OpenEO Processing Engine⚓︎

EOEPCA supports earth observation processing via the openEO standard. We recommend reading up on openEO to understand the basic motivation.

In openEO terminology, a processing engine is referred to as a ‘back-end’. Various open source and proprietary back-ends already exist. A back-end implements the openEO API on top of a specific technology stack. This means that implementations often differ in terms of the capabilities and processes of the openEO API that it implements.

Back-ends also offer a number of ‘collections’ which represent (earth observation) datasets that can be processed on that backend. Hence, data access and processing can be optimized in a specific instance.

openEO Geotrellis⚓︎

This component will be further documented in it’s own repository.

It is built on top of the Geotrellis Scala library, utilizing Apache Spark for high performance distributed computing.

This backend targets various container orchestrators such as Kubernetes and Apache YARN and can also run standalone in a simple Docker container. For HPC, we recommend to simply orchestrate multiple standalone instances using e.g. SLURM, or explore more complex options.

openEO XArray⚓︎

This backend can be found here.

It is built on top of XArray, which relies on Dask for distributed processing.

Dask can be deployed in a Kubernetes environment, or standalone.