Recent weeks have seen the release of three significant white papers on growth in data centers. All have received pretty wide attention in the trade press and even in national media.
Jonathon Koomey produced probably the most prominent of these studies, at the request of the New York Times, which probably guaranteed that it would be widely read. Koomey, who has long been an influential figure in energy debates in the U.S., makes a real contribution in this report (Growth in Data Center Electricity Use 2005 to 2010), as it appears to be the first effort to document how much energy data centers use. More than that, though, Koomey also tracks changes in energy intensity in data centers over the period 2005-2010.
Among his conclusions, “Electricity used in US data centers in 2010 was significantly lower than predicted by the EPA’s 2007 report to Congress on data centers. That result reflected this study’s reduced electricity growth rates compared to earlier estimates, which were driven mainly by a lower server installed base than was earlier predicted rather than the efficiency improvements anticipated in the report to Congress.”
I know that early reports highlighting the greater energy intensity in data centers tempted me to credit industry efforts to improve energy use.
Jeff Paschke, Rick Kurtzbein, and Michael Levy of Tier1Research published a report, “North American Multi-Tenant Datacenter Supply Emerging Major Markets – 2011” that suggests that we’re building more data center space than ever. As part of this study, the authors looked at data center construction in the Top 50 Metropolitan areas in North America from 2000-2010. Forty-six of these markets show growth. Some of the markets have seen astounding growth, including Las Vegas (46 percent), Raleigh-Cary (41.8 percent), and Riverside, CA (29.8 percent). The report does an excellent job analyzing potential drivers that cause regional building booms. The report says, “ It turns out there is a rough correlation between metropolitan population size [Metropolitan Statistical Area (MSA) in the United States and Census Metropolitan Area (CMA) in Canada] and datacenter market size. This makes some sense because population size correlates to the business activity and number of jobs in a particular metropolitan area. But GDP of a market also is a factor. And other factors also come into play when determining the potential datacenter market size, such as severe weather risks (e.g., hurricanes), operational costs (e.g., power costs, labor costs, state and local taxes, etc.), fiber availability and the types of business located in a city. For example, financial services and Internet businesses are heavy users of datacenters, while manufacturers are typically not.”
The third white paper, released by Juniper Networks, calls attention to gridlock in Washington, DC. Juniper Networks surveyed federal IT professionals in June 2011 to obtain their predictions for the change in capacity demand, and their readiness to address the associated management, security, and performance issues.
The paper includes many noteworthy observations, but I was taken by the lack of confidence federal employees expressed in the government’s consolidation plans.
- 10 percent of the respondents believe that the government will achieve its consolidation targets by its 2015 goal
- 23 percent think there will be more federal data centers in 2015, rather than fewer
Juniper included a list of recommendations with its rather pessimistic numbers.
All three white papers are worth reading, and I’ll be posting the papers or links to them on our website as soon as possible, for your convenience.
That said, the picture painted by the three whitepaper appears to be one of uncontrollable growth, with only a slowing economy as a limit or drag.