University data storage options
Services provided by the University for storing and curating data
The University and Division offer several storage facilities that may be of use in your research.
Usage notes for the Microsoft OneDrive storage offered by the University
All University researchers with a Nexus account have a 5TB OneDrive share that can be accessed over the web or synchronised (either in full or partially) to your local computer.
For usage instructions see:
Files are versioned within OneDrive - 30 days of rollbacks are available, either for the whole drive or for individual files - right click on the file in the macOS Finder or Windows Explorer and choose Version history.
For important files you are strongly recommended to back them up via some additional mechanism - due to the selective sync (i.e. files remain in the cloud until you open them) you cannot rely on automated backups of your computer to protect these files, so you are advised to copy them to either force OneDrive to always store the file on the device being backed up (see below) or to copy the file(s) to some appropriately protected (e.g. encrypted or departmentally/divisionally provided shared) storage
All files are irretrievably deleted 60 days after the SSO for the OneDrive account expires. You must ensure that all important research data is transferred from the OneDrive account well in advance of you (or your research group member) leaving the University.
Cloud First Files
By default OneDrive will automatically move files to the cloud service, leaving a stub on your local device. Files/folders not locally stored will be visible In the file browser by the presence of a cloud/arrow symbol. This mechanism of operation has two side effects:
- These remote hosted files will not be openable on your device without a network connection. If you know you will need these files when you are unable to access a network you should download the file before leaving the network - do this by simply opening the file. It will now stay on your device until OneDrive decides that it needs to free up space. To ensure the file is permanently held locally, right click on the file/folder and choose the option 'Always keep on my device'. You can reverse this by choosing the 'Free up space' option.
- Cloud located files will not be backed up by any locally running backup applications. If you have a particularly important file that must be backed up then you can either force it to be always stored locally (see point 1), or take steps to manually back it up by copying to an additional location.
Usage guidelines for storing research data in Microsoft Teams
Each team within Microsoft Teams is allocated 5TB of storage, suitable for all data classifications. Microsoft provide a 30 day recovery facility for files stored within Teams, see below for details on browsing the history.
This shared space is useable by all team members so any one could delete/damage files accidentally and if team members have folder synchronisation setup onto their desktop/laptop computer then any malware on that host may effect your Teams hosted files.
You should also be aware that should all members of the Team leave the University, then all data held in that Team will be irretrievably deleted after 60 days.
The Teams bulk recovery system is not fine grained - it can only restore the state of the whole Team storage to a particular point in time.
If you find you have damaged a Microsoft Office file then you can use the application to view the version history of the file to recover earlier versions.
For other file types, if you navigate to the folder containing your damaged/lost file in Teams and then use the 'Open in SharePoint' toolbar option (it may be hidden in the '...' menu).
If your file has been deleted then use the 'Recycle bin' menu item on the left to view the deleted files.
If your file has been damaged, then click on the '...' in the line for that particular file and use the 'Version history' menu option to browse the available versions of the file.
Users of the BMRC compute cluster have twol storage options open to them, a high-performance research data file system attached to the BMRC compute cluster and a read-only 'archive' platform
Live research storage
The BMRC compute cluster uses a high-performance storage platform utilising DDN GPFS hardware. This allows multi-GB/s concurrent access to your files across the compute cluster, but does not offer any backups of your data. You should consider it to be the equivalent of the WIN hosted 'scratch' file system, although it is more resilient to hardware failure than our scratch server.
Storage at BMRC is allocated on a per-project basis, to ensure good segregation of data from other projects and is purchased directly from the BMRC team by the project lead/grant holder. The project lead should contact the BMRC team to request the creation of a project folder and the necessary access control groups.
At the time of preparation, the cost for this storage is £60 per TB pa, but please confirm with BMRC before basing any grant application on this figure as this is reviewed on a regular basis.
As this store is not backed up you will need to take additional steps to protect important files and data.
Raw MRI/MEG scan data collected on a WIN scanner should already be curated by WIN (and XNAT) so ordinarily would not need additional protection.
If you are processing data from a third-party, then if you are confident that you can obtain the data again at a later date this would also not need backing up.
Processing scripts should be kept under some form of version control in a central resource, for example WIN's GitLab.
For meaningful derived data there are two immediate options for protecting this from damage; BMRC's archive facility and WIN hosted storage and archive facilities, but other storage options may be of use.
Taking a backup of your data to WIN's storage is as simple as using a secure file transfer program; 'scp', 'rsync' are supported options, but BMRC also supports Globus for file transfers of large datasets. Once on WINs storage you may send it to tape if desired.
BMRC also offer a charged for 'archive' facility. This is a low performance single copy/same-site storage location charged at a significantly lower rate to the high-performance storage. It is presented as a read-only storage location on the principal servers, so provides protection against accidental deletion or damage but offers no disaster protection. Contact email@example.com for details on charging and to request archival of data.
Divisional warm archive/non-critical file store solution
MSD-IT provide a storage platform, 2TS, which is suitable for small to mid scale research data storage. It is not deployed in a high-availability arrangement, so is not suitable for storing data that must always be accessible.
Details on costs and the service in general can be found on the MSD-IT 2TS page.
University provided file storage system for research data
In March 2023 the University will begin roll out of a new service for storing in-progress research data. This product has been called Research File System (RFS) and will provide access to a storage location from any University network location. At launch, the facility will not be approved for confidential data, so should not be used by WIN researchers to store personal information (e.g. human MRIs). As and when this approval changes it may become an option for your project's data storage.
Project page - https://projects.it.ox.ac.uk/research-file-service-rfs
University operated data archiving service
The University provides a cloud based secure archiving facility, branded DigiSafe. This service provides multiple-offsite backups of both 'warm' and 'cold' data (e.g. occasionally accessed (warm) or rarely/kept for compliance reasons (cold) data. Data storage is charged on a rental basis and is managed at a Departmental/Centre level.
For pricing details see the DigiSafe page, and if you are interested in using this service contact firstname.lastname@example.org to discuss.