CLARIND-UdS: Data Preservation
The deposited data are stored on a running virtual machine. They are secured against
- Physical loss
- Change of the data
- Unauthorized access
Protection against physical loss
The virtual machine run on a distributed hardware environment in two separate buildings. The storage is realized with RAID disks. In addition, we the inhouse IT takes daily snapshots of the virtual machines, and keeps them for at least six months at a secured place. IN case of data loss, the virtual machine can be reconstructed from a snapshot.
The servers for the virtual machines and the backup servers are located in different buildungs and only accessible to technical staff. The snapshots are scheduled as follows: 4 snapshots per day for the first 95 days, allowing a dense rollback, some snapshots are kept upto two years. The integrity of the snapshots is checked weekly. There is a backup on hard disk drives every six month. Those backups are stored in a fire-resistent vault (fire resistency class LFS 30P (EN 15659)).
Protection against change of data
The repository monitors the integrity of all archived files using two kinds of checksums (MD5 and SHA1). The checksums are controlled on a daily basis, and in case of a checksum mismatch a backup copy of the resource will be restored.
Protection against unauthorized access
The virtual machine of the repository is run behind a firewall. From the outside, only the standard http and https ports are accessible, and they only provide access to resources under free licenses. A remote login to the machine is only possible via Secure Shell. The server is only accessible to the repository staff.