Jump to content

ETD Guide/Technical Issues/Backups; Mirrors

From Wikibooks, open books for an open world

An archival infrastructure for ETD should not only consider document format or the use of digital signatures, but also a consequently run concept for mirroring and backups.

Mirrors ensure, that archives run stable, regardless of usage, bandwidth and location.

A common server concept for an ETD archive is the following one:

  1. Production server
  2. Archive Server (Storage Unit, like IBMs Tivoli Storage Management System)
  3. Public Archive server
  4. Archive Mirror at different geographical location.

The production server is the server that communicates with the users, especially the authors. Here the uploads and metadata processing is done.

The archive server ensures that a permanent incremental backup of the public archive server is done. Only system administrators have access to the archive server. This secures the access to the ETDs and their digital signatures. It prevents users from manipulating ETDs and their authenticity and integrity.

The public archive server itself is the server that is known as the document server from the outside. Here the ETDs can be access by users, the retrieval is available. We advise to secure this server with a special RAID (Redundant Array of Independent Disks) system, that allows a security in case of a hardware failure of the disk, that might be caused by extreme usage of the server. A RAID system holds internally 2 equal copies of the server distributed on independent hard disks. So if one copy cannot be accessed due to a head crash or other hardware inconsistencies, the second copy will take over functionality and operate as the first copy. The user itself will not realize, which copy is actually in use.

Additionally to the archive server there should be another storage management system server, an archive mirror that mirrors the original archive. This is important if due to environmental accidents e.g. fire, earthquake or anything else destroys the original archive server. So a data loss, or the loss of the full archive can be prevented. This backup archive should be located at another geographical position, even at another continent.


Next Section: Dissemination of ETDs