Verne Global

Data Center |

13 October 2017

Why can't the US government close its data centers?

Written by Peter Judge (Guest)

Peter Judge is the Global Editor at Datacenter Dynamics. His main interests are networking, security, mobility and cloud. (All blogs by Guest bloggers are their own)

Most organisations are finding that running their own data centers with dedicated hardware costs a lot, and closing and consolidating them is a great way to save money. IT tasks can be done more efficiently in virtual servers, in shared, colocation spaces or even on the public cloud, while dedicated in-house hardware spends too much time sitting idle.

Government agencies, more than most others, are accountable and have stringent budgets. And they also have racks and racks of dedicated in-house servers. They ought to be gaining masses of savings through tidying them all away - but it seems they aren’t. Let’s try and understand why.

The US government is well aware of the benefits of consolidation, virtualisation and cloud. It has set out clear directions to make savings by reducing its IT spend.

Back in 2010, the Obama administration decided it could save a lot of money by closing down and consolidating its data centers. The FITARA Act was about saving money, and it included the ‘Data Center Consolidation Initiative’, which specified that public bodies should operate fewer and better facilities.

Government agencies got busy looking at what savings could be made, but in 2015, there was a surprise: a check by the Government Accounting Office (GAO) found the number of data centers the government was running and paying for had actually gone up. There were 11,700 data centers - 2,000 more than had previously been known about!

This wasn’t about new data centers arising, so much as it was about government accountants getting better at finding the data centers their departments were running. The definition of data centers was extended to include small server rooms - which are among the least efficient ways to do IT. The opportunities for savings actually grew!

With that target in sight, things got more serious. A new Data Center Optimization Initiative (DCOI) memo was sent round by the Office of Management and Budget, which gave more detailed guidance about how to implement the demands of FITARA. Things like virtualisation and automated server management were spelt out.

Progress was made and the GAO reported good news and bad news this year:

The good news: in March, the GAO announced that $2.8 billion had been saved through consolidation under FITARA.

The bad news: most of those savings were made by just four agencies - Commerce, Homeland Security, Defense and the Treasury. Huge savings were being missed in the other bodies.

A second GAO report in September found that beyond those four, almost no agencies have met the targets set out in the DCOI.

What’s the problem? Data center consolidation requires investment, and man-hours, to save money down the line. Most of these agencies are struggling to find enough budget and people to put in, to get the savings out later. Savings from consolidation are perhaps difficult to explain and demonstrate ahead of the fact - and the accountability processes in the public sector would make it hard to proceed without those concrete demonstrations.

The GAO thinks that the evidence is there - look at the savings made by the leading agencies. So it’s suggesting that the current administration extend the deadline for the FITARA act and the Data Center Consolidation Initiative. To me, that looks like a sound suggestion. It would generate business for cloud providers and colocation services, while saving the tax payer money. Let’s see if the advice is heeded...

Note: You can read more of Peter's blogs at Datacenter Dynamics here.


Sign up for the Verne Global newsletter

Opinion, thought leadership and news delivered directly to your inbox once a month.