PUB - Publikationen an der Universität Bielefeld

The architecture of Kiel Data Management Infrastructure (KDMI) is setup to serve from the data creation process all the way to the data publication procedure. Accordingly the KDMI is managing data at the right beginning of the data life cycle and does not leave data unattended at this very crucial time. Starting from the chosen working procedure to handwritten protocols or lab notes the provenance of the resulting research data is captured within the KDMI. The provenance definition system is the fundamental (see figure 1) capturing tool for working procedures. The provenance definition is used to enable data input by file import, web client or hand writing recognition. The captured data in the provenance system for data is taking care of unpublished in house research data created directly on site. This system serves as a master for research data systems with more degrees of freedom in regard to technology, design or performance (e.g. GraphDB, etc). Such research systems can be regarded as compilations of unpublished data and public domain data e.g. from World Data Centers or archives. These compilations can be used to run statistical data mining and pattern finding algorithms on these specially designed platforms. The architecture of the KDMI ensures that a technical solution for data correction from the slave systems to the master system is possible and improves the quality of the stored data in the provenance system for data. After the research phase is over and the interpretation is finished the provenance system is used by a workflow based publication system called PubFlow. Within PubFlow it is possible to create repeatable workflows to publish data into various external long-term archives or World Data Center. The KDMI is based on the utilization of persistent identifiers for samples and person identities to support this automatized publication process. The publication process is the final step of the KDMI and the management responsibility of the long-term part of the data life cycle is handed over to the chosen archive. Nevertheless the provenance information remains at the KDMI and the definition maybe serves for future datasets again. Unattended data may get lost or be destroyed