Our cloud infrastructure provide users with innovative methods to efficiently process and distribute data in a matter of seconds. Thanks to the EO Finder tool, users can order and process the data automatically using API or manually with GUI. Depending on users experience and needs 4 complementary processing modes available to all registered users may be used

1. Processing on the Virtual Machine: users run local processing chain on a dedicated VM instance (or a VM cluster using tool like Kubernetes or Docker) interacting with the archive via local access interfaces (S3, NFS) or OGC interfaces (WCS/WMS/WMTS).
    
2. Cluster computing (PGaaS - Product Generation as a Service): users run tasks in a serverless environment, triggering the next processing steps via an API operating on the archive using local access interfaces (S3, NFS) or OGC interfaces (WCS/WMS/WMTS). Our processing is container-based, so we are able to process up to several hundred thousand products for the customer in a very short time. Thanks to attractive price models the user saves his time and we process.
    
3. Processing in the shared code development environment (Jupyter): users run their code interactively accessing the archive using preinstalled libraries (EOlearn, GDAL) or directly.
    
4. Processing of data using external applications (i.e. GIS): users access earth observation data using OGC interfaces (WCS/WMS/WMTS) and use data processing capabilities of the SentineHub to do data processing on the fly.

Currently, a number of processors are available to automatically generate Analysis-Ready Data from Sentinel-1 (terrain-corrected backscatter, interferometric coherence) and Sentinel-2 (surface reflectance with sen2cor). Additional processors can be installed, delivered by the user in the form of a docker image. The use of the dockerized application allows for server-less data processing with own workflow rules.

Sample automated processing can be found on CREODIAS webpage: S2scenes.