Use of grid computing for modeling virtual geospatial products
Aijun Chen,Liping Di,Yaxing Wei,Yuqi Bai,Yang Liu
DOI: https://doi.org/10.1080/13658810902733666
2009-01-01
International Journal of Geographical Information Science
Abstract:Earth science research and applications usually use Distributed Geospatial Information Processing (DGIP) services and powerful computing capabilities to extract information and knowledge from large volumes of distributed geospatial data. Conceptually, such processing can be abstracted into a logical model that utilizes geospatial domain knowledge to produce new geospatial products. Using this idea, the geo-tree concept and the proposed geospatial Abstract Information Model (AIM) have been used to develop a Grid workflow engine complying with geospatial standards and the Business Process Execution Language. Upon a user's request, the engine generates virtual geospatial data/information/knowledge products from existing DGIP data and services. This article details how to (1) define and describe the AIM in XML format, (2) describe the process logically with an AIM, including the geospatial semantic logic, (3) conceptually describe the process of producing a particular geospatial product step by step from raw geospatial data, (4) instantiate AIM as a concrete Grid-service workflow by selecting the optimal service instances and data sets, and (5) design a Grid workflow engine to execute the concrete workflows to produce geospatial products. To verify the advantages and applicability of this Grid-enabled virtual geospatial product system, its performance is evaluated, and a sample application is provided.