A Rapid Turn-around, Scalable Big Data Processing Capability for the JPL Airborne Snow Observatory (ASO) Mission

The JPL Airborne Snow Observatory (ASO) is an integrated LIDAR and Spectrometer measuring snow depth and rate of snow melt in the Sierra Nevadas, specifically, the Tuolumne River Basin, Sierra Nevada, California above the O'Shaughnessy Dam of the Hetch Hetchy reservoir, and the Uncompahgre Basin, Colorado, amongst other sites. The ASO data was delivered to water resource managers from the California Department of Water Resources in under 24 hours from the time that the Twin Otter aircraft landed in Mammoth Lakes, CA to the time disks were plugged in to the ASO Mobile Compute System (MCS) deployed at the Sierra Nevada Aquatic Research Laboratory (SNARL) near the airport.

ASO performed weekly flights and each flight took between 500GB to 1 Terabyte of raw data, which was then processed from level 0 data products all the way to full level 4 maps of Snow Water Equivalent, albedo mosaics, and snow depth from LIDAR. These data were produced by Interactive Data analysis Language (IDL) algorithms which were then unobtrusively and automatically integrated into an Apache OODT and Apache Tika based Big Data processing system. Data movement was both electronic and physical including novel uses of LaCie 1 and 2 TeraByte (TB) data bricks and deployment in rugged terrain. The MCS was controlled remotely from the Jet Propulsion Laboratory, California Institute of Technology (JPL) in Pasadena, California on behalf of the National Aeronautics and Space Administration (NASA). Communication was aided through the use of novel Internet Relay Chat (IRC) command and control mechanisms and through the use of the Notifico open source communication tools.

This talk will describe the high powered, and light-weight Big Data processing system that we developed for ASO and its implications more broadly for airborne missions at NASA and throughout the government. The lessons learned from ASO show the potential to have a large impact in the development of Big Data processing systems in the years to come.