In April of this year, Google announced that it is taking the next step in making its data centers greener and cleaner. The company indicates that it has been carbon-neutral since 2007, and it has covered its energy consumption with 100% renewables since 2017.
Many large corporations have undertaken similar commitments, covering the equivalent of their total electricity use with renewable energy from Power Purchase Agreements (PPSa). These PPAs match their total electricity consumption to the output of a new “additional” renewable facility built on their behalf. However, only the total volumes match, not the actual physical flows of power. For example, if a company were to offset its 100 megawatthours (MWh) of consumption with 100 MWh produced by a solar facility, then at times unused surplus solar would be sold into the market while at other moments (nights, for example) the company would be buying system power from the grid at whatever carbon intensity the grid was offering at that moment.
Carbon intelligent computing
To address that challenge Google has been working to create “24 X 7 carbon-free energy,” developing a “carbon-intelligent computing platform,” that has already been deployed in a large data center. This approach will enable Google to shift the timing of non-critical various compute assignments to periods when renewables are more abundant on the grid.
How is this achieved?
Every day, Google’s carbon-intelligent platform creates and compares two forecast for the following 24-hour period. The first forecast predicts the likely utility dispatch curve - that tells generating plants when to operate - and hourly average carbon intensity of the associated generating assets. A second internal Google forecast evaluates the predicted hourly energy consumption necessary to perform specific compute tasks over that same duration.
Google then uses both forecasts to create an optimised hourly compute hierarchy that is best aligned with times of the lowest-carbon electricity supply. Preliminary results suggest that carbon-aware load shifting strategy is effective, and that there are enough elastic compute tasks capable of being moved to make the strategy work.
Where might this go?
Google indicates that the first stage of this approach looked at shifting compute tasks with a given data center to different and less carbon-intensive timeframes on the grid. The next and more complex stage is to actually move the tasks between data centers, shifting that compute load to the regional power grid that is least carbon-intensive at any given time. The company plans to share the results of these endeavors in future research publications at an unspecified date.
Is there an easier way?
The 1968 Comedy starring Jerry Lewis called “Don’t Raise the Bridge, Lower the River” suggests that there are often both easier and harder ways to achieve the same outcome. In the case of Google, it’s working on this compute shifting challenge because it finds itself with data centers all over the world, with differing degrees of carbon intensity in each associated power grid. In this unavoidable scenario, such a laborious approach then makes sense, and Google should be lauded for it.
But for companies looking to decarbonise their own compute requirements, there may be another, simpler way to accomplish the same outcome. And that is simply to locate one’s computing tasks in a data center situated in a power grid that is already carbon free – and low cost besides.
One such location is Iceland, where Verne Global has located its high-performance data center for just that reason (the Icelandic grid’s high reliability standards don’t hurt either). Customers outsourcing their compute jobs to Iceland essentially accomplish the exact same outcome as Google. With a power grid supplied by 100% hydroelectric and geothermal power, datacenters are always green, without the need for complex forecast matching, algorithms, and scheduling gyrations. Just sayin.’