EOhopS: EO data hosted processing for science

When the European Space Agency decided to introduce the European science community to cloud-based services for earth observation data, it approached cloudeo to build the necessary platform. The idea was to include so-called Third-Party Mission (TPM) – i.e., not provided and hosted by ESA¬ – data in this cloud-based approach: currently, users have to physically download such TPM data from ESA archives or the data providers' servers. This process uses extensive bandwidth resources and requires large storage and data management capacities on the users' side.

Cloudeo with many years of experience and expertise in cloud-based Data-as-a-Service (DaaS), Software-as-a-Service (SaaS), and Platform-as-a-Service (PaaS) projects and products. They developed their Service API to create a single interface, allowing platform-independent access to large files (typical for Earth Observation data). In essence, the cloudeo Service API makes homogeneous access to the data and simplifies access for the crowdsourcing application. The idea is that, rather than acquiring expensive software licenses and purchasing large, bandwidth- and storage-heavy EO imagery data packages, customers can pay for the usage and processing of such files using virtual workstations that include time-based software licenses. This model was very well suited for the EOhopS project; mainly, because the embedded metering supports pay-per-use and application-specific business models, allowing Third-Party providers to monetize their data without concerns about their IP rights.

The project's scope was to provide scientific users with access to these third-party EO data in a hosted processing infrastructure (similar to the cloudeo Workbench) using Virtual Machines. Simultaneously, the users were also invited to upload their content, like in-situ measurement data or scientific algorithms, to this virtual Workbench. This concept allows users to process these vast volumes of data in the cloud and to download only the desired results to their local computing environment, thus saving costly bandwidth and storage resources while at the same time protecting the copyright of the TPM providers.

This model came with many logistical challenges that needed to be resolved, such as license conditions of EO Data-as-a-Service, allocation of remote processing capabilities to users, and secure connections of this cloud processing infrastructure TPM data providers for ordering and archive access, etc.

Since this was an exploratory program, projects were selected not only on their scientific merit but also based on how well they were able to use the new technological affordances, namely the capacity to process large amounts of data remotely, to upload their content to a remote environment for analysis, and to make only the processing results available for download.

Only a limited number of projects could be served within ESA resources allocated to the service; projects were accepted, based on their matching with the above objectives and a first-come-first-served basis. The demand exceeded the project capacity: As of now, all EOhopS resources have been allocated, and no further project applications are currently accepted.