Curating the web’s deep past – Migration strategies for the German Continental Deep Drilling Program web content

On timescales beyond the life of a research project, a core task in the curation of digital research data is the migration of data and metadata to new storage media, new hardware, and software systems. These migrations are necessitated by ageing software systems, ageing hardware systems, and the rise of new technologies in data management. Using the example of the German Continental Deep Drilling Program (KTB) we outline steps taken to keep the acquired data accessible to researchers and trace the history of data management in KTB from a project platform in the early 1990ies through three migrations up to the current data management platform. The migration steps taken not only preserved the data, but also made data from KTB accessible via internet and citable through Digital Object Identifier (DOI). We also describe measures taken to manage hardware and software obsolescence and minimise the amount of maintenance necessary to keep data accessible beyond the active project phase. At present, data from KTB are stored in an Open Archival Information System (OAIS) compliant repository based on the eSciDoc repository framework. Information packages consist of self-contained packages of binary data files and discovery metadata in Extensible Mark-up Language (XML) format. The binary data files were created from a relational database used for data management in the previous version of the system, and from websites generated from a content management system. Metadata are provided in DataCite, GCMD-DIF, and ISO19139/INSPIRE schema definitions. Access to the KTB data is provided through download pages which are produced by XML transformation from the stored metadata.