Water is getting more attention as a critical resource for a data center. Compass Data Center's Chris Crosby has a post on Water.
Water, Water, Everywhere…
As Coleridge made clear, water is essential for survival. Since data centers hadn’t been invented when he published The Rime of the Ancient Mariner in 1798 he didn’t mention them. But if they had, he certainly would have included them in his epic poem since water is too often a common requirement for their operation. This water dependency is an area of consideration that businesses should factor into their plans for upcoming data centers.
A high percentage of today’s data centers use water-based cooling methods to keep them from becoming the equivalent of a Hopi sweatbox in the desert. Although evaporative cooling, whether through traditional towers or “advanced” swamp coolers, remain highly effective cooling methods, when you’re planning a new data center you may want to consider the impact of the weather and water availability on your decision.
Chris mentions the drought in Eastern Washington in 2001.
The other element that water can have a substantial impact on is your data center’s power. Specifically if your new site will be serviced by a utility using hydro-electric power. I was recently reading an article that stated, “Quincy, WA—Big Data Centers Leverage Abundant, Inexpensive Renewable Energy”. Fair enough. But Quincy is located in Eastern Washington, an area that is largely desert-like and relies exclusively on the power produced by the Wanapum and Priest Rapids dams. What would happen if this area fell into a period of pro-longed and severe drought?
We actually have a pretty good idea of what would happen since the area experienced a severe drought in 2001 and a slightly lesser one in 2005. In both instances, water restrictions forced the lay-offs of thousands of nearby aluminum factory workers as the water available wasn’t enough to generate the power to run the factories’ smelters. In both cases, electrical prices also rose substantially as power had to be imported from California.
What is interesting is the price of power was so high that Alcoa figured out it could make more money selling power than selling Aluminum.
ALCOA:
"Why Sell Aluminum When We Can Make a Bigger Profit Selling Electricity?"Jun 25, 2001
Alcoa Aluminum's main smelter in Washington state has been shut down, as have the smelting operations of four other aluminum companies on the Northwest Coast.
They weren't shut down because of a lack of buyers for aluminum –nor for a lack of workers. The big aluminum producers shut down simply because they could make more money by reselling electricity than they could by producing and selling aluminum.
Alcoa, for example, had longterm contracts to buy electric power from the federal government's Bonneville Power Administration at a rate of $22 a megawatthour. But deregulation of the electric power industry opened the road to enormous price increases. By last winter, electric power was selling on the open market at rates running between $250 and $500 a megawatthour. Shutting down aluminum production to resell electricity let Alcoa pocket the difference.
The result was an enormous increase in Alcoa's profits. Its profits in the first quarter of this year 120 million dollars - were over ten times as high as what they had been in the same period a year ago, before they started selling electric power.