EOSC-Life Final EOSC-LS Cloud Observatory Report

This report details the evolution of the support mechanisms available to the Biological and Medical ESFRI research infrastructures (BMS RIs) and how the allocated cloud resources have been used at this stage of the EOSC-Life project by the BMS Host Institutions (BMS HIs). We analyse the activities carried on to date within the Work Package 7 (WP7) tasks:

• T.7.1 Cloud Resource Planning and Integration
• T.7.2 Cloud Training Programme
• T.7.3 Sensitive Data Hosting and Processing Policy Framework
• T.7.4 Managed Platforms for Data Analysis
• T.7.5 Operations and Support Helpdesk

and provide our point of view on how the WP7 work will evolve during the remaining months of the project. In general this deliverable does not repeat reporting of activities already reported in the initial version of the Cloud Observatory Report (D7.11), except where is needed to provide context.

Our observations within WP7 from the first 3 years of the project are:

• There is a considerable need for upskilling. Cloud technologies are still complex and require substantial understanding of IT infrastructure basics to set up and operate.
• A recent shift to container technologies unifies and simplifies distribution of applications and services, but requires additional skills from operators and scientists. A similar effect can be seen in the shift between traditional file and directory based storage and object storage for modern, cloud-native applications and services.
• An alternate approach is the use of services and tools that abstract away the need to understand how the underlying infrastructure works.
• This bears a concrete risk that Life Scientists and their Institutions lose the ability to develop, run, and operate their own infrastructures because there will be a lack of organic specialisation from within Life Sciences domain to facilitate future development of infrastructure and workflow management development.
• All approaches still face challenges in the appropriate storage and handling of sensitive data. While service-based workflow offerings allow ease of access, they make reasoning about (transient) data storage more difficult.