Title: A Cloud-based Data Processing Foundry to Expedite SAR Science Products
Presenting Author: Hook Hua
Organization: Jet Propulsion Laboratory / Caltech
Co-Author(s):
Hook Hua, Paul Rosen, Gerald Manipon, Michael Starch, Piyush Agram, Eric Gurrola, Eric Fielding, Tom Farr, Lan Dang, Justin Linick

Abstract:
Existing and upcoming SAR missions are expected to generate order of magnitude more data volumes than other Earth observing missions. However, SAR data is often processed "by hand," using expert knowledge, and slow and labor-intensive processes. As a result, growing volumes of SAR instrument data remain unprocessed to higher-level science products. The Advanced Rapid Imaging & Analysis (ARIA) science data system has been prototyping cloud computing-based processing of SAR instrument data. The approach has shown to provide scalable compute and storage functionalities to handle the increasing needs of SAR science data processing. We are now exploring a data processing Foundry approach for SAR processing, where a cloud-based science data system can be leveraged to "on-ramp" processing and analysis of multiple SAR instrument datasets. Characterizing the "on-ramping" process is needed to model the development and operational costs of integrating additional SAR data processors, benchmarking, integration & testing, and operational data production. Once on-ramped, an on-demand cloud processing approach will enable investigators to select and quickly produce SAR science products for use in science and applications. ARIA prioritizes a disaster response application, but we are also prototyping ecosystems applications as part of the Foundry modeling.