Big guns for hire: Supercomputing-in-the-cloud
Some companies are slashing costs and improving production by using on-demand high-performance computing services
By Todd R. Weiss
July 13, 2010 06:00 AM ET
Computerworld - For much of its existence, Woodward Governor Co. has been building fuel nozzles and other parts for jet aircraft engines the old-fashioned way -- carefully designing them, building physical prototypes out of raw materials, conducting tests and finally creating production parts.
Typically, it took the Fort Collins, Colo., company weeks to produce workable plans for specialized new nozzle designs. Now, that time frame has been trimmed down to several hours by cutting out the steps needed to make physical prototypes.
Instead, Woodward is using powerful cloud-based supercomputers that can run complex software to quickly create and evaluate designs that can go right into production. In other words, the nozzle designs can go from creation to production 80% faster because the company no longer needs to first build models to prove the designs, a change that's projected to save the company about a half-million dollars each year.
Your on-demand HPC shopping list
Perhaps your company needs high-performance computing (HPC) power to speed up its design or production processes, but you're not able to make the financial commitment for the expensive hardware, additional power and cooling, and dedicated IT administrators needed to run it all. If so, here are some options in the growing pay-for-what-you-need supercomputing market.
- SGI: In February, Silicon Graphics International Corp. launched an on-demand, cloud-based HPC service called Cyclone. It's designed for companies that need extra computing power to run technical applications. "In the first week of the announcement, we had 15 calls from companies on three continents" with inquiries about Cyclone, says Christian Tanasescu, SGI's vice president of software engineering. "Certainly this is our plan -- to help companies bring in supercomputing capabilities, of lowering the barriers for this kind of computing." Cyclone marks the first time that SGI, a server manufacturer, is offering IT services, Tanasescu says. The Fremont, Calif.-based vendor has signed on some customers in the academic market so far, but no further details are being released, according to an SGI spokesman.
- HP: Hewlett-Packard Co. is testing a public HPC cloud with Intel Corp., Yahoo Inc. and several academic organizations, according to Ed Turkel, HP's manager of business development for scalable computing and infrastructure. The project, which includes the company's HP Labs division, consists of open-source research with connected supercomputers from all the participants, he says. The project, called OpenCirrus.org, is available now for companies that want to access it and experiment with it. The amount of computing power from each of the participating groups varies, but each contributes access to more than 1,000 processor cores. The total computing power of about 16,017 CPU cores is presently available through the cloud test bed. The project is still under development and no details were available on when or if it would ever be available for use by paying customers.
- IBM: Herb Schultz, IBM's deep computing manager, says that the company had predicted that a market for on-demand HPC services would materialize and started preparing the necessary systems about five years ago. Customers who want to buy on-demand supercomputing services from IBM can register online and then access the services through a secure, Web-based system. Users can upload data or run their executables as needed. IBM can isolate a given customer's data and put a wall around it so there are no security issues, Schultz explains.
Security warnings
One caveat about supercomputing in the cloud is that user companies need to be absolutely sure that their proprietary data will be stored, run and analyzed securely by HPC vendors, says Dan Olds, an analyst at Gabriel Consulting Group. To accomplish that, customers need to be upfront about their data protection needs and ask their vendors questions to ensure that their data is handled with the security that they require, he says.
"They want to opt for more well-known vendors if they have concerns for how sensitive their data is," says Olds. "And think about something like jet-nozzle design," he says, referring to Woodward's HPC projects. "There are a fair number of people out there that would like to get a sneak peak at that early information, such as the people who are bidding against them. So it's important that you get assurances about security."
"Woodward had some fairly robust computer workstations internally" to produce designs, but the company didn't have anything powerful enough to run large simulations, says Robert Graybill, an HPC consultant who worked with Woodward to help add on-demand supercomputer power to the company's production capabilities. Woodward's pilot project lasted 10 weeks, and additional testing will continue through October, followed by reviews of the project to determine if and how Woodward might use on-demand supercomputing in the future.
Graybill is the CEO and president of McLean, Va.-based Nimbis Services Inc., an intermediary that helps match businesses that need more computing horsepower with HPC providers that can deliver it on an as-needed, pay-as-you-go basis. (Woodward IT representatives were not available for comment.)
Dan Olds, an analyst at Gabriel Consulting Group Inc. in Beaverton, Ore., says on-demand supercomputing is an idea that's blossoming. It's particularly useful for businesses that sometimes need processing help but can't afford to buy their own full-time, on-site HPC equipment.
"And it really opens the door for other users where this isn't even on their horizon yet," Olds says. "They could do these things on their own with the modest equipment they have, but it literally could take years." Alternately, users could simplify the problems they're trying to solve so their projects could run on systems they can afford, "but that might not give them the results they need," Olds says. On-demand supercomputing helps companies avoid both those compromises without breaking their budgets.
Free Insider Guide