Last year Google announced a goal to power all its cloud data centers with carbon-free energy 24 hours a day by 2030. This Wednesday Google Cloud pitched an option to help tide its climate-conscious customers over until it reaches that very ambitious goal.
The company has published data that shows what percentage of time Google data centers in most of its cloud availability regions run on carbon-free energy on average. Its metric for representing this is CFE%, or carbon-free energy percentage.
“Just like the potential differences in a region’s price or latency, there are differences in the carbon emissions associated with the production of electricity that is sourced in each Google Cloud region,” a Google blog post read.
“The CFE% will tell you on average, how often that region was supplied with carbon-free energy on an hourly basis. Maximizing the amount of carbon-free energy that supplies your application or workload will help reduce the gross carbon emissions from running on it.”
If carbon emissions are part of your company’s computing infrastructure decision making, you can use this data as part of your cloud region selection process. Many other factors go into that decision – things like latency, cost, cloud-provider feature availability, data storage location regulations, and so on – and Google’s new data can add carbon emissions as another factor.
Google Cloud’s announcement features a statement by Patrick Flynn, VP of sustainability at Salesforce, a customer that’s eager to use CFE% to “prioritize locations that maximize carbon-free energy.”
Historically, Google has relied on long-term renewable energy Power Purchase Agreements to bring clean energy to the local grids that power its data centers. Google says it has fully matched its total energy consumption with wind and solar PPAs since 2017.
Wind and solar are both considered renewable energy sources. The company's 10-year goal is to use carbon-free energy sources around the clock. As a Google spokesperson pointed out to us, "carbon-free" is different from "renewable," in that it applies to more types of energy sources. Geothermal energy, for example, is considered carbon-free but not renewable.
But every grid Google’s data centers are on still relies on a mix of generation methods and there’s no Google Cloud region that’s powered by carbon-free energy around the clock today.
The CFE% metric is calculated using data about hourly generation feeding a grid and clean energy on the grid produced through a Google PPA. Average hourly CFE percentage for one year (2019) was aggregated to produce the numbers released Wednesday.
Of all Google Cloud availability regions for which data is available, us-west1 in Oregon is the greenest, with a CFE of 89 percent. The dirtiest is asia-southeast1 in Singapore, with a CFE of 3 percent.
There’s no CFE% available for most other Asian availability regions except Taiwan (19 percent). So, unless your choice is between Singapore and Taiwan, the Singapore metric isn’t very useful at this point.
Since Google has yet to reach its goal of powering all its data centers with carbon-free energy around the clock, it included each region’s grid carbon intensity metric, or emissions factor, which shows average carbon emissions per unit of energy generated on a grid.
Google suggests you can use the emissions factor to compare two regions with similar CFE%. It’s a way of knowing how carbon-intensive your cloud infrastructure will be when it’s not running on carbon-free energy.
In an example Google used, while Frankfurt and the Netherlands both have a CFE% of 61 percent, the grid in Netherlands has a higher emissions factor. So, if have to choose between Google Cloud’s europe-west3 region, hosted in Frankfurt, and europe-west4 region, hosted in the Netherlands, and there are no other reasons to go with one or the other, you’ll go with europe-west3 if you want to be in a greener data center.