Title: The Automated Event Service: Efficient and Flexible Searching for Earth Science Phenomena
Presenting Author: Thomas Clune
Organization: NASA GSFC

Abstract:
Existing NASA data providers generally only support searches based upon metadata (e.g. location, time, physical quantity, etc.) rather that data (i.e. physical values). This constraint drives researchers to expend significant portions of their time and effort downloading and managing large subsets of data for local examination. This inefficient process is then wastefully replicated by other investigators pursuing analogous queries. The Automated Event Service (AES) is a data analysis environment that exploits Big Data technologies to facilitate custom searches for user-defined phenomena in Earth Science reanalyses data sets. AES is a scalable system that combines (1) an extensible next-generation database with direct support for multidimensional data and built upon a shared-nothing computing architecture, (2) a high-level Event Specification Language (ESL) and (3) a collaborative environment known as the Collaborative Workbench (CWB), which was funded by a separate AIST grant. Because AES co-locates data and computational tasks/queries, searches typically proceed several orders of magnitude more quickly than the traditional approach. Queries and results can be readily shared with collaborators for refining search criteria and/or further analysis. After a brief summary of the capabilities of AES, we will present our experience in using AES for two science scenarios: Blizzard and Somali Jet. We conclude with a discussion of our plans for further extensions.