In May, a researcher with Hewlett-Packard Laboratories (HP Labs) presented a paper at the (ASME) 2010 4th International Conference on Energy Sustainability, in Phoenix, AZ, the US, entitled Design of Farm Waste-Driven Supply Side Infrastructure for Data Centers. The paper drafted a working model of a data centre that runs on power supplied by the manure of cows on a dairy farm.
Not long after that, Computerworld Singapore was given an interview with one of the celebrated authors of the paper, Chandrakant Patel, HP Fellow and Director of the Sustainable IT Ecosystem Laboratory at HP Labs. Patels supply and demand side approach to the creation of sustainable IT ecosystems was a primary guiding factor to the HP Labs proposition in the paper.
Below is the expurgated transcript of the interview we did on Vesak Day (May 28, 2010), 8.30-9.15 a.m. Singapore time.
When did this idea come to you guys at HP Labsa data centre fueled by manure?
Chandrakanth Patel: Our work in the data centre actually started in the late 1990s. We were looking at managing energy in a key resource centre but at that time we had started out with a focus on demand side management. So we were looking at things on the usage side. Asking such questions as: How do you use sensing and control to manage energy as a key resource?
We did a lot of work and by round 2002-2003 we had more or less concluded that demand side management was well under control. So we started looking at the supply side sometime in the mid-2000s. When we looked at the supply side, we started to ask: Where does energy come from? What is the source of energy? And in that context we started to look at a variety or sources and thats when we looked at biomass, nuclear, geothermalevery energy conversion technology that we could find in the books.
What we said thenand even today we say sowas we had to make sure that the application of whatever technology or energy conversion means we chose and worked with at the time had to be done in conjunction with other sources in a micro grid.
Let me sort of paint a picture. We envisaged a data centre with demand side control systems inside the data centre that take the service level objective for workloads coming in. That service level objectivereturning and responding to the workload in a given time. That way we would take the service level objective and, based on the response time needed, wed allocate the right resource in the right track, and provide the right amount of cooling in the right places. As for things we didnt want, wed shut down.
Sign up for Computerworld eNewsletters.