Data Center Journal

VOLUME 44 | JUNE 2016

Issue link: http://cp.revolio.com/i/686088

Contents of this Issue

Navigation

Page 10 of 24

8 | THE DATA CENTER JOURNAL www.datacenterjournal.com t ypically, the metrics used in our world are $/kW of IT capacity and $/sf of real estate. Some folks reflect all costs in $/kW. Others simply talk about the data center in $/kW and leave the land for the building ($/sf ) out of the equation entirely. In both cases, however, to drive a low $/sf and/or $/kW, "economy of scale" is the assumed catalyst that spurs the cost-effective metric of measurement. is model is used both by providers of multi-tenant data centers (MTDCs) and prefabricated modular units and by many enterprises building their own facilities. Although the belief that building at scale is the most cost-efficient method of data center development appears logical on the surface, it relies on a fundamental require- ment: boatloads of cash to burn. It's fIrst cost, not just tco In terms of data center economics, no concept has garnered more attention, and less understanding, than total cost of ownership (TCO). Entering the term "data center total cost of ownership" into Google will provide you with over 11.5 million results, so obviously people have given this topic a lot of thought. To a large degree, the problem lies in the nature of the com- ponents that comprise the TCO calculus. Because of the longitudinal elements that are part of the equation—energy costs for example—the perceived benefits of design decisions sometimes hide the fact that they are not worth the cost of the initial invest- ment ("first cost") required to produce them. For example, a common area where we find this to be the case is in the quest of many operators and providers to achieve lower PUE. Although certainly admirable, incomplete economic analysis can mask the impact of a poor investment. An example of how the failure to properly examine the impact of first cost on a long-term investment is a company's choice of cooling methodology. A recent analysis by Romonet, a provider of data center analytical soware, entitled Trad- ing TCO for PUE? delivers just such an example. e company chose New Mexico, owing to its favorable atmospheric condi- tions, as the location to examine the value of using an adiabatic cooling system in addition to air-side economization in a hypothetical location. New Mexico's cli- mate is hot and dry and offers substantial "free cooling" benefits in the summer and winter as Figure 1 shows. rough the use of an adiabatic sys- tem, the site would benefit from over four times the free-cooling hours than without such a system. Naturally, the initial reac- tion to this cursory data would be "get that cooling guy in here and give him a P.O." And if we looked at the perceived cost sav- ings over a 10-year period we'd be feeling even better about our $500K investment in that adiabatic system, since it appears to have saved us over $440K in operating expenses. Unfortunately, any analysis of this type needs to include a few things such as discounted future savings (otherwise known as net present value, or NPV) as well as the cost of not only system main- tenance but also the water used, and its treatment, over the 10-year period. When these factors are taken into account, our $500K investment in an adiabatic cooling system actually resulted in a negative payback of $430K! e point of this example is that very oen the failure to account for the long-term impact of an initial decision can permanently preclude companies from exercising alternative business—not just data center—alternatives. Figure 1. Free-cooling hours mean that the compressors are off—which on the surface means "free," as they are not drawing electricity.

Articles in this issue

Links on this page

Archives of this issue

view archives of Data Center Journal - VOLUME 44 | JUNE 2016