Service level definition

Introduction

Edinburgh DataShare is an online digital repository of multi-disciplinary research datasets produced at the University of Edinburgh, supported and administered by the Research Data Support team within Information Services. It is designed to be scalable and highly resilient. It is available from the University website at:

• https://datashare.ed.ac.uk

Functionality

DataShare is an online system which allows University of Edinburgh researchers to publish, share, describe, embargo, and license their data assets for discovery and use by others via the Internet. The repository includes a metadata schema compatible with repository harvesting protocols, a user interface for deposit and administration, search and browse facilities, item-level usage statistics, time-stamped submissions and persistent identifiers.

Depositors may include current and retired University staff, University visitors or collaborators associated with a University research project, enrolled postgraduate students, or undergraduate students approved as depositors to a collection by their supervisor.

Depositors may:

• Register for the service using their University login.
• Request new collections to be established within the repository.
• Be given a range of privileges to administer Collections, including giving other registered users deposit permissions, and moderating submissions, under special circumstances.
• Upload data files, including documentation and multiple formats. Submit each dataset or item with descriptive metadata, subject to quality control. Larger or more complex deposits consisting of many datasets may be ingested by arranging batch import. 
• Embargo the availability of a data item until a future date or indefinitely. Metadata remains publicly viewable, and the system allows the depositor to accept or decline public requests to access data under embargo. 
• Request withdrawal of a data item from public view or supersede it with a newer version. A public metadata landing page will remain. 
• Use the citation and permanent identifier generated by the repository to cite the dataset or data item in published papers.
• Attach an open data licence to each data item by default or alternatively include a statement of intellectual property rights.

The service operates according to a set of policies, which are available from the home page of the repository. These include a "Data and metadata policy", submission policy, preservation policy and a depositor agreement.

Availability

Edinburgh DataShare is available 24x7. Target availability is 99% measured over a year, excluding planned maintenance periods. Core hours of support are 9AM to 5PM during University working days.

There is no regular scheduled maintenance period, with all routine maintenance announced on the service home page in advance. Any reduction in performance or resilience during maintenance periods will be managed to reduce its effect, normally resulting in only a few minutes' partial service loss.

Resilience

The service is designed to be highly resilient. Edinburgh DataShare runs on a server at the Mary Somerville Data Centre in Edinburgh using storage from the University Storage Area Network. Edinburgh DataShare is automatically replicated to a backup container sited at Appleton Tower, 2.5 km away from the Mary Somerville Data Centre.

Backup

The Edinburgh DataShare repository contents will be backed up to tape daily and database snapshots will be available on disk for two days. Database snapshots will also be backed up to tape. Daily tape backups will be retained for 30 days and a monthly archive copy will be kept for a year.

Disaster Recovery

The primary solution in the event of a major failure at either Kings Buildings or Appleton Tower is to continue service at the remaining site. In the event of data corruption affecting both sites (data corruption or deletion being automatically replicated) or a major external event, a service copy can be restored from tape backups.

Support

User support is provided by the Research Data Support team, as part of the comprehensive Research Data Service. Any problem reports or queries should be emailed to us via the Contact box above.

Enquiries can also be sent to IS.Helpline@ed.ac.uk where they will be passed to Research Data Service staff for resolution.

Staff are available for consultation or extended support according to demand. Extensive online documentation and guidance is available via:

• https://datashare.ed.ac.uk