2003 Seattle Annual Meeting (November 2–5, 2003)

Paper No. 12
Presentation Time: 11:15 AM

WEB SERVICES AND GEOLOGIC DATA INTERCHANGE


COX, Simon J.D., CSIRO Exploration and Mining, ARRC, PO Box 1130, Bentley, WA, 6102, Simon.Cox@csiro.au

Web-hosted Spatial Data Infrastructures are being developed in many jurisdictions. With the emergence of standardised data encodings, in particular using XML, interest has shifted from catalogues to web interfaces directly to data. The Open GIS Consortium (OGC) is designing interfaces in the form of standardised syntax for operations composed as request-response message-pairs carried over http. Published specifications include Web Map Service (WMS) which concerns maps transported as digital images, Web Feature Service (WFS) which provides access to geographic objects and their properties, and Geography Markup Language (GML) which provides XML components and a basic data model to encode descriptions of geographic objects. Under the Sensor Web Enablement initiative the same pattern is used for interfaces to dynamic sources, such as sensors and simulators.

The OGC Web Service interface (OWS) specifications define services in terms of their public interface, and do not prescribe the internal structure of either server or client software. Implementations, therefore, are usually in the form of facades on top of existing data services such as GIS and RDBMS on the server side, and as additional import options in desktop map-viewers on the client side.

Because of the focus on the public interface between software, interoperability within a specific application domain, such as geology, is enabled by standardisation of the encoding of the response payload. Optimum behaviour requires agreement not only at the level of the low-level syntax (XML, GML Feature-property model) but also in the detailed information models that are being serialised (e.g. a common model and encoding for boreholes, or for mineral deposits). The community language is defined as a GML application schema. The intention is that many servers with differing private data models (e.g. RDBMS schemata) and different persistence mechanisms (e.g. RDBMS vs. OO-DB vs. XML-DB) can all serve geologic information according to the same data model serialised in a common way. In this context different sources can be used according to the usefulness of the data, rather than the technology used by the host organisation.