Using spatiotemporal patterns to optimize Earth Observation Big Data access: Novel approaches of indexing, service modeling and cloud computing

Abstract

Based on our GEOSS Clearinghouse operating experience, we summarized the three Earth Observation (EO) Big Data access challenges, namely, fast access, accurate service estimation and global access, and two essential research questions: are there any spatiotemporal patterns when end users access EO data, and how can these spatiotemporal patterns be utilized to better facilitate EO Big Data access? To tackle these two research questions, we conducted a two-year pattern analysis with 2+ million user access records. The spatial pattern, temporal pattern and spatiotemporal pattern of user-data interactions were explored. For the second research question, we developed three spatiotemporal optimization strategies to respond to the three access challenges: a) spatiotemporal indexing to accelerate data access, b) spatiotemporal service modeling to improve data access accuracy and c) spatiotemporal cloud computing to enhance global access. This research is a pioneering framework for spatiotemporal optimization of EO Big Data access and valuable for other multidisciplinary geographic data and information research.