University of Reading Research Data Archive

Preservation Policy

Retention

Datasets deposited in the Archive will be retained for a minimum of ten years from the date of deposit or for the retention period specified in any applicable third party policy or contract term, whichever is longer. Exceptions to this policy will be made only if the Archive is required to withdraw data in accordance with any applicable laws or policies or in response to a proven complaint.

Once the minimum retention period has expired, a Dataset will be subject to review by the Archive administration, and may be withdrawn if deemed to be no longer in scope of the Collection Policy.

Metadata Records will be retained indefinitely, including those for withdrawn Datasets. Where a Dataset has been withdrawn, the Metadata Record will be updated to reflect this change in status.

Preservation

The Archive will endeavour to ensure continued readability and accessibility of Datasets by undertaking migration of data files to new formats where necessary. It may not be possible to guarantee the readability of some unusual file formats.

The Archive is managed to Information Security standard ISO27001. The original bit stream is retained for all items, in addition to any upgraded formats.

The Archive holds two copies of deposited data files in separate data centres, and places a third copy in escrow with an independent third party.

Withdrawal

A Dataset may be withdrawn on expiry of the minimum retention period following standard collection management procedures; in response to an identified breach of law or policy; or on receipt of a valid and proven complaint, in accordance with the Archive Takedown Policy.

Acceptable reasons for the exceptional withdrawal of data include:

Unless there is good reason to permanently delete a Dataset, withdrawn Datasets will not be deleted, but removed from public view. The Metadata Record for a withdrawn item will be retained indefinitely, and modified to include:

Version control

Once deposited, a Dataset cannot be updated by the addition, removal or modification of data files. Data files are allocated a checksum to facilitate the detection of alterations. Errata and corrigenda lists may be included with the original record if required.

A new version of a Dataset may be deposited. In this case:

Closure of service

In the event of the Archive being closed down, all content will be transferred to an alternative archive providing an equivalent level of preservation.