Wednesday, June 2, 2010

Cloud Ain't Easy

If there's one thing I've learned over the past decade working with grid computing and virtualization technologies is that innovation isn't easy. Cloud computing is an innovation initiative. It's not one thing but rather the implementation, configuration, integration, and automation of several things. It's not avaialble from one provider, off the shelf, in an easy to install and support package.

No. Cloud Computing is more like a jungle. Lots of custom scripting and some custom development. Lots of building interim and bridge utilities. Lots of pushing vendors to deliver what you need rather than what they want to provide. Add to it the confusion over what cloud is (as recently as Vegas InterOp in April I heard someone state cloud was just another name for server virtualization) and it's no surprise it's a jungle.

Cloud only makes sense when it's done right. It needs a solid architecture at it's foundation. It needs a strategy and roadmap to ensure it's built from the start to deliver real business value. It needs to align with business and technology goals. And it needs to pay for itself where nobody wants to pay for infrastructure. It can be done, just ask Bechtel, but it's not easy.

First there's the skills gap. Many, if not most, organizations plan to go the cloud computing route on their own. The underlying strategy is vendor driven mistaking the vendor's best interests for their own. Whether it's Amazon, VMWare or Citrix I assure you tying to their strategy only ensures you'll be a profit generator for them. The most prominent skills gap is in enterprise architecture which is where cloud's rubber should be hitting the road. But we don't have enough EA's in the world and especially not enough with cloud skills. So instead companies forge ahead on their own with predictable results:
- lack of synchronization between technology teams
- lack of security standards and policies which address the technology footprint of cloud
- poor or missing governance structures
- growth of stealth IT in the form of expensed SaaS and even IaaS cloud services
- lack of understanding by the business how cloud can enhance what they do today

Second there's the focus on cost takeout. True, cloud optimizes the infrastructure stack which means more bang for the buck. But doesn't lower cost mean you can perform more iterations for the same budget? Learning is an iterative process so if we can learn more for the same dollar, don't we become smarter? Isn't this the foundation of innovation? Trying something new, tweaking it, and trying again until we have the better mouse trap?

Third there's the belief that IT controls technology. Technology goes wherever the money is, and today with a credit card anyone can move to the cloud.

Fourth and most scary there's the "Cloud at any cost" crew who are innovating in the cloud without proper planning and strategy, exposing their companies to massive risk which cannot be quantified. Who owns the data? Is the company properly indemnified from patent infringement? Where does the data go? Who has the authority to ramp usage up or down? Is any regulated data involved? What if the cloud vendor disappears or goes bankrupt?

Cloud ain't easy. It requires planning and strategy. It requires both a business and technology point of view. Most of all it requires people to realize as a new, disruptive technology the old rules don't apply.  It's time to throw out the nice, easy to use cookie cutters we've created for the Internet.

Cloud simply ain't easy, but it sure is compelling!

No comments:

Post a Comment