Navy's top IT official takes issue with the federal government's focus on cutting the number of government data centers, says it misses the big picture on IT cost control.
New York's 32-Story Data 'Fortress'
(click image for slideshow)
The Navy's top IT official took issue with the federal government's myopic focus on cutting the number of government data centers as a way to reduce costs, saying it diverts attention from the overall view of managing IT cost effectively.
"Stop worrying about the number of data centers. It's absolutely not the metric (that matters)," Navy CIO Terry Halverson told a group of government and industry executives and legislative aides at a technology briefing at the U.S. Capitol on July 10. "The metric is how much am I paying to store all the data collectively?"
Halverson said, instead of focusing on the number of data centers, his approach would be to "take $1.4 billion out of my data center costs -- the money is gone" -- and then channel investments where they're needed most. There are big savings in managing your data better, using better options, he said. "The trick is knowing the value of the data and where the users are."
"There are some places where it absolutely makes sense, given the new technology, to put a data pod right next to the user -- (it's) the cheapest way to do it. That's not where your costs are," he argued. He urged federal IT leaders instead to "look at what you're doing with your data; look at the best way to meet your mission with that data; and then look at the pricing. That's where we have to focus our time and stop counting the number of data centers we've closed."
Speaking at a cloud computing forum sponsored by the Information Technology & Innovation Foundation, Halverson derided the number and nature of government data center definitions. "We've had 25 or 30," he said, that have come out under the Federal Data Center Consolidation Initiative, describing facilities that in many instances are little more than the size of a telecommunications closet.
One of the fundamental flaws in the logic suggesting that closing data centers will result in government IT savings, said Halverson, is the fact that "we aren't a business. Sometimes what a business can do when they close a data center is sell the land, they stop the lease -- I don't get that savings (under government accounting practices)," he said. "Most of my data centers aren't standalone buildings. They are in buildings that are used for other things. So closing a data center isn't going to get me the savings you hear other people talk about."
Another factor lost in the debate about closing data centers, he said, is the importance of scale.
"I have one network that has 897,000 people on it. That's a big number," he said. "People ask, 'why can't you do this faster?'" Pointing out that the Navy must maintain network operations from Antarctica to space, let alone across the seas, he said, "If I patched everything perfectly, running at optimum speed ... physics says it takes a certain amount of time to get that done. You really have to understand all parts of your environment," and "not just pick pieces of your environment and say I'm going to do 'X.' It's not going to work."
Halverson wasn't alone in suggesting during the forum that the focus on data centers is failing to deliver promised savings. Congressman Gerry Connolly (D-Va.), ranking member of the House Subcommittee on Government Operations, faulted the Office of Management and Budget (OMB) and government agencies more broadly for their lack of overall progress in reining in federal IT spending.
"I'm disappointed in the federal government," in the way it manages IT, said Connolly. "This matters, because, if you look at the $81 billion a year we spend on IT acquisition and maintenance, at least $20 billion of it is lost (in) sunk costs and old, antiquated legacy systems."
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.
InformationWeek Must Reads Oct. 21, 2014InformationWeek's new Must Reads is a compendium of our best recent coverage of digital strategy. Learn why you should learn to embrace DevOps, how to avoid roadblocks for digital projects, what the five steps to API management are, and more.