Google Search's Carbon Footprint: Exactly How Much CO2 is Released?

Recently, attempts by giant search engine firm Google to promote its eco-friendly policies have blossomed into a debate about the "real energy value" of one tiny search. Just a simple point and click, a few keywords in that simple text box - but identifying how much energy is used in a simple search transaction gets a little complicated. For starters, there's the difference between measuring the client-side (energy used on your own computer) and the server side (what energy Google generates back at its collection of supercomputers).

That's just the beginning of a vigorous argument over something not a lot of people cared about before powerful voices began speaking for the urgency of eco-initiatives and greening IT. Well, it matters now. So how about a few opinions on just how green a Google search is?

click to enlarge

It's important to note that even those claiming higher energy use for a search (including the original Harvard professor, according to reporting) aren't singling out Google but identifying a general phenomenon...in fact, Google has introduced some cutting-edge responses to the issue of overloaded servers that are driving a beneficial force to ease the "power strain" from so many users piling onto a search engine (any search engine) at once.

Then there was the response of those affiliated with the company: a report on the online Data Center Dynamics site shows a high-level Google official calling 7 grams "many times too high". The counter-claim here suggests that because of the extremely fast response of a Google server, the real number is closer to .2 grams! And, this argument produces a different relative "real-time" task for setting up an equivalent: estimating an average daily human food intake, the Google representative claimed the search would account for roughly the energy that a human body burns in 10 seconds!

A casual bystander might offer a few more suggestions, especially is he/she were an interjecting type: every search is different. What kind of search are we talking about? One that only turns up a few results or one that returns thousands? What about keyword length? And what if the server needs to do extra work to parse out returns for a .PDF page, or some other special result? What about the one-shot "Lucky" search option?

The point is that the specific median of this debate over a "simple search" is hard to define, even if we restrict our measurements to just the server side. One reason is because it's hard to estimate how a server works in response to any client activity. Another is that the search itself is a varying element.

But these recent differences of statement have given us some handy benchmarks to work with to define how much energy our computers actually use, and to illuminate the server side costs of client activities in general.

Some new possibilities are opening up out of this kind of research and debate over the specifics of greening IT. One such is the outcropping of "dark" search engine options like EcoSearch, an alternative search engine software operation that does not light up as many pixels on the screen. Experts have differing opinions on the saved value from a simple search on one of these "Eco-engines" but its appearance shows us how much consumers and others are now looking at collective energy use issues. We can learn from these news stories and emerging "anecdotes" in pioneering changes that we can all be proud of.