Prototype of HPC/Data Infrastructure for On-demand Services
The project will build a prototype for Data/HPC services based on Earth sciences cases. The consortium will develop and provide new services to discover, manage and process spatial and environmental data (in particular from earth surface, atmosphere, ocean). The project will develop datasets, adding into the EOSC catalogues. It will optimize and industrialize treatments workflows to allow the largest reusability, and provide an open access to standardized HPC services. The consortium will improve the FAIRization processes and open access to the environmental initial and processed datasets. The project will also progress on sustainability and evolution of services from an experimental and restricted usage to a large and standardized one (in a production phase). It will explore a distributed model for data transfer and resource allocation between two european computing centers, CINES in France and CSC in Finland. On the use cases, the project will develop data post-processing methods, coupled with HPC capabilities, which will be deployed as a service for several end-users such as scientific communities, public authorities, private players and citizen scientists. The data generated and services created will be available on the relevant EU portals (EU Open Data Portal/EUDAT/EOSC) and will be preserved using the long-term preservation services of EOSC. The PHIDIAS project will then propose a generic workflow for massive scientific data by combining computing, dissemination and archiving resources in a single framework.
CSC collaborates in this project with 12 PMs on AI and on the fly computing workflow.
Other partners include French computing centers and environmental research organisations, including SYKE.
The project will allow CSC to test and develop new approaches with regards to AI and will be used as a pilot for EuroHPC.