Processing of Geospatial Data for Science and Industry

The amount and quality of data generated by Spatial Data Infrastructures (SDI) is on a sharp rise. Efficiently processing this data is one of the foremost goals of geoinformatics. By employing the powerful D-Grid infrastructure, processing data and models from different areas of SDI is facilitated and expedited.

The strategic goal of the GDI-Grid project (GDI is shorthand for "Geodateninfrastruktur", the german phrase for spatial data infrastructure) is integration of technologies from SDIs and Grid computing. Chained Workflows executed in the Grid are slated to replace mostly monolithic processing methods in conventional SDIs.

From a technical point of view, GDI-Grid combines technologies currently available to SDIs and Grid computing and gridifies computation of geospatial data. To achieve this goal, data, models, services and security mechanisms are combined into Grid-compatible workflows.

In modern distributed processing of geospatial data, interoperability is achieved by implementing interfaces according to the OGC web services specifications. To adapt these specifications – which stem from geoinformatics – to Grid standards, architectural differences and commonalities had to be determined.

With standards-compliant and sustainable development in mind, the project then designed and implemented a platform that adheres to security requirements in D-Grid. OGC web services were transformed into WSRF web services and thus received a direct interface to the D-Grid computing infrastructure. These newly-developed web services as well as already existing services were combined to workflows – processing chains orchestrated by a specialized component, the workflow engine. After checking for syntactical and semantical validity, the workflow engine monitors job execution and data flow. All elements of a workflow are – very much in the same way as the workflow engine itself – web services and can be accessed as such.

An additional challenge for GDI-Grid is integration of commercial software products – brought into the project by industrial partners like ESRI Germany or Stapelfeldt Engineering. It is a project goal of GDI-Grid to make commercially-licensed software available in D-Grid in a scalable fashion. In cooperation with the D-Grid integration project, we are working on fundamental processes and solutions.

Use cases for flood simulation, noise propagation simulation and emergency traffic routing help evaluating the Grid-based SDI created in the project. A practicability and marketability study serves as an indicator for the commercial value of the project results and helps ensure sustainability.

Example Scenario “Flood Simulation”
Man-made alteration of water bodies and the current climate change are some of the triggers for the increased frequency of flood catastrophes. Laws and regulations in EU and national governments mandate large-scale simulations of flood events for endangered areas. Additional requirements for flood countermeasures include risk analysis and evacuation plans.

By using the computing resources available in D-Grid, detailed simulation of large target areas as well as data conversion and enrichment can be facilitated. The resulting simulation services comply with current standards in geospatial data processing and can be incorporated in a Grid-based SDI.

Example Scenario “Noise Propagation Simulation”
Similar to flood simulation, simulation of noise propagation in urban areas is a processing-intensive task performed on very large data sets. Also, EU directives for assessment of environmental noise require simulation and mapping of noise in urban environments. The results of such noise assessment have major economic implications since they can significantly influence property value.

Thus, detailed simulation of noise propagation is in high demand. In the GDI-Grid project, complex noise propagation simulations are evaluated regarding their parallelization potential. Specialized simulation software – developed by project partner Stapelfeldt Engineering – is adapted to the Grid. Testing of the Grid-based simulation is performed with test data sets of variable size, the largest being roughly the size of a large German city.

Example Scenario “Evacuation procedures”
In the third “proof of concept” scenario, simulation and optimization of evacuation procedures will be performed using the D-Grid infrastructure. While conventional routing algorithms have already been highly optimized for use in commodity routing applications (such as car navigation), evacuation procedures are significantly more complex. The fact that during a catastrophe, the road network and evacuation situation constantly change has to be taken into account –disaster prediction is an integral parts of the scenario.

These complex and interconnected calculations can be greatly accelerated using the Grid, improving the quality of disaster simulation and therefore saving lives during an actual catastrophe.