Long term preservation of scientific data: Lessons from jet and other domains

All fusion devices generate large amounts of data, which must remain accessible throughout and beyond the lifetime of the device. This need for long-term data preservation creates major software and hardware challenges. Issues surrounding long-term data preservation will be discussed, and the main challenges in data curation will be outlined. The paper will then review the approaches taken at JET to ensure accessibility of all of JET's data over almost three decades. Using the JET experience and the wider context, the paper will then suggest how new experiments should plan for the long-term preservation of their data.

► Data from a fusion experiment must be preserved for decades, and will be required beyond the lifetime of the experiment. ► General principles of long-term data preservation have been identified in domains outside fusion. ► Key challenges are media obsolescence, data organisation, and file format obsolescence. ► Data from JET has to date been preserved for almost three decades: the steps taken to achieve this are discussed. ► Recommendations for new devices in planning for the long-term preservation of their data are made.