The inexorable growth in the volume of data that is produced is undeniable but putting a precise number on the future level is problematic and estimates vary greatly. But if you consider that 90 percent of all the data in the world has been created in the last two years it gives you some idea of the magnitude of this escalating challenge. This evolution in power hungry computer infrastructure must be set against the challenge that all industry faces in addressing the threat of climate change that demands a transformation in how the world produces and uses energy.
The data sector is increasingly looking at ways to improve its sustainability as national and international targets loom towards the end of the current decade. The utilisation of renewable energy is increasing, whether directly or using power purchase agreements, but the more enlightened organisations are looking beyond that and towards the decoupling of sectors that involves a greater integration with the power sector.
It is now 13 years since one of the leading quintet, Google, made a voluntary commitment to become carbon neutral. They validated this by calculating its global carbon footprint, purchasing high-quality carbon offsets, and working with a third-party to certify the calculations and validate its offset portfolio. But since that time its ambitions have evolved. Last year marked the third year in a row that it matched its energy usage with 100 percent renewable energy purchases.
“This is an important milestone on the way to our ultimate goal: to source enough carbon-free energy for our operations in all places, at all times,” Ana Radovanovic, technical lead for carbon-intelligent computing at Google, says. “Now, we are working toward 24/7 carbon-free energy everywhere we have data centres, which deliver our products to billions of people around the world. To achieve 24×7 carbon-free energy, our data centres need to work more closely with carbon-free energy sources like solar and wind.”
New carbon-intelligent computing platform
To achieve, this Google are adopting a tactic common within the power grid community called demand management. The concept is that by monitoring the power factor of the power grid, as well as their own control parameters, individual, intermittent loads would switch on or off at optimal moments to balance the overall system load with generation, reducing critical power mismatches.
Google’s latest step on the path to sustainable digital infrastructure is a new carbon-intelligent computing platform that has been developed by a small team of engineers and mirrors this approach. “We designed and deployed this first-of-its-kind system for our hyperscale data centres to shift the timing of many compute tasks to when low-carbon power sources, like wind and solar, are most plentiful,” Radovanovic, adds. “This is done without additional computer hardware and without impacting the performance of Google services that people rely on around the clock.
“Shifting the timing of non-urgent compute tasks—like creating new filter features on Google Photos, YouTube video processing, or adding new words to Google Translate—helps reduce the electrical grid’s carbon footprint, getting us closer to 24/7 carbon-free energy.”
It works by performing two daily forecasts at every Google data centre. One of the forecasts, provided by partner, Tomorrow, predicts how the average hourly carbon intensity of the local electrical grid will change over the course of a day. A complementary Google internal forecast predicts the hourly power resources that a data centre needs to carry out its compute tasks during the same period. The two forecasts are then combined to optimise hour-by-hour guidelines to align compute tasks with times of low-carbon electricity supply.
“Early results demonstrate carbon-aware load shifting works,” Radovanovic continues. “Results from our pilot suggest that by shifting compute jobs we can increase the amount of lower-carbon energy we consume.”
The first version of this carbon-intelligent computing platform focuses on shifting tasks to different times of the day, within the same data centre. But it is also possible to move flexible compute tasks between different data centres, so that more work is completed when and where doing so is more environmentally friendly. “Our plan for the future is to shift load in both time and location to maximise the reduction in grid-level CO2 emissions,” Radovanovic concludes. “Our methodology, including performance results of our global rollout, will be shared in upcoming research publications. We hope that our findings inspire other organisations to deploy their own versions of a carbon-intelligent platform, and together, we can continue to encourage the growth of carbon-free electricity worldwide.”