|START Conference Manager|
Virtually no INSPIRE dataset is kept and managed managed in the INSPIRE Data Store itself. Usually the data is kept in a different system and has to be transferred to the INSPIRE Data Store. Here the well-known ETL Process comes into play. The INSPIRE Data Store in turn is in most cases a spatial relational database and as the base data model usually differs from the INSPIRE data model the ETL Process usually includes a model transformation as well. This transformation is not the last one. In order to deliver a harmonized data with a WFS, the relational data has be transformed again, now into a GML data stream. Basically this is another ETL process, which unfortunately is really tricky and exceeds the functional limits of some WFS, especially when it comes to complex GML data models. Although deegree is able to handle even complex data models, the data transformation consumes considerable CPU time and database queries. Already today deegree contains a functional approach to solve that problem, the so called “BLOB Mapping Mode”. This mode however has some limitations which now have been overcome with introduction of the “High Performance Mode”. The High Performance Mode generalizes the approach of BLOB mode and largely solves its limitations. Handling and Performance of an WFS is significantly improved. The lecture examines the above-mentioned problems with ETL processes in general and the challenges of configuring an INSPIRE conformant WFS. The conceptual extension of the High Performance Mode is explained as well as the performance gains achieved with the first prototype.
Topic Area: [2.5] Technical solutions for optimizing and/or extending the INSPIRE technical framework Abstract Type: Oral Presentation
Comments: WFS, INSPIRE, complex GML data models, performance, ETL, database, DBMS
START Conference Manager (V2.61.0 - Rev. 4840)