Mike Manos discusses data center site selection, you need to “kick the dirt” to find what is real

At Gartner’s Data Center Conference, Mike Manos made an excellent point that “75% of the data center costs are effected by site selection.” Great architecture is designed to a site characteristics.  But, the status quo is to design data centers that are built based on past experiences.  Green data centers need to be designed to fit with site characteristics.

Mike wrote a post on site selection.

Kickin’ Dirt

December 21, 2009 by mmanos


I recently got an interesting note from Joel Stone, the Global Operations Chief at Global Switch.  As some of you might know Joel used to run North American Operations for me at Microsoft.  I guess he was digging through some old pictures and found this old photo of our initial site selection trip to Quincy, Washington.

As you can see, the open expanse of farmland behind me, ultimately became Microsoft’s showcase facilities in the Northwest.  In fact you can even see some farm equipment just behind me.   It got me reminiscing about that time and how exciting and horrifying that experience can be.

Kicking the Dirt.

Many people I speak to at conferences generally think that the site selection process is largely academic.   Find the right intersection of a few key criteria and locate areas on a map that seem to fit those requirements.   In fact, the site selection strategy that we employed took many different factors into consideration each with its own weight leading ultimately to a ‘heat map’ in which to investigate possible locations.

Even with some of the brightest minds, and substantial research being done, its interesting to me that ultimately the process breaks down into something I call ‘Kickin Dirt’.   Those ivory tower exercises ultimately help you narrow down your decisions to a few locations, but the true value of the process is when you get out to the location itself and ‘kick the dirt around’.   You get a feel for the infrastructure, local culture, and those hard to quantify factors that no modeling software can tell you. 

Mike makes an excellent point for the decision on site selection.

Once you have gone out and kicked the dirt,  its decision time.  The decision you make, backed by all the data and process in the world, backed by personal experience of the locations in question,  ultimately nets out to someone making a decision.   My experience is that this is something that rarely works well if left up to committee.  At some point someone needs the courage and conviction, and in some cases outright insanity to make the call.

Are you willing to take a risk in site selection?  Most aren’t.  But, the leaders are, and they are the ones who are first to go where others haven’t and have lower costs.  Mike has said the cost of the land was a great deal as no one thought of the land as a data center site.  Google are the others who have this down.