Google Wants To Schedule Data Center Workloads

April 25, 2020 Off By Naveen Victor

Photo Credit: Google Cloud Products

Google is resorting to ingenious ways of reducing its carbon footprint. The search giant employs the services of massive data centers located around the world to maintain its search engine, YouTube, email, cloud storage solutions and other services.

Since 2007, the company’s operations have been deemed carbon neutral, meaning that there isn’t a net release of carbon dioxide to the atmosphere. Last year, marked the third year in a row that Google has matched its energy consumption with 100% renewable energy purchases.

Now, its taking its green initiative a step further by working toward a 24×7 carbon-free energy system for all its data centers. In order to achieve this, its power hungry sites would need to to rely heavily on renewable power sources that include solar and wind.

Photo Credit: Google

One of its cleverer initiatives involves the use of a carbon-intelligence computing platform. Developed by a small team of engineers, this system will shift the timing of many compute tasks of data centers to a period when power sources like wind and solar are most plentiful.

The company doesn’t foresee this impacting the performance of its services like search, Maps and YouTube. In fact, it won’t even require the use of additional computer hardware to operate. Shifting the timing of when non-urgent compute tasks like video processing or photo editing are executed, is essential to Google’s quest in achieving its 24×7 carbon-free energy.

Photo Credit: Google

The way it works is, Google’s carbon-intelligent platform compares two types of forecasts for each following day. The first is provided by a company called Tomorrow, which predicts the average hourly carbon intensity of the local electrical grid. The other comes from Google, in which it predicts the hourly power resources that a data center needs during that period.

Google isn’t stopping here either. Eventually, it plans on moving flexible compute tasks across data centers where the yield of their renewable power sources are at their peak. This allows the company flexibility in assigning tasks to data centers that have power to spare, bringing the company ever closer to its goal of 24×7 carbon-free energy.