granularity of their billing and offering instant […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: