Editor's Note:
Robert E. McFarlane is a principal at Shen Milsom & Wilke, an international firm that provides consulting and technology design on large and small projects, including information technology data center planning. McFarlane’s expertise is in the design of data center infrastructure.  ENR approached McFarlane with a series of questions about the reports of electrical problems at the National Security Administration’s data center in Utah, seeking to better understand the kinds of issues that may be involved. Several other experts ENR approached were eager to respond as well, but prevented from doing so by non-disclosure agreements they had signed with the NSA.

All we seem to know about the new NSA data center in Utah is that it is supposedly 100,000 sq ft of raised floor space in a million sq ft complex, that it is a 65 MW facility, and that for about fourteen months something has been going drastically wrong with the power system. Big firms were contracted to design and build it—firms said to have a lot of experience with complex, high-tech government projects.  So what has gone wrong? 

This is a National Security Agency project, and everyone involved is under strict non-disclosure so we can’t even verify the little that has been reported. Speculating would be foolish:  If the people directly involved haven’t been able to figure it out, it would be ridiculous for us to assume to know the answers.  What we do know is that we, the taxpayers, are presently on the hook for a $1.5 billion “white elephant” in Utah that doesn’t work, can’t yet be used, could well cost millions more before it can even be opened and, if present conditions are any indication, could be fraught with numerous other problems after it actually goes into use.

Since we can’t identify the cause of these apparent electrical “faults”, and the commensurate “explosions” that are said to be sending surges through the system and destroying equipment, maybe we can take a forensic approach and learn from the contractor’s dilemma by asking questions. It won’t tell us what’s actually wrong at the NSA, or change the outcome of this project, but it could help us avoid similar situations ourselves if we ask and answer appropriately on our own jobs.

Q 1:     Is this just another project the government doomed to failure from the start?

That would be an easy answer, particularly in light of the problems being seen on the ObamaCare website.  Anyone experienced in government projects knows that the procurement process and specifications can mandate ill-advised or unnecessarily complex designs and onerous procedures. There are also fee constraints that often discourage using the most experienced experts (Partners and Principals) by stipulating unrealistically low compensation levels. And there can also be incentives to use firms or contractors that may be based more on politics than on technical acumen.

Then there’s the bureaucracy that can eat up resources, make questionable decisions, and push contractors to shortcutting and cost-cutting.

But the firms involved in this project tout their experience with large federal projects. They should be well aware of the quicksand lurking in the mirage of a government contract, and how easily one can be sucked into it. They should also be experts in handling the bureaucracy and efficiently dispensing with the overwhelming paperwork. 

Q 2:     Did the government keep making changes?  Were the designers over-ruled?

We have now heard that major changes were dictated for the healthcare website only weeks before the “go live” date. We don’t know how many other times the website requirements were changed over the course of development, but this is something owners tend to do. But the government does it even more often, as multiple people are put “in charge” and re-think requirements, adding capabilities that weren’t included in the original specs.