Hamilton wrote in a recent blog post that US data centres were consuming power at a rate equal to 1.5% of America's total power consumption in 2005. The EPA predicted it would jump to 3% or more by 2010. Truth be known, actual power consumption four years ago was between 1% and 1.5%. It continues to remain the same today.
The primary factor in keeping power consumption constant while continuing to expand the number and size of the world's data centres is one of technology. In other words, technology has increased energy efficiency nearly 50% over the last decade.
One of the most important technological advancements has been cloud computing. It has allowed the industry to grow considerably without increasing the amount of power consumed and, what's more, it has opened the door to more green options. According to Hamilton, we have now reached a point where new data centres are capable of generating more power than consumed by way of renewable sources. Doing so means the data centres can actually contribute to the power grid rather than taking from it.
Hamilton cited a new US data centre being built by Facebook in Altoona, Iowa. Although Facebook is not generating its own power on-site, it is partially funding a nearby wind farm project that will provide 100% of the power needed to run the facility. The data centre will be connected to the grid for those times when wind is just not getting the job done, but it is expected that the wind farm will generate more than enough on most days.
Mr Hamilton makes the case that the open door to more green options does not necessarily mean data centres should be generating their own power on-site. His belief rests in the fact that data centre operators are experts in hosting digital information, not producing electricity and maintaining power generation capabilities. He sees far too many things that could potentially go wrong if data centres got into the power generation business.
He believes the approach taken by Facebook is a far better way to go. Let data centre operators focus their attentions on more efficient computer systems that will require less power, while allowing power companies to focus on generation and transmission. That is the best way to serve everyone's needs with as few problems as possible.
The EPA was incorrect in its predictions largely because it did not take into account innovations and efficiency. Since then, we have learned our lesson about making predictions for the future. Instead, we are focusing our attention on making things better in any way we can. That's the way it should be.