Microsoft publishes methodology for data centre cooling lifecycle assessment

Microsoft has developed a methodology to assess the lifecycle emissions, energy and water impacts of various data centre cooling techniques.

The study will help Microsoft make decisions about new data centre designs and cloud operations, but it also aims to support transparency in the tech sector : Microsoft has published its findings in Nature and created an open research repository where others can add in their own data and scenarios to conduct their own lifecycle assessments.

“What we’re trying to do here is tell the industry, ‘Here’s how you build an end-to-end life cycle assessment that takes cooling into account. And here is a tool for you that you can customise to your specific needs and then make a decision,” Husam Alissa, director of systems technology in Cloud Operations and Innovation at Microsoft and leader of the lifecycle assessment study, explained.

Immersion vs cold plate cooling

Microsoft’s research compared the water consumption, energy demand and greenhouse gas emissions associated with different types of cooling techniques, across the entire lifespan of data centres.

According to its finding, renewable-powered cold plate cooling, a method through which a metal plate cooled with liquid is attached to a central or graphics processing to absorb its heat, reduces emissions and water consumption compared to air cooling.

This newer technology, which Microsoft has started to deploy across its data centres, is estimated to reduce greenhouse gas emissions and energy demand by around 15% and water consumption by 30 to 50% across the datacentres’ life spans, the study has found.

However, the tech giant also looked at other cooling technologies, including one and two-phase immersion, where servers are installed vertically in a coolant bath.

“Our intention is not to say, ‘this is the right technology.’ They all could be. There are different circumstances that make you use a technology,” Alissa added.

More work needed to study lifecycle impact of AI chips

In a blog post about the research, Microsoft warns that the study only analysed the cooling needs of central processing units of CPUs, used for general compute power – not the specialised chips designed to handle AI workloads (generally graphics processing units of GPUs). 

“The team is working on a follow-up to examine life cycle impacts of AI chips and expects to see similar improvements with advanced cooling methods,” the company added.

Learn more about efforts to quantify the environmental impacts of AI in this CSO Futures webinar recording with AVEVA, the Green Software Foundation and Tech Mahindra.