IN21D-04:
Use Cases for Server Operators Extending the Open-Source Data-Access Protocol (DAP)
Tuesday, 16 December 2014: 8:45 AM
James H R Gallagher, OPeNDAP, Inc., Butte, MT, United States, David W Fulker, OPeNDAP, Inc., Boulder, CO, United States, Brian Blanton, Renaissance Computing Institute, Chapel Hill, NC, United States, Steven Businger, University of Hawaii at Manoa, Honolulu, HI, United States and Peter Cornillon, University of Rhode Island Narragansett Bay, Narragansett, RI, United States
Abstract:
On the premise that EarthCube must incorporate data-access (Web) services that are effective even in big-data contexts, we articulate three use cases where a common form of data reduction, namely array-subset selection, falls short. These cases—addressing climate-model downscaling for native-Hawaiian use, real-time storm-surge prediction for U.S. coastal areas, and analysis of sea-surface-temperature (SST) fronts using satellite imagery—share three traits: a) each requires access to vast and remote volumes of source data, though the end-user applications need much less (by orders of magnitude); b) the volume reduction cannot be realized solely via subsetting, especially if limited to subarray-specification via index constraints; c) each data-reduction need can be met by extending a well-used data-access protocol (DAP) to embrace new data-proximate (I.e., pre-retrieval) server functions; and d) the required new functions will be useful across many geoscience (EarthCube) domains. Reflecting OpenDAP progress on designing this extension—dubbed ODSIP for Open Data-Services Protocol, to be prototyped under an NSF/EarthCube award—this talk sketches the near-source operations needed for the three use-cases, highlighting potential for abstraction and thus broad applicability.
