Kettles and kilograms
- 13 January 2009
By Molly Webb, Head of Smart Technologies, The Climate Group.
I have a lot of questions about this headline, some of which are addressed in the frenzy of articles and blogs around this headline (see my google tag on delicious), especially the rebuttal saying the original study never mentioned Google!
My first reaction is to try to put headlines like this one into perspective. Here are my initial thoughts - it would be great to hear yours.
Let's start with the kettle:
- What kind of kettle? I live in the UK but I am American and no Americans I know use electric kettles. So we use an old fashioned kettle on a stovetop. And apparently that stovetop will take approximately the same amount of energy as the electrical version, though it probably depends on your gas or electricity supplier.
- How much water is in your kettle? Overfilling the kettle will mean wasting electricity, but boiling just the right amount is more efficient. (Who does this?)
Next, the search calculation:
- Was the calculation based on actual electricity consumption? Or was it based on estimates of average electricity consumption of data centres, networks and PCs? The figure will change depending on region, efficiency of networks and power supply, among a number of other factors.
- What's included? Should we include the use of your PC in the google search? Do you have one monitor on, or two? What kind of monitor (LCD or CRT)? Or are you using your blackberry or mobile device to search? Who is 'responsible' for the device you use to access the internet? - Google (the service provider), the OEM, the teleco or you for buying the device and using the service?
- Who's numbers do we believe? Right now it's a Harvard prof against Google - claims, counter claims and counter-counter claims. This is very hard to calculate, as we learned in SMART 2020. And there is no standard for measuring network energy consumption, though standards for PCs are further along.
- What's the baseline? Should we compare the .2 seconds for this search with driving to the library and spending a few hours with the card catalogue and heating the library while you're there (but also keeping the stacks cool)? How many other people are in the library with you, or searching at the same time?
- Is efficiency enough? Google and other IT companies are striving to make their data centers and equipment more efficient (Climate Savers Computing initiative, the Green Grid) but is that enough? Google is also aiming to make renewables cheaper than coal (RE<C) and investing in the smart grid, an area that we pointed out could lead to a couple billion tonnes of CO2e reduction by 2020 (SMART 2020)
- What does the .2 grams CO2 compare to? According to the report author's statistic of 7g CO2 per search, a Google search uses the same energy consumption as watching an online video. But can we compare it to Yahoo or MSN searches? Is gCO2 per search the right way to think about this? What about Google's energy consumption per employee? Or Google's energy consumption per year compared to its contribution to GDP? How do we measure companies and their services? There is no standard for this aside from the CDP, which is a good start on the company side but doesn't address services. If we use the numbers on Google's blog, one kilometer (0.6 miles) driven in a standard car produces as many greenhouse gases as a thousand Google searches - at 200 million searches a day that's 200,000 cars driving 1 km or 2,000 cars driving 100 km a day. That's a lot of cars! But.
- What if Google succeeds in completely greening the electricity that supplies its data centres and therefore its search? Right now, if we use Google's numbers, it looks like .2g/.0003kwh implies that the carbon intensity of the supply of that power is about 660 g CO2e, or about equivalent to the average carbon intensity of natural gas in Europe (ie: a lot better than coal at over 1000 kg CO2 and a lot worse than nuclear at nearly zero). This .2g (or 1-10g depending on who's you believe) figure will change dramatically depending on the supply of electricity to Google's data centres, to the network equipment that transmit data, and to your own home or office where you're using your laptop.
- What's included in the 2%? That doesn't include home electronics which other reports have estimated could consume up to 45% of electricity used in UK households (source: best named report ever, The Ampere Strikes Back, by the Energy Savings Trust). In other words, we might have bigger problems such as 30" TV screens.
- What about the other 98%? Can we put energy and carbon of ICT in perspective? SMART 2020 (The Climate Group, GeSI and McKinsey report) estimated that ICT is 2% of CO2e globally (confirmed by Gartner) and is growing in real terms in spite of a host of energy efficiencies due to sheer demand for ICT. However, transport is responsible for closer to 9% of emissions, and buildings and industry consume most of generated electricity globally. SMART 2020 found that applying ICT more rigorously to energy efficiency in industry, transport, buildings and the generation/distribution of power could lead to 15% of emissions reductions. Policies that drive this will be needed (see Digital Energy Solutions Campaign and Technology CEO Council's 'behind the green' in the US, and the EU's 2008 Communication on ICT and climate change)
What is needed? I could list many actions that are needed to achieve the 15% reductions across the economy, but on the 2% (ICT's own footprint), SMART 2020 pointed out that the second bullet point below is crucial - but this may take some time as it is certainly challenging.
A standard way of measuring ICT products alongside home electronics devices that mean we can compare the energy consumption of each (or even better, require smart meters in the home that measure consumption in real-time and allow us to retire high-consumption devices)
- A standard way of measuring IT and telecoms services so that we can compare for example whether we should use one cable or telecom company over another, or our laptop for searching vs our handheld device, or so that companies know whether their phone/internet service is more energy efficient than another.
- A price on carbon (but I risk becoming a broken record)
- And finally, I'd like to see governments and companies tackle the challenge of real time energy monitoring across infrastructure so we can make much better decisions based on efficiency alongside other business/social priorities.
How many times did I use Google in the process of writing this post? Searching for all those articles I bookmarked and a few data points/reports/links may have been 15 searches and therefore 3g CO2e according to Google or 105 g CO2e according to Dr. Alex Wissner-Gross.
At least my avatar in Second Life wasn't the one doing the searching! Anyway, I'm going to put the kettle on...