I'm not sure I'd say LOCKSS has any resemblence to RAID, other than the
mirroring aspect of RAID-1 (remember that RAID is used to increase the
data throughput of a storage array, not just provide backup capability).
But it is a fascinating concept! From the description it's more like
conventional redundant backup, except that it relies on an essentially
blind process which, when implemented, will result in some quantity of
backup copies in existence somewhere.
This works fine with the kind of material it is designed for (print
journals deployed as web-available files) but is it applicable to the
kind of media storage that AMIA members are concened with?
-- Eric Wenocur
Lab Tech Systems
Ronald Murray wrote:
> The Stanford developed LOCKSS system addresses the threat of data
> loss at the level of the entire system:
> It seems to be the underlying concept behind RAID implemented at the
> whole system level. See also the recent RLG DigiNews article for a
> bottom-up analysis of the digital data loss issue:
> Ron Murray
>>>> [log in to unmask] 01/27/06 9:20 PM >>>
> Thank you for the technical info.
> The issue here has to do with the possible total loss of the data.
> Archivists must be certain that any new system that they buy into is
> totally reliable OR does it also require Backup?
> At this time, there is no electronic system that is 100 percent
> reliable so Backup is a given. RAID-6 may be the exception.
> What upsets me is that some people seem to think that RAID-5 is 100
> percent reliable--and, IT IS NOT!
> Therefore RAID-5 should have 100 percent Backup or RAID-6 should be
> used--with no Backup necessary.
> What is wrong with my reasoning here?
> Jim Wheeler