Novel IT

The goal of this task is to establish, upgrade and develop the required infrastructure for EvoLand. As a basic infrastructure, we will rely on public clouds such as WEkEO, CDSE, and similar, supported when required with hyperscalers such as AWS and GCP. The operational services Terrascope, Euro Data Cube and Sentinel Hub (currently running on several clouds) will serve as a middle layer and algorithms will access these via the open source openEO interface. The openEO API allows switching easily between different providers or backends. It allows a uniform approach in data discovery, data cube processing, batch job management, and easy integration of the processing workflow to streamline the move from experiment to operational production. This will provide developers with access to various full and global archives of Sentinel-1 and -2, processed to the ARD level.

However, significant upgrades are needed to provide the capabilities required for EvoLand such as i) automated orchestration of openEO processes to relevant back-ends, ii) scaling capabilities of data processing chains to allow processing of ML features at a global scale, in the appropriate format (i.e. xarray, COG, zarr, etc.) and iii) integration of the novel EO data in the infrastructure. The most important part, however, is the efficient integration of all the components, including the existing back-ends of project partners, to ensure the re-use of the already developed pipelines and not start from scratch. The result will therefore go well beyond the scope of the project – both due to the upgrades of the used cloud-native data processing services, and open-source tools as well as the development of new methodologies for sustainable and reusable algorithm development.