RDA DOI Use Cases
The RDA has created a set of use cases to facilitate structured DOI management for the archive. Considerations include data security/resiliance, dataset version control, dynamic dataset extension, and dataset depreciation.
Create a new DOI for a dataset
- Requirement: All files must have one-to-one match on tape (offline) and online storage
- Mint a new DOI through DataCite
Complete dataset replacement (new data from the provider)
- Requirement: Update dataset metadata to describe changes on the existing dataset landing page.
- Assign a new DOI for the updated dataset
- Legacy version:
- Maintain in tape archive and remove from online storage
- Create a new landing page for legacy DOI informing user that data has been replaced by a new version, and that the legacy version of data can be retrieved from tape upon request
- Update legacy DataCite metadata so DOI resolves to new landing page
Routine DOI dataset extension in time
- Archive new files
- New files inherit existing dataset DOI
- Need to regularly update temporal coverage in DataCite metadata
Removal of DOI dataset
- Create a new "depreciated" dataset landing page to explain dataset status and end user options
- Dataset file set is preserved and files can be restaged for access
- Dataset file set has been deleted from system, explanation provided
- Update DataCite metadata so the dataset DOI now resolves to the special “depreciated” dataset landing page
Small scale data replacement within a dataset
- Files to be replaced are removed from the online storage (disk hosted) dataset file set
- Files to be replaced are moved to new names that designate version history on the decsdata directory, and set to status "V" -version controled. These files are no longer visibile to the general user community
- IVC is the "internal version control" number and DTS is the "date/time stamp" are saved as history in database
- Re-assign IVC across complete file set
- Outcome:
- Original files permanently preserved
- Replacement files published for public access
- Version history maintained for internal provenance tracking, and can be used to support end users if needed
- DOI remains the same