Thursday, June 9, 2011

Whether you believe in the cloud or not, it’s coming.


The article from Barton George originally appeared here

Whether you believe in the cloud or not, it’s coming. That being said, it’s not a phenomenon that will fill skies of IT departments tomorrow, but rather, it is starting out as another tool in IT’s bag of tricks. As time passes over the next three to five years however, cloud computing will increasingly become a greater part of the portfolio of compute models that IT departments manage, sitting alongside traditional computing and virtualization. 

Cloud Computing Today

If you were to graph the distribution of compute models being used today by IT departments in large enterprises, it would look something like the chart below. Today, traditional computing and virtualization are where most of the distribution lies, with a little bit of flirting with the public cloud in the case of software-as-a-service (SaaS) applications for areas like human resources (HR), customer relationship management (CRM) and email. Additionally, to ITs consternation, public cloud usage can also be found in the form of the unsanctioned use of third-party cloud sites by various business functions. On the private cloud side its existence in large organizations is presently negligible.


Over the Next Three to Five Years

Over the next three to five years the above distribution will flatten out and shift to the right and will resemble the graph below. Private cloud will represent the largest compute model utilized, but it will be equally flanked by virtualization and public cloud. You’ll notice there will still be a decent amount of resources that remain in the traditional compute bucket representing applications that are not worth the effort of rewriting or converting to a cloud platform.


Evolutionary vs. Revolutionary

One of the things to note with this new distribution is that the lines between virtualization and private cloud will start to blur (there will also be a blurring between private and public clouds, as hybrid clouds become more of a reality in the future, but that’s another story for another time). There are two ways to go about setting up private clouds, or really any type of cloud: evolutionary and revolutionary. 

The evolutionary approach starts with virtualization and is appropriate where large investments in that area have been made and when you are talking about traditional enterprise applications. With virtualization serving as the foundation (see the graph below), additional capabilities are then layered on, such as usage-based billing/chargeback, workload life cycle management, dynamic resource pooling and a self-service portal for users. 

One of the key aspects of the evolutionary approach is that at every step along the way, every capability added brings greater efficiencies and agility. You do not need to wait until you meet the full definition of a private cloud to derive value and you can stop anywhere along the way.


You Say You Want a Revolution?

The other way to get to the cloud is the revolutionary approach. This is appropriate for Greenfield opportunities within organizations, and is targeted at nontraditional, Web 2.0 applications that are "cloud-native" — applications written for deployment in the cloud. These revolutionary solutions will often be delivered as an integrated, turnkey unit (see graph below).


You Don’t Need to Choose

Rather than adopting one or the other, most organizations will use both approaches to get a private cloud. At this stage the evolutionary approach will be the predominant way of getting a private cloud. As more and more "new world" applications are developed for the cloud, the balance will begin to tip in favor of the revolutionary approach.
Whether evolutionary or revolutionary, over the next three to five years cloud computing will become a greater part of IT’s compute portfolio. In the longer term however, the elements and characteristics of cloud computing will be absorbed into IT and it will simply represent the way computing is handled.

No comments: