Notes from the Consultant’s Jungle

The challenge in using green energy alternatives such as wind or solar power in data centers is power density.   Data centers consume a lot of power with respect to their physical footprint, so they need very large solar arrays or large wind turbine farms to generate the necessary power for the facility.  Consequently, the amount of time required to reach the positive ROI point with the solar or wind power equipment is typically quite long.  This economic detail is often the inflection point for deciding whether to apply green energy alternatives in the data center plan.

A December post on the Data Center Knowledge blog reports new data centers powered entirely by the wind.  In particular, the article focuses on the Other World Computing site in Woodstock, Illinois (northwest of Chicago), which claims to operate the entire facility solely on the power from a single wind turbine.

OWC’s website does claim that the site is primarily powered by the wind turbine.  The website goes on to say that the site has four layers of power, although along with the turbine, municipal power, and generator backup they claim the UPS as a power source.  We think that claiming the UPS as a source of power for the data center is a bit of a stretch.  OWC invests a lot of web page real estate talking about its wind turbine.  Kudos to OWC for making such a strong statement and investment in the effort to create a green data center.  We would add though, that there is some fine print (though not necessarily related to OWC) to consider in this story.

First of all, the OWC data center is a small one.  From the publicly available information, it likely has a critical load of less than half of a megawatt.  This is a very small facility by most common standards.  Indeed, other examples claiming to be fully wind-powered are similarly small.  Microsoft’s Virtual Earth application is hosted in a single data center container, which is said to be powered by a wind turbine.  A data center container is even smaller than the OWC example mentioned above.

Another example of a data center claiming to be 100% wind powered is Midwest Data Center in Rock Port, Missouri.  Again, this data center is small.  The payload floor is 1,000 square feet and the critical load is likely below 200KW.

Without going further to support the point, suffice to say that the applicability of dedicated wind or solar power for data center operation is limited to very small facilities… unless of course there is an option to leverage a very large (acres) wind or solar farm on nearby property.  That is also provided that the wind blows hard enough and often enough in that location to produce the planned sustained amount of power for the facility.

This brings us to the second critical point in this discussion which is ROI.  In the OWC example, the operator is expecting a 10-14 year ROI period, assuming the wind blows strongly enough through that period to support the modeling assumptions.  We think this is an ROI period much longer than most would find favorable (and, we would suggest, longer than the productive lifespan of the facility before requiring substantial renovation).  Similarly, the cost to acquire this energy source, when compared to the payload size of the data center facility itself, is an excessively large percentage of the total cost to build the facility that most planners would decide to not take this path.  It would be an investment of principle in a commitment to be green, rather than one that would otherwise please the accountants.

Indeed if one looks at the web sites of the companies mentioned here, there is ample web real estate invested to promote the “Green-ness” of these sites and to fluff the feathers over the use of wind turbines.  These companies have made a strong commitment to natural renewable energy and are reaching out to Customers who have a similar set of principles.