Internet-centric businesses could cut their energy costs as much as 40% by shifting their data dynamically to wherever energy prices are lowest at any particular time, according to a study from researchers at MTI, Carnegie Mellon University and Akamai Technologies (NASDAQ: AKAM).
Akamai helped contribute real-world routing data to test the thesis in an MIT PhD student’s 2008 paper. Researchers analyzed 39 months of electricity price information in 29 major cities and found prices fluctuated widely in different cities at different times. The researchers designed a routing scheme that would take advantage of those fluctuations, taking into consideration the cost of moving data around the country dynamically, and saw substantial savings.
Bill St. Arnaud at CANARIE, the Canadian research concern, spoke with me about this concept (and others) in this podcast (you can also read this article). This concept has been demonstrated as far back as 2006, when the APEC TEL (Asia-Pacific Economic Cooperation Telecommunications and Information) Working Group used rapidly provisioned dynamic lightpaths to migrate live virtual machines from Calgary to Chicago (by way of South Korea!) without adding more than about 1 second of application downtime.
The rising popularity of both green initiatives and cloud computing is helping to draw more attention to the energy-aware Internet routing model. And the participation of influential telecom vendors like Akamai is helping to prove the real-world business case.