 |
Question of the WeekThe Naked Scientists: Science Radio & Science Podcasts |
 Science Questions RSS Feed |
Click here to listen
|
How much energy is used when you do a Google search? Christianne |
|
We put this to Eric Teetzel, Program Manager for RE(less than)C at Google:
|
March 2009 |
|
I can't remember the answers, but someone came up with a figure which was bandied around in the press about 6 weeks ago - and looked absurdly high. A few days later Google released its own estimate which was something like 100x less than the previous figure - and felt more realistic. Take the total power used by Google's servers (or whole business if you like) and divide by the number of searches. Presumably you want to apportion the "idle time" between searches also to a search? Or not? A vastly greater energy must be consumed by all the people across the world who now stay up later, with their computers, lights and heating on doing the "Googling" and surfing when a few years ago they'd have been tucked up in bed using very little energy! - techmind - 24th Feb 09
This was first reported in The Times: http://technology.timesonline.co.uk/tol/news/tech_and_web/article5489134.ece Then Google responded: http://www.usnews.com/blogs/fresh-greens/2009/1/13/google-dispels-energy-usage-claims.html?s_cid=rss:fresh-greens:google-dispels-energy-usage-claims - paul.fr - 24th Feb 09
I know this isn't a very scientific reply but I find it hard to believe that is how much power is used. I could justify having a shower every day on that basis considering how many searches I do.
- turnipsock - 25th Feb 09
All these people googling to come with an answer can't have done much for the environment!
- Richard1964 - 26th Feb 09
To have this topic resemble anything like a balance, the Google method should be compared to the energy input needed to have knowledge shared in the pre-electronic/digital era. I feel that the chopping down of trees, production of ink, smelting of metal to manufacture printing presses etc. won't come to seem very efficient. Just a thought..... - wannabe - 27th Feb 09
I was just going to say - I wonder how much energy has gone into researching this answer...?
- chris - 28th Feb 09
the number cited by BBC is 0.0003 kWh instead of 0.3 kwH per search. 0.3kwH is definitely too high.
See the whole discussion | Make a comment- a listener - 12th Mar 09
| |
RE(less than)C is an initiative that we started to advance technology and renewable energy, to make it cost competitive with fossil fuel power generation. We’ve done the calculations internally and I think anybody that’s tried to do carbon accounting understands there’s a lot of complications and nuance. The basic premise is that one Google search uses about 0.0003KWh worth of electricity. That’s ± some and that then translates, based on emissions load, into somewhere around 0.2g of CO2 per search that we answer. To actually do the things that we do all of our online services require machines. Those machines are basically all housed in facilities we call data centres. Those machines are typically servers and networking equipment. The way in which we do the energy calculation per query – we look at not just the exact machines that touched the query as it comes in our data centre but we also look at allocating networking routing costs as well as what we would call just ‘overhead.’ It takes energy to build the index to be able to effectively answer the query as they come in. We also allocate those costs across all of our search presence. That’s how we come up with the number of 0.2g per search or 0.0003kWh worth of electricity.
The contents of this site are © The Naked Scientists® 2000-2010. The Naked Scientists® and Naked Science® are registered trademarks.