With this information in mind, we talk with IT about where they tend to operate on the "technology continuum." In other words: How often do they turn over hardware? How aggressive are they in acquiring the newest technology? How flexible are they in appropriating funds and staff to go along with the new business or technology as it occurs? We then overlay all this information with our knowledge (some public and some under non-disclosure) as to how the industry is moving. Finally, we work up actual layouts to determine the amount of space that will be realistically required over the projected time period, both for technology and for infrastructure. Considering how hardware is getting smaller, but at the same time proliferating, and the move many companies are making toward server consolidation and virtual computing, this is not a simple task. And it is far from a "rule-of-thumb" process.
But no matter how confident we are with our analysis and projections, we are always looking to locate the data center so there is opportunity for future expansion into "soft space" -- area that can be taken over without having to move a department with difficult or expensive amenities and with nothing on the separating wall that would make its removal problematic, such as power panels or the communications demark. In many cases, this area is even equipped with raised floor, piping and power feeds in order to get heavy construction done at the least expense and with minimal later disruption.
The one "rule of thumb" we can provide is that it is generally accepted today that, for a major data center meeting the Uptime Institute's Tier 3 or Tier 4 standards, you should plan on the mechanical/electrical support space being at least another 50% (and probably closer to 100%) of the technology room space. In other words, whatever space you predict will be needed for cabinets and other technical hardware, double it.
This was first published in November 2005