It’s no secret that access to endless energy is the lifeblood of data centers. At the same time, energy costs vary widely depending on location. What if data centers could use that price variability to their advantage? That’s the question asked by researchers at MIT, Carnegie Mellon, and networking company Akamai. According to the researchers, Internet businesses that swallow up energy (think Google and Microsoft) could save millions of dollars by using an Internet-routing algorithm that moves data to wherever electricity costs are lowest at a particular time.
There are many reasons for price variability–changes in supply, time of day, consumer demand, and fuel prices. By taking into account the cost of rerouting information versus the potential savings from lowered energy costs, researchers found that companies could potentially cut energy consumption by 40%. With data center usage expected to quadruple in the next decade, these findings could have positive implications for the continued growth of data center-reliant organizations. And as companies like Apple, which recently announced that it is building one of the world’s largest data centers in North Carolina, become more invested in energy-intensive cloud computing initiatives, cost concerns are only going to become more common,
There are some caveats to the energy-routing system. It cuts costs, but not necessarily overall energy use or pollution. And as Jonathan Koomey, a Lawrence Berkeley National Laboratory scientist, explains “The trick is to be able to control these systems well enough and to create controls that are cheap enough to be able to take advantage of
the arbitrage opportunity available from differential electricity
prices, without affecting reliability or latency”.
We don’t have yet have this control, but in the meantime data centers have a number of other options to cut energy use and potential costs. IBM, for example, is working on a super-energy efficient data center that will use 50% less energy than existing data centers. And yes, that’s without sending data all over the place in the hopes of cutting costs.
[Via MIT Technology Review]