Synchrotron X-ray experiments generate massive and diverse datasets across multiple scientific disciplines, necessitating robust and sustainable data processing solutions. As the scientific community increasingly relies on open-source software, the challenge of maintaining and preserving the sovereignty of data processing tools becomes critical. This talk will address the unique needs of synchrotron facilities and present a strategy adopted by the data reduction and analysis group at Synchrotron SOLEIL. We will discuss the importance of maintaining full control over software and libraries, avoiding dependencies on external services that may change their economic models.
Our approach emphasizes the use of stable packaging systems, such as Debian-based distributions, over e.g. rolling release solutions such as conda. We have contributed to over 400 Debian Science packages, including JupyterLab and Napari, to ensure long-term stability and accessibility. Additionally, we have developed specialized software tailored to synchrotron-related data processing and provide data analysis services using open-source tools, including on-demand virtual machines in the browser.
This talk will also outline criteria for defining long-term support for software and services, aiming to foster collaboration within the scientific community to create a sustainable and independent ecosystem for scientific software development and distribution.