Running tape, or similar backups to ensure the long term preservation of all the data stored in the Dataverse is an implied responsibility that should be taken most seriously.
In addition to running these disk-level backups, we have provided an experimental script that can be run on schedule (via a cron job or something similar) to create extra archival copies of all the Datafiles stored in the Dataverse on a remote storage server, accessible via an ssh connection. The script and some documentation can be found in scripts/backup/run_backup in the Dataverse source tree at https://github.com/IQSS/dataverse . Some degree of knowledge of system administration and Python is required.
Once again, the script is experimental and NOT a replacement of regular and reliable disk backups!