Saturday, August 7, 2010

CIO Cloud Initiatives

With VMworld just around the corner, I've been reflecting on EMC World's focus of "the journey to the private cloud." Even today, some of the statements made during the opening keynote by Joe Tucci still don't sit well with me.

In the keynote, Joe showed a slide that had a list of ten top CIO initiatives for 2010.  The third initiative was Cloud Computing, followed closely by Storage Virtualization.  People can argue endlessly on how accurate surveys such as this one are, but from a hype standpoint, Cloud Computing is definitely up there.

I question how many of those CIOs have an accurate understanding on the nature and skills of their IT organization and how that impacts the feasibility of internally implementing a private cloud... successfully at least.  I've always pictured the "journey to the private cloud" as a model similar to Maslow's hierarchy of needs.

On the lowest level, you have the technical competence of staff.  Is your organization properly staffed with the right type of IT people to architect, implement, and administer a private cloud?  Or are there a substantial number of fiefdoms staffed by "next-next-finish" administrators?  Any gaps here would need to be resolved before a private cloud could be successfully implemented.  The resulting architecture is going to be most likely more complex and automated than the current environment, and this will amplify any skills gaps.

The next level is the maturity of workflows and processes.  Are the steps for implementing new IT systems well defined?  Is there a good understanding on where time is being spent for new system implementations?  More importantly, are the processes relevant and actually followed, with appropriate budget and technical approval in place to ensure that new systems are properly financed and designed?  How is chargeback handled?  Without good processes and workflows, all private cloud provides is the automation of broken processes and a further lack of fiscal responsibility.

The third level is the organization's current state of virtualization.  What is the current percentage of systems virtualized?  What barriers exist to increase this percentage?  Is the environment stable, performant, and well managed?  If your organization is bad at virtualization, it will be terrible at private cloud.

If the organization is successful on those three levels, then the private cloud is likely the next step to take to further increase the business value of IT.  In fact, organizations in that state will gradually adopt a private cloud model without any external influence as they continue to automate and consolidate the routine architectures and implementations.  

However, many organizations aren't there yet... at least, not enough to make Cloud Computing the third most popular CIO initiative.  I'd argue that many times when CIOs say that Cloud Computing is important, what they're really saying is that their current IT infrastructure is painful to work with, expensive, and slow to change.  The hope is that once this new buzzword architecture is implemented, their IT will suck less.

In a way, this is similar to the "server consolidation" CIO initiatives during the infancy of virtualization.  Lots of organizations had initiatives to reduce the server footprint.  The resultant architecture ended up being Windows VMs running on VMware GSX Server running on Windows 2000/2003 (since, after all, the current footprint was primarily Windows so those departments were assigned this initiative).  While meeting the "requirement" of server consolidation, these implementations resulted in little to negative business benefit.

Lets make sure today's private cloud implementations don't end up the same way.

No comments: