Skip to Content

Shrunken Servers Aim for a Greener Internet

Intel teams up with a startup to create a server twice as efficient as those that power websites and apps today.
January 31, 2012

As the cloud becomes more pervasive—driving everything from social networking to mobile apps—the computers that power it must guzzle more and more energy. Today, startup company SeaMicro, chip maker Intel, and electronics giant Samsung unveiled a new computer design that could make the data centers that power cloud services dramatically more efficient.

The new server design uses half the energy and takes up a third of the space of servers that do the hard work in most data centers today.

Mozilla, the nonprofit browser maker, and CompSec, which builds cloud computing infrastructure for federal intelligence agencies, are already using the new designs.

SeaMicro, based in Santa Clara, California, has received $60 million in venture funding and a $9.3 million U.S. Department of Energy grant intended to encourage greater efficiency for cloud infrastructure.

Until now, the startup was known for making small, efficient servers suitable only for low-powered jobs in a data center. These servers use Intel’s Atom chips, originally developed for mobile computers. The new design keeps many of the same efficient features but swaps in a conventional server chip from Intel’s Xeon line, allowing it to do more computationally taxing tasks like running the vast databases at the heart of social networks or online stores.

“We were able to address an important portion of this rapidly growing market [before], but by no means all of it,” said SeaMicro’s CEO and cofounder, Andrew Feldman, at an event on Tuesday in San Francisco. “This allows us to address the entire data center.”

Many people in the industry are eagerly awaiting the arrival of servers built around smart-phone chips, which are highly energy-efficient, as a way to cut data center costs. Those chips are made by a variety of companies, including Qualcomm, Marvell, and Texas Instruments, using designs licensed from U.K. company ARM, a shift that threatens to dilute Intel’s dominance of the server chip market. Hewlett-Packard announced late last year that it would launch a line of servers with cell-phone chips inside, in partnership with startup Calxeda.

Mini motherboard: SeaMicro has developed custom chips (the four along the base of this motherboard) that take over the work of components scrapped to save space and energy.

Intel’s partnership with SeaMicro could help it fend off that threat. ARM-based chips are promising but not yet available in versions capable of the more powerful 64-bit computing tasks that data centers require, points out Jean Bozman, an analyst with IDC. SeaMicro’s approach, which combines Intel’s proven 64-bit server chip with a stripped-down, more efficient design, could gain a foothold first. “This is already shipping to customers,” she points out.

SeaMicro can deliver energy and space savings without a mobile chip because it found ways to scrap components that make servers inefficient, said Feldman at Tuesday’s event. “We are ripping off gobs of components that suck power like vampires,” he said. “Our patented technology allows us to eliminate all but three components from a [server’s] motherboard: the processor, the RAM, and our own chip.”

Conventional motherboards also feature network cards, BIOS chips (which help a computer start up), fans, USB connectors, LED lights, and other components that constantly use power and sap efficiency, said Feldman. SeaMicro’s chip does the job of the needed components—such as BIOS chips and network links—in a more efficient way. For example, the SeaMicro chip can power down its networking capabilities when they’re not needed. “It’s an exercise of removing the death by a thousand cuts that you normally get,” said Feldman.

As a result, the motherboard inside one of SeaMicro’s new servers is just 11 inches long and five inches wide. A conventional server motherboard is around two feet long by 18 inches wide.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.