What services does RADAR offer?
With RADAR, research data can be archived or published. Archived datasets are stored securely for user-defined periods. For example, the German Research Foundation (DFG) recommends keeping research data for a period of 10 years. In RADAR, data providers can choose retention periods between 5 and 15 years. By default, archived datasets and the descriptions (metadata) are not publicly accessible unless the data provider specifies otherwise. Published datasets are made publicly accessible. Optionally, data providers can define an embargo period. The data can then only be accessed publicly after this period has expired. RADAR assigns a Digital Object Identifier (DOI) to each published dataset. This makes it possible to uniquely identify a dataset, connect it to publications and reliably cite it. RADAR guarantees secure storage of at least 25 years for published datasets. In order to improve the findability and interpretation of the data, datasets can be described with discipline-specific metadata. With RADAR, data can be reviewed before publication, for example as part of a peer review process. For this purpose, data providers in RADAR can create non-public links and pass them on to reviewers. In addition, RADAR provides an API that allows RADAR customers to offer the service via their own user interface.
How much does the service cost?
In addition to an annual contract fee, costs comprise usage-based fees, which depend on data volumes and the service used. The costs for data archiving depend on the amount of data stored and the retention period. For the publication of datasets, RADAR offers a one-off payment based on the data volume. The current price list can be found here. These prices apply exclusively to universities and research institutions. When applying for funding, applicants can add the costs for the publication or archiving of research data with RADAR to the application and thus procure funding for making use of RADAR services.
How can others access the datasets?
Published datasets can be viewed publicly and access is free of charge. The license terms specified by the data provider apply to the reuse of the dataset.
Archived datasets are generally not publicly accessible (dark archive). However, the data provider can grant individual registered users access to a dataset.
How long are published and archived data stored in RADAR?
For archived datasets, the data provider determines the retention period (5, 10, or 15 years). The dataset is then securely archived by RADAR for this period (dark archive). The data provider can access the dataset at any time.
For published datasets, RADAR guarantees the availability of the data for a period of at least 25 years. If necessary, the data provider can define an embargo period. In that case, the dataset will only be published after this period has expired.
What measures for the long-term preservation of datasets does RADAR take?
The data transfer to the storage layer is carried out via standardized protocols (SFTP, GridFTP). Data is stored using modern disk and tape systems. These offer several petabytes of storage space. A storage management system (high-performance storage system - HPSS) ensures efficient data access and allows cost-effective long-term storage. System reliability is ensured by multiple copies stored at different locations using different technologies. The use of Redundant Array of Independent Disks (RAID) and the system monitoring itself ensure the data security of the administrative components, such as databases, in the system. The data itself is available as BagIt structures in ZIP containers, i.e. the data is stored in well-documented containers. Each archived dataset consists of a file that is self-descriptive in itself (containing the data itself and the descriptive metadata record). For the interpretation of the BagIt structures no special software is necessary.
RADAR currently only ensures the physical preservation of datasets (Bitstream Preservation), a functional long-term preservation with format migration is currently not planned.
RADAR does not guarantee the permanent usability or interpretability of the data contained in a dataset, as these depend on the data formats selected by the data provider. RADAR therefore recommends the use of file formats that are suitable for long-term preservation.
Is the data stored redundantly?
Yes, three copies are stored at three locations (2 x Karlsruhe, 1 x Dresden). Different software and hardware systems are in use. The systems in Karlsruhe and Dresden are organizationally and administratively completely separated from each other.
Is the integrity of the data checked during the retention period?
Before being transferred to permanent storage, datasets are provided with a checksum, which is checked after each duplication. By comparing checksums, errors can be detected and eliminated during data transmission. When accessing a dataset, the checksum is calculated again and compared with the stored value in order to identify possible data consistency errors. If an error should be detected, RADAR accesses the second copy at the other location. The storage media used are replaced at the latest when the manufacturer's warranty expires. The affected datasets are migrated to new storage media and checked to avoid bit errors ("fixity checks").