SFB 1404/2: Transparente Multi-Center Datenanalyseworkflows für die Erdbeobachtung (TP B05)

DAWs for satellite-based Earth observation (EOW) are characterized by large data volumes and high heterogeneity and complexity of the analysis steps. Appropriate scheduling of such work-flows must be adaptive and consider the location of data files, which exceeds the capabilities of common scheduling engines. We overcame this limitation in FONDA I by designing a location-aware scheduling algorithm, which makes porting and adapting EOWs much easier. However, many current EOWs require huge input datasets making downloads infeasible, while datasets are often distributed across different online platforms. We will accordingly research the foundations of a federated, multi-center enabled EOW engine and exploit this federated design to answer crucial questions on long-term changes of agricultural land due to its importance for food security and sustainability.