My Cup of IT: The Biggest Loser?

By: Steve O’Keeffe

Like you, last week I read the patient’s chart on the 2012 budget and what it means for IT. Spending pulse quickens by $1.5 billion up $3.5 billion, down $2 billion. Amputations at OPM, NASA, and Education. Huge infrastructure cholesterol scares at Army Corps, VA, HUD, and SSA. DoD puts on the white coat as lead shrink in cyber anxiety.

But, it’s the prescribed data center diet that left a lump in my throat. Hats off to OMB for its commitment to EHR – government transparency and new medicine. OMB examines the patient – the data center count skyrocketed 385 percent since 1998. Dr. Kundra then plots the data center spike directly against the IT spending hike in the same period. This is a “fatty foods, high blood pressure” connection – the intended takeaway here is “cause and effect.” OMB sees pork in data centers. And, following the first lady’s lead, Dr. Kundra’s putting Uncle Sam on a data center diet.
While doctors are famous for their illegible scrawl, Dr. Kundra’s prescription is written in block capitals. Publishing the results of the 2010 Federal Data Center audit, OMB goes beyond the 2,094 aggregate number to break out the data center count at each agency. And, if that’s not enough, Dr. Kundra sets us up for our very own version of The Biggest Loser telling each agency how much weight to drop by 2015. We’ve published the chart from the 2012 Budget Analytical Perspective below and added a couple of columns to rank agencies by absolute and relative target data center cuts, red columns on right.
Agency Name # DCs in ’10 Target # DCs in ’15 Absolute Ranking Relative Ranking
Department of Commerce 41 23 7 7
Department of Defense 772 428 1 6
Department of Energy 89 83 11 15
General Services Administration 15 3 10 3
Department of Health and Human Services 185 131 4 9
Department of Homeland Security 24 2 10 1
Department of Interior 210 120 2 8
Department of Housing and Urban Development 2 1 14 5
Department of Justice 65 50 8 11
Department of Labor 20 18 13 14
National Aeronautics and Space Administration 79 57 6 10
National Science Foundation 2 1 14 5
Nuclear Regulatory Commission 3 1 13 4
Small Business Administration 4 2 13 5
Department of State 361 282 3 12
Department of Transportation 35 31 12 13
Department of Treasury 42 29 9 8
U.S. Agency for International Development 2 1 14 5
U.S. Department of Agriculture 46 7 5 2
Hardly surprising, DoD tips the scale as the sumo with 772 data centers. And, according to the OMB diet, the armed services have the most weight to drop 344 data centers. It’d be interesting to get that breakout by service branch… Interior, State, HHS, and USDA all make the scales creak. But, cutting the data by relative versus absolute weight loss targets reveals DHS, USDA, and GSA face the toughest pound-for-pound targets.
To be sure, the diet’s not going to be an easy pill for agencies to swallow. Dr. Kundra deserves high marks for giving a definitive diagnosis no qualitative TQM or BPR prevarication here. This is a bold new treatment. But, questions remain. Incentives where’s the stick, carrots, and is OMB eating its own dog food? Accuracy of the scales is counting data centers a meaningful metric? Viability of the nutritional substitution plan where’s the South Beach cookbook and what happens if the “Open Table” is bare? Do data center “owners” really own them – or is there more than one person on the scales?
As Kundra, M.D. points to data center consolidation as the funding source for his cloud vision, only time will tell if he succeeds in getting Feds to shed their IT bulge. By being so direct about the number, the reduction goal, and the timeline – what happens if the pounds don’t come off? 2015 is over the horizon – who’ll be in charge at OMB and the agencies to get back on the scale? Unlike the reality show, there are many, many variables in this equation. If we don’t get it right, then who’s the biggest loser?

Leave a Comment

One Comment

Leave a Reply

Profile Photo Mark Forman

I like this notion of saving money by taking advantage of IT trends, and kudo’s to Vivek for driving change. But let’s call it when the emperor ain’t wearing any clothes.

The commercial world learned years ago that you can’t save money by moving a servers from different buildings into one big building, because the biggest cost driver is labor and that’s tied directly to the complexity and number of applications being managed. No reduction in complexity…no savings. Consolidating data centers will at best provide incremental savings and could end up costing money if we don’t have enough IT infrastructure PMs available for the facility moves, nor enough people who have the ITIL experience needed for this, and technology migrations implied here are substantial. So there is a major risk of cost increase and little real potential savings.

The largest cost driver in IT data centers is the shear complexity of managing disparate client-server apps, not the number of buildings that house them. The revolution in IT is about commoditization, and the real cost savings and performance gains result from shutting down client-server apps while moving to shared services at the enterprise or global level…aka web services via cloud providers. The new approach to savings in data center consolidation is to shut down applications, not move them onto bigger machines in fewer buildings. I think Vivek understands this, but it is very hard to get agencies to adopt new business standard approaches that come with cloud or enterprise services…something about “not invented here”…

The beauty of Vivek’s reform plan is coupling data center consolidation with the cloud strategy. It would be good to know how many applications, not buildings, are being shuttered.