Efficient interoperation of user-generated geospatial model based on cloud computing

Abstract

Numerous geospatial models have been developed. Development of technologies that fa cilitate sharing geospatial model becomes highly desirable. Although geospatial infrastructures built using standards-based distributed services have become the default computing paradigm adopted, most of them have been built fo llowing a top-down approach that only official providers are entitled to deploy and preserve geospatial resources including geospatial models. Since the mechanisms to deploy geospatial model as a web service in geospatial infrastructures are technologically complex, there has been limited participation from users, resulting in a scarcity of sharing geospatial model resources. Furthermore, the poor efficiency of geospatial model online implement remains the pains in most of geospatial infrastructures for transmitting and computing geodata in standard OGC service. Nowadays, the development of Cloud computing including associated spatial computing becomes the promising technology which is essential for enabling computing-intensive and data-intensive geospatial research and applications. Citizen sensing activities are accumulating data, information and models at a comparable or faster pace. To utilize these emerging technologies for addressing the limitations mentioned above, we present a distributed fr amework based on OGC specification and extend it with a service deploy component. This component improves ad hoc integration, deploy ment and register of geospatial model resources by two distinguishing approaches. Geospatial models can then be deployed, shared and accessed through tools and sy stems compliant with WPS by end users. Moreover, we modified the interface specification of WPS and WFS conforming to the MapReduce parallel computing mode and designed a self-adaptive task regulating mechanism for Hadoop and achieved great acceleration of the data-intensive computing in geospatial model. Finally, we illustrate our technique by the Geospatial Data and model platform addresses availability and efficiency of geospatial model auto-deploy mechanism.