IT key to Columbia investigation

Computerworld - When the space shuttle Columbia disintegrated over East Texas six months ago this Friday, NASA began an unprecedented effort to use IT to locate and log debris scattered over nearly 1,000 square miles.
According to Dave Whittle, chairman of the National Aeronautics and Space Administration's Mishap Investigation Team at the Johnson Space Center in Houston, by the time NASA wound down its debris-collection effort in May, searchers had collected 84,000 pieces of debris -- roughly 40% of the shuttle. Ninety-eight percent of that debris was "geo-located" -- found by means of Global Positioning System (GPS) technology. And information about the debris was stored in a Microsoft SQL Server database.
A Team Effort
The Environmental Protection Agency handled the geo-location and data logging because of its responsibility for hazardous material cleanup. Don White, the EPA's on-scene coordinator for the Columbia debris recovery at the agency's field office in Dallas, said NASA tapped the EPA and the contractor it uses for IT support on major environmental cleanup projects -- Weston Solutions Inc. in West Chester, Pa. -- to handle debris data collection. Weston had developed a field data collection program running on Hewlett-Packard Co. iPaq Pocket PCs, according to Brad Morgan, IT project manager on Weston's EPA contact.
By the time the data collection effort was finished, Morgan said, Weston and the EPA were fielding between 250 and 280 data collection teams a day, each equipped with an iPaq and a GPS receiver. Roughly 40% of the iPaqs featured an integrated GPS receiver, which made entry of geo-location data automatic.
At the end of each day, the EPA teams would synchronize their data with a SQL Server database set up by Weston. Kristin Ingram, chief of the information sciences branch at the Johnson Space Center, said the information from the EPA database was merged with a NASA database that includes a shuttle parts list. The data was then stored in the Shuttle Interagency Debris Database (SIDD). The SIDD runs on two Dell Inc. 8450 servers, each housing four Pentium III Xeon chips with 2GB of RAM and 18GB of storage. Additional storage was provided by dual Dell PowerVault systems with a capacity of 1TB each.
Data replication between the EPA and the SIDD SQL Server databases was done through bulk-merge replication rather than by transaction. Ingram said she found the process frustrating, since SQL Server proved to be "cantankerous" in merge mode.
Ingram said SIDD played a key role in refining the debris search on a daily basis by showing patterns in the distribution of key parts within the debris field. Those patterns helped narrow the search for Columbia's data recorder -- the equivalent of a commercial airliner's "black box."
NASA's earth science information directorate at the Stennis Space Center in Hancock County, Miss., helped turn the SIDD data into visual information with geographical information systems (GIS) technology, said Kirk Sharp, a GIS expert at Stennis.
Sharp said Stennis used GIS software from Environmental Systems Research Institute Inc. in Redlands, Calif., to create visual representations of the debris field.