Database Server Models for WMS and WFS Geographic Web Services - Geographical Information Systems: Trends and Technologies - page 147

WFS Client

GetCapabilities

DescribeFeatureType

GetFeature

XML GML GML

WFS Server

Fig. 2. Basic operations of the WFS Specifi cation.

Selection criteria are defi ned using a fi lter. To get relevant data, a
client specifi es an object identifi er in the fi lter. The WFS server receives
the GetFeature query, determines the correct database, creates and sends a
SQL statement to the database and formats the results. The fi lter is used to
manage the “Where” clause of a SQL statement. Spatial data are handled
with the spatial schema defi ned in the ISO 19107 norm. Before sending
results to a client, the server transforms objects using a GML3 format.
The WFS specifi cation emphasizes the exchange of geographical objects.
Since data are distributed worldwide for a tremendous set of applications,
there is no reason to defi ne a unique and standard database schema.
Therefore it is of prime importance that the data model should be provided
with the data set. Unfortunately, that means that the cognitive charge of
understanding the data model is directed towards end-users. In this context,
we present some works devoted to the improvement of interoperability
and the database schema defi nition.

Data-model-oriented web services

The proposition, named mdWFS, defi ned in Donaubauer et al. (2007a,b),
defi nes a data-model-oriented web service. This proposition tries to take
advantage of both: data interoperability as defi ned in OWS specifi cations
(OGC Web Services) and the expressive power of conceptual data models.
The main idea, used from conceptual data models, is the defi nition of a
conceptual language to describe spatial data. Since the defi nition is provided
at the conceptual level, it is independent of a specifi c implementation system
or transfer format such as XML or GML. A wrapper (i.e., a compiler) is
the ideal tool to move from a generic representation to a specifi c one. The
semantic interoperability is provided by a representation of conceptual
models. The representation of a conceptual model is translated from a
schema A into a schema B. The model driven approach consists in four steps
illustrated in Fig. 3 (Donaubauer et al. 2007a): specifi cation of an application
domain, specifi cation of a conceptual schema associated with its UML meta-