We noticed you're browsing in private or incognito mode.

To continue reading this article, please exit incognito mode or log in.

Not an Insider? Subscribe now for unlimited access to online articles.

Intelligent Machines

Computer Clusters That Heat Houses

A novel water-cooling system makes it more efficient for computers to heat buildings.

They have been used to model climate change, forecast economic trends, and simulate the intricate complexities folding proteins. Now IBM has something new in store for high-performance computers: heating buildings.

Keeping cool: A novel on-chip water-cooling system developed by IBM could make it efficient for data centers to provide waste heat for buildings.

Thanks to a novel on-chip water-cooling system developed by the company, the thermal energy from a cluster of computer processors can be efficiently recycled to provide hot water for an office, says Bruno Michel, manager of advanced thermal packaging at IBM’s Zurich Research Laboratory, in Switzerland. The goal, he says, is to improve the energy efficiency of large computing clusters and reduce their environmental impact.

A pilot scheme involving a computer system fitted with the technology is expected to save up to 30 tons of carbon dioxide emissions per year–the equivalent of an 85 percent carbon footprint reduction. A novel network of microfluidic capillaries inside a heat sink is attached to the surface of each chip in the computer cluster, which allows water to be piped to within microns of the semiconductor material itself. Despite its close proximity to the circuitry, there is no danger of leakage, says Michel, because the capillaries are hermetically sealed. By having water flow so close to each chip, heat can be removed more efficiently. Water heated to 60 °C is then passed through a heat exchanger to provide heat that is delivered elsewhere.

IBM has spent several years developing the microfluidic cooling technology, and it plans to test it in partnership with Swiss Federal Institute of Technology, in Zurich. A 10-teraflop computer cluster consisting of two IBM BladeCenter Servers in a single rack will be used by the university’s Computational Science and Engineering Lab to model fluid dynamics for nanotechnology research. The water will then be plumbed into the university’s heating system, where it will help heat 60 buildings. “This is the first large-scale system,” says Michel. “It’s about one-twentieth of the size of an average data center.” Ultimately, he says, the technology could help address the energy problems posed by large data centers.

Up to 50 percent of the energy consumed by a modern data center goes toward air cooling. Most of the heat is then wasted because it is just dumped into the atmosphere. There have been a few efforts to recycle the heat generated by conventional data centers. For example, a nine-story, 18,500-square-meter data center being built in London by the hosting company Telehouse Europe will provide heating for nearby offices. Other companies, including IBM, have used excess thermal energy to heat green houses or swimming pools. But reusing waste heat is expensive because usually only relatively low temperatures can be harvested, says Frank Brand, director of operations of the Dutch data-center engineering firm Imtech. “You can only get about 30 to 35 degrees Celsius,” he says.

In contrast, because water is many times more efficient at capturing heat than air, water cooling can deliver much higher temperatures, says Michel. Water was once commonly used to cool mainframe computers, but this merely consisted of piping cold water through server cabinets to cool the air near the racks.

By some estimates, information technology infrastructure is responsible for as much as 2 percent of global carbon emissions, putting it on a par with aviation. And some experts say that this figure is set to double in the next five years.

“It’s more efficient to heat water and move it somewhere else than it is with air,” says Jonathan Koomey, a project scientist at Lawrence Berkeley National Laboratories and a consulting professor at Stanford University. In 2005, data centers were responsible for 1 percent of global electricity–a doubling of 2000 levels, Koomey says. But he’s not convinced that the figure will continue to grow. “There are many ways to improve the efficiency of data centers,” he says. For example, better management of computer centers can improve efficiencies dramatically. “We have servers that on average are running at 5 to 15 percent of their maximum load,” Koomey says. “Even if the server is doing nothing, it’s still using 60 to 70 percent of its power.”

Brand also notes that “air is a much cheaper way to do the cooling” and that modern data centers consume far less energy than do their older counterparts for cooling.

The trend toward stacking processors on top of each other to increase their power density is another reason why IBM is pursuing this sort of microfluidic water cooling, says Michel. Such three-dimensional chips will pose serious problems for traditional air-based cooling systems, he says.

Be the leader your company needs. Implement ethical AI.
Join us at EmTech Digital 2019.

Register now
More from Intelligent Machines

Artificial intelligence and robots are transforming how we work and live.

Want more award-winning journalism? Subscribe to Insider Plus.
  • Insider Plus {! insider.prices.plus !}*

    {! insider.display.menuOptionsLabel !}

    Everything included in Insider Basic, plus the digital magazine, extensive archive, ad-free web experience, and discounts to partner offerings and MIT Technology Review events.

    See details+

    Print + Digital Magazine (6 bi-monthly issues)

    Unlimited online access including all articles, multimedia, and more

    The Download newsletter with top tech stories delivered daily to your inbox

    Technology Review PDF magazine archive, including articles, images, and covers dating back to 1899

    10% Discount to MIT Technology Review events and MIT Press

    Ad-free website experience

You've read of three free articles this month. for unlimited online access. You've read of three free articles this month. for unlimited online access. This is your last free article this month. for unlimited online access. You've read all your free articles this month. for unlimited online access. You've read of three free articles this month. for more, or for unlimited online access. for two more free articles, or for unlimited online access.