Key points are not available for this paper at this time.
Today various data formats and different browser based 3D globes and other viewer implementations are available. To enable interoperable visualizations between them and to integrate the distributed data, a standardized interface on the server side is necessary for the portrayal of 3D geospatial data. We therefore present a framework implementing the new 3D Portrayal Service standard (3DPS) of the Open Geospatial Consortium (OGC). The focus of the paper is on the View conformance class of the 3DPS that supports the delivery of server-side rendered high resolution images. An introduction to the main aspects and features of the 3DPS standard is given. The framework controls the workflow from the storage of geographical raw data to the creation of the visual end result. The framework is open for extensions giving the ability to utilize different technologies. On the other hand the framework encapsulates server technology and starts server instances, which scale to handle a massive amount of requests. The framework implementation also inherits the 3DPS specific implementation overhead, allowing the framework user to focus on implementing his technology. Building upon the framework, two prototypical applications are presented, which use different renderers, data formats and storage solutions. Based on these two prototype service implementations, a distributed data set is integrated in a web-based visualization. It shows that geospatial data can be served by different independent data providers without a data integration on the data model level. The results are a significant step forward to a 3D distributed geospatial data infrastructure based on open standard service interfaces. In addition, the use of different data formats produced interesting insights in their performance for server-side rendering. This work will be beneficial for researchers wanting to support 3DPS GetView without having to take care of the standard's overhead and the implicitly needed server implementation.
Gutbell et al. (Mon,) studied this question.