We were also running 24,000 physical servers when virtualizing would eliminate the need for a whopping 6,000. Our environment was relying on the most expensive storage for every piece of data, when tiering our data and reducing duplication could lower our storage costs by 50 percent. Some applications were running on expensive Unix platforms, even while open standards were five or 10 times cheaper. Now we're saving more than $300 million per year-all of which leaves budget dollars that can be allocated to innovation rather than maintenance.
And of course all of this work to increase our data centers' efficiency is good for the planet, since we no longer needed to build the new data center we were planning. (While many companies are building new so-called "earth-friendly" data centers, we believe that the greenest data center is the one you never build.)
As of today we're about 70 percent of the way there, and we expect to finish standardizing and streamlining our systems in late 2012. We aren't devoting our resources to this project with a specific goal-to make a particular change or release an upcoming product. That's the point. Rationalizing the underpinnings of every technology that Dell uses to conduct business will increase our agility and speed, and speed is the goal.
When speed is the goal, you have to ensure you're measuring the right things to get you there. Many IT leaders focus on development timelines and delivery metrics, while business people tend to measure time to launch. But when or whether a project meets its delivery timeline really isn't relevant in the end. The purpose of IT is to create new value streams for the enterprise, so the question you must ask and answer is whether a given project achieved the benefits promised.
That's why it's not enough for today's CIO to understand the role of technology in business - he or she must understand and enable the business itself. Technology and the business are the same thing. Saving money is too expensive; the agility tax is too high. Delaying it will only raise it, probably at a moment when your company can least afford the bill.
Sign up for Computerworld eNewsletters.