The project has completed the initial stage of collecting, and analyzing the end-users requirements and provided the detailed work to identify the main driving lines on which the platform implementation work shall be organised. Moreover, the Business Process Modeling and Notation in the EO4EU project has been developed, along with the approach and methods used to collect the requirements.
Regarding the implementation of various components, the project has developed a Knowledge Graph (design, implementation and deployment), and documented the ingestion of EO data, leveraging AI technologies. Moreover, a first version of the data fusion engine, as well of the AI/ML models marketplace and of the EO4EU has been designed and developed. The ML methods examined have undergone extensive evaluation and fine tuning and produced two main ML models (self-supervised learning and learnt compression). The models are available and integrated in EO4EU workflows. Finally, a portal with an initial version of every customer facing service (OpenEO API, Data Visualisation Dashboard, Command Line Interface, VR/XR component, etc.) has been delivered.
On the integration front, the project focused on extracting the specifications of the services, integrating the service-oriented architecture. The integration analysis examined the integration of software components and their computational and storage needs as well as the functional needs of data in the various pre- and post-processing forms. All the development followed Infrastructure as Code (IaC) methodology, automating the software provisioning, configuration management, and application deployment. A unified storage solution in the platform by setting up a fast and reliable object storage service has also been designed. There has been extensive effort to identify the best technological solution for the process of storage, integration and deployment of the source code of the software components. A census phase of all the applications involved in the project has also been in operation. A pipelines of key components has been implemented as a starting pointtests automatically.
The project Use Cases have been defined in detail, workflow and needed resources (data, processing, results provision) are clearly identified and technical implementation guidelines formulated. All the proposed UCs have been analysed. Successively use case resources have been broken down along the project timeline, for both development and operation phases, by means of the definition of the so-called deployment plan. Finally on overall platform occupation scenario (detailed by month) has been provided joining the different use case's needs, for what regards storage (traditional high performance) and processing (traditional, high performance). The project has defined a Pilot Plan with the timeline and the detailed execution steps of each UC, steps classified into Data Access; Data Preparation; Data Processing; and Results analysis/testing/validation. The Pilot Plans were supported by gradual findings and information collection. A list of main (champions) and then secondary users was compiled and kept updated, starting from the ones expressing their interest in the proposal phase. In total 26 users (including project partners but excluding possible multiple individuals per end-user institution) have been attracted and involved. To explore their availability to contribute to the specific UC, an ice-breaking questionnaire was formulated and circulated. Results were summarized and made available to partners to complete the view in terms of user needs. Successively, more UC-specific technical questionnaires were prepared and circulated to the end users.
