• A new feature could almost miraculously cut energy consumption of data centers by 30%
• Interrupt request suspension dynamically alters CPU power usage and can be done via OS
• Hyperscalers are likely to be the big winners and it will be interesting to see how it impacts AI

Data centers reportedly account for between 2-4% of total electricity consumption around the world, something hyperscalers are understandably looking to reduce wherever possible.

Potential solutions include implementing next-generation architectures like Hyperconverged Infrastructure (HCI) and employing advanced cooling techniques.

Professor Martin Karsten at the Cheriton School of Computer Science, within the University of Waterloo in Ontario, Canada, has a cheaper, easier solution. He claims that data center energy consumption could be cut by up to 30%, simply by changing a few lines of Linux code.

Small change, big impact

Working with Joe Damato at Fastly, Professor Karsten has developed a small, non-intrusive kernel change of just 30 lines of code that uses IRQ (interrupt request) suspension to reduce unnecessary CPU interruptions and improve Linux’s network traffic processing. This tweak has now been published as part of Linux’s newest kernel, release version 6.13.

This code change, which reportedly improves Linux networking efficiency and boosts throughput by up to 45% without increasing latency, is based on a research paper called “Kernel vs. User-Level Networking: Don’t Throw Out the Stack with the Interrupts”, which Professor Karsten authored with former master’s student Peter Cai in 2023.

“We didn’t add anything,” Professor Karsten said of the code change. “We just rearranged what is done when, which leads to a much better usage of the data center’s CPU caches. It’s kind of like rearranging the pipeline at a manufacturing plant so that you don’t have people running around all the time.”

The Professor believes this small adjustment could have a huge impact. “All these big companies – Amazon, Google, Meta – use Linux in some capacity, but they’re very picky about how they decide to use it. If they choose to ‘switch on’ our method in their data centers, it could save gigawatt-hours of energy worldwide. Almost every single service request that happens on the Internet could be positively affected by this.”

Aoife Foley, IEEE senior member and professor in the School of Mechanical and Aerospace Engineering at Queen’s University Belfast, welcomes the potential savings but observes that it’s going to take a lot more than just changing a few lines of code to address the broader energy challenges.

“There is a long way to go yet,” she says. “These facilities represent huge electricity demands, adding pressure to electricity grids and increasing the challenge of energy transitions, especially in smaller countries. Although it is impossible to calculate precisely, the entire ICT sector is estimated to account for about 1.4 per cent of CO₂ emissions globally. Infrastructure and operations leaders have a responsibility here and need to consider the unnecessary waste associated with data storage and commit to generating power from more renewable sources.”

Yandex recently released an open source tool called Perforator, which takes a similar approach to Professor Karsten’s research, helping businesses optimize their code, reduce server load, and ultimately lower energy and equipment costs.

Sergey Skvortsov, who leads the team behind Perforator, told us: “This latest research confirms what we have long believed: optimizing code is one of the most effective ways to reduce data center energy consumption. Perforator helps businesses identify and fix inefficient code, cutting CPU usage by up to 20% and reducing infrastructure costs – without sacrificing performance. With data centers consuming up to 4% of global electricity, tools like Perforator can play a crucial role in making tech infrastructure more sustainable.”

You might also like

• Why the data center of 2030 will look very different from today
• Data centers are transforming waste heat into community energy assets
• The AI boom is spurring an energy crisis: is there a sustainable way out?