Someone recently said to me that no two data centers are alike. There’s probably a good deal of truth in that, or maybe, it’s just that most who design or use data centers like to think that what they design, build or occupy is totally unique. Those who are on the bleeding edge of design have technology on their side of the argument. On the other side, the proliferation of modular data centers is leading the way to some manner of convergence and standardization. Several industry publications like Tier 1 and DataCenter Knowledge (see the article of 2/6/2012, http://www.datacenterknowledge.com/archives/2012/02/06/the-state-of-the-modular-data-center/) have made note of this. According to Jason Schafer, Research Manager at Tier1 Research, “If data center owners and operators are not at least exploring and considering modular components as a means for data center expansions and new builds, they are putting themselves at a significant disadvantage from a scalability, cost and possibly maintenance standpoint.”

Certainly economic factors are considerable when developing a data center and incentives and tax rebates have become increasingly important. At one time in the not-to-distant-past, the optimum location for a data center was determined by the availability of telecommunications infrastructure and, to a lesser extent, power. Now that fiber is pretty much ubiquitous, power availability, quality and cost has become increasingly more crucial. Until such time that servers use less power (and there are some incredible innovations on the way from SeaMicro [owned by AMD] and Calxeda) and create less heat, the need for power probably won’t abate. Other environmental threats such as earthquakes, lightning, tornados as well as contingent risks such as proximity to airports, rail and hazardous materials will continue to be factors in deciding where to locate data centers. However, it seems to me that beyond these operational issues, there is a greater imperative for site selection.

Critical mass seems to take the lead in the argument. Herd mentality plays a role in assuaging the concerns as to whether or not a data center will be successful in one location as opposed to another. If there are other data centers or service providers in a particular locale, it just makes sense to join them. That story has a certain amount of validity right up to the point at which the market reaches saturation and that results in the commoditization of product/services.

Underlying all of the concerns, evaluations and analyses is the business case. Beyond the physical attributes of the site, what are the business motivations that ensure that the venture will be a home run? It all comes down to researching the demand prior to committing to the location. Determining the customer based served, and how it will be served is essential. Understanding the demand for services required is indispensable. This sounds insultingly obvious, but I’ve witnessed so many occasions when second-hand, anecdotal information is cited as support rather than actual research based on either data (IP Addresses) collected, or direct personal interviews with potential end-users or service providers. Research builds the business case, which is supported by every positive attribute of the site, and feature.