IT reliability at stake during NASA launch

Information technology vendors will be under the microscope tomorrow during the first NASA space shuttle launch since 2003’s Columbia disaster.

Silicon Graphics Inc. (SGI) and Nortel PEC Solutions equipment are vital components of the Discovery shuttle’s Return to Flight mission, in which seven crewmembers will travel to the International Space Station and test new safety procedures.

In preparation for the flight, NASA officials have been modeling and remodeling what-if scenarios of potential dangers on SGI servers at several NASA research centers. The Columbia mission did not have support systems that had the same computational ability of the computers that will be used in the Discovery launch, officials said.

“If they were to notice a problem, like a tile flying off, they would have sufficient computational ability to do repairs,” said Dave Parry, SGI senior vice president and general manager of the Server and Platform Group.

Ames Research Center’s SGI-built Columbia supercomputer, named in honor of the lost crew, has hit processing speeds of 51.87 teraflops, making it the second fastest supercomputer in the world.

“NASA did some mock testing demonstrating that they could propose solutions in a 24-hour window that would allow them to react in the case of something going wrong,” Parry said.

And should the Columbia computer malfunction, NASA has a backup system installed at Goddard Space Flight Center in Greenbelt, Md., he added.

The Columbia shuttle disintegrated upon re-entry to earth because a piece of foam broke off a protruding part of the external tank and struck the shuttle's left wing at more than 500 miles per hour during liftoff.

This time, to weigh the seriousness of any damage, a team of NASA engineers will compare observations of the problem area with historical engineering data and, if necessary, run complex computational fluid dynamics calculations on the Columbia supercomputer to predict changes in heating due to the damage.

Walter Brooks, chief of the NASA Advanced Supercomputing Division at Ames Research Center, said his team has increased computational capacity by a factor of 60 in the two and a half years since the Columbia shuttle’s destruction.

Meanwhile, Nortel PEC routers will support mission-critical communications throughout the flight, allowing NASA research centers to share data from the shuttle once that information reaches the ground. The network is not onboard the craft.

Scott Gibbs, general manager of the civilian segment at Nortel PEC, said the company’s network is 99.99999 percent reliable.

“Failure is not an option,” he said.