Data Rods: High speed, time-series analysis of massive cryospheric data sets using pure object databases

Abstract

A data rod is a conceptual data object, defined as the time-series of data at a specific set of gridded coordinates. Ideally, a single data rod encapsulates all the information known at that location through time. The Data Rods project sought to store collections of data rods in standalone or distributed databases, providing users with direct access to large collections of remotely-sensed data (as opposed to only allowing direct access to the metadata). Temporal data analyses could then be performed by querying for the set of data rods that cover the spatial area of interest. This query process would implicitly retrieve only the data necessary for the desired analysis, avoiding the overhead of downloading and filtering extraneous data. Combined with server-side processing and userfriendly analytical tools, the project's overarching objective was to reduce the research cycle time from weeks or months down to days, perhaps less.

Multi-Source Data Processing Middleware for Land Monitoring within a Web-Based Spatial Data Infrastructure for Siberia

“…A more complex task is the provision of styling information for the visualization of data, as this information is not stored in the original data file and has to be generated manually for each dataset. Technical limitations did not occur during implementation, but further work would be helpful in investigating possible improvements in data storage: for example, raster databases [49,50] or no-SQL databases for storing and analyzing time-series data [51]. Data access could also be improved in ways, such as the implementation of the OGC Web Coverage Processing Service specification [52], which provides processing during data access [53].…”

Global Satellite Observations for Smart Cities

“…The concept of data rods (Teng et al 2012;Gallaher and Grant 2012;Rui et al 2012Rui et al , 2013) can be applied to this challenge. Teng et al (2012) proposed two general solutions: 1) retrieve multiple time series for short time periods and stitch the multiple time series into desired single long time series and 2) reprocess (parameter and spatial subsetting) and archive data as one-time cost approach.…”