Select your localized edition:

Close ×

More Ways to Connect

Discover one of our 28 local entrepreneurial communities »

Be the first to know as we launch in new countries and markets around the globe.

Interested in bringing MIT Technology Review to your local market?

MIT Technology ReviewMIT Technology Review - logo

 

Unsupported browser: Your browser does not meet modern web standards. See how it scores »

{ action.text }

It may be stretching the point a little to call it, as Wired’s Cade Metz does, the “world’s fastest nonexistent supercomputer.” Amazon’s supercomputer–it built one recently atop its Elastic Compute Cloud (EC2)–exists alright, although it is virtual. Most salient, though, is the fact that Amazon promises to bring supercomputing power to, if not the masses, then at least to anyone with a big question and a decently-sized grant.

Amazon has been building virtual supercomputers on its cloud for a while now, entering the ranks of the top 500 fastest in 2010. Recently, though, Amazon’s offering climbed to become the 42nd fastest in the world. Amazon has been a major player in cloud services since 2006.

Though Amazon’s offering pales in comparison to some of those 41 faster supercomputers–Amazon can run at 240 teraflops, handling 240 trillion calculations a second; Fujitsu’s K Computer runs at 10.51 petaflops, or 10 quadrillion per second–EC2 represents a real potential disruption in the supercomputing market. Recently, a “Top 5 Pharma” customer (Amazon won’t specify who; indeed it seems to be avoiding much public comment about the supercomputer) ran a cluster with Amazon for seven hours at a peak cost of $1,279, per Ars Technica.

That may not sound cheap on the face of it, but consider the alternative: building your own computer cluster capable of similar feats. CNET dug up some numbers: the hardware to support K Computer runs $20 million, and Fujitu foots a $10 million electricity bill each year to run the thing. By contrast, running the EC2 supercomputer 24/7 for a year would cost you $11 million.

Jason Stowe, CEO of Cycle Computing, put it this way to Wired: “It’s just absurd… If you created a 30,000-core cluster in a data center, that would cost you $5 million, $10 million, and you’d have to pick a vendor, buy all the hardware, wait for it to come, rack it, stack it, cable it, and actually get it working. You’d have to wait six months, 12 months before you go it running.” By that point, he says, your research question may have refined itself, and you may need more cores, or fewer.

There are some naysayers to the whole cloud supercomputing idea. Some say Amazon will never give you the customer service you might need for fine-tuning questions. “You need a certain level of support, help with things like loading data off out disks and tweaking the performance of the cluster to suit our needs,” one loyal customer of Penguin Computing, a competitor, tells Wired.

But the bottom line is this: if you don’t need the fastest supercomputer in the world, if you don’t need a lot of hand-holding, and if you’re on a budget, Amazon’s option may be just the thing you need. The economics of the cloud (see our interactive diagram from 2009 for a vivid illustration) are–like a supercomputer itself–overwhelmingly powerful.

1 comment. Share your thoughts »

Tagged: Computing

Reprints and Permissions | Send feedback to the editor

From the Archives

Close

Introducing MIT Technology Review Insider.

Already a Magazine subscriber?

You're automatically an Insider. It's easy to activate or upgrade your account.

Activate Your Account

Become an Insider

It's the new way to subscribe. Get even more of the tech news, research, and discoveries you crave.

Sign Up

Learn More

Find out why MIT Technology Review Insider is for you and explore your options.

Show Me