Archival storage <-- was Re: [EE]: To class
Russell McMahon email (remove spam text)
This is a good enough topic to deserve its own subject heading.
Do YOU know where YOUR data is?
> > Storing floppies that you value in a suitable environment can help.
> Storing anything you value on a floppy without a decent backup is a bad
I agree, with qualifications.
Floppy diskettes can fail at any time and sometimes do. (This is also true
of ANY storage medium). I make at least two copies of anything critical that
I commit to floppy. I have very seldom had problems but it does happen.
I backup small items of value to floppy occasionally as I work if I am using
a laptop and at the end of a session. I change the floppy with some
regularity. I also backup across a network quite regularly and then make CDR
copies of the networked backup on occasion. At major checkpoints in a
project I send email copies to a friend in a city 80 miles South of here and
he stores the files (he tells me) on floppy disks in his sock drawer. (These
are usually PCB schematic and PCB files and assembler source and the sizes
are usually sub floppy sized.)
As a guide. Put the amount of effort into backups that is justified by the
pain that you won't feel when your entire system crashes irrevocably. I know
someone who literally lost a whole thesis to a hard disk crash. ALL HARD
DISKS CRASH - sometime between 30 minutes and about 10 years after first
use. I have seen new HDDs fail at 30 minutes, and 1 week. (And a new laptop
die after 1 day, and ....)
> CD-Rs are much more reliable and will probably outlive the drives to
> read them.
This is a dangerous assumption (even though it MAY be correct).
The early CDRs used a Cyanine dye and claimed archival lifetimes of
typically around 10 years. The next gain was to Pthalo-Cyanine (goldsurface)
and they claimed lifetimes of approaching 100 years. Both claims proved
untrue. There has been much written in this. Partly the chemistry used is
based on what patents you hold. The latest top end CDRs from reputable
manufacturers claim unused storage lifetimes of 20 years and written
lifetimes of 25 years. These are based on accelerated aging tests and
knowledge of how they THINK the chemistry will behave.
Early CDRs failed far sooner than claimed.
I have seen CDRs that have been used for a successful install become "blank"
within a year of being written. On contacting the supplier (a large
organisation that distributed in house software this way) they were totally
unsurprised, assured me it was a common occurrence, and sent another disk.
After this experience the same organisation now uses exclusively Kodak CDRs.
(Standard disclaimer - no shares, profits, associations etc ....). YMMV.
Storage conditions (temperature certainly, also perhaps humidity and ???)
will affect CDR lifetime. CDRWs may behave differently. Note that lifetime
guarantees are USELESS. They may replace a CD but the lost data is (usually)
far more valuable. How many people actually try to claim for failed CD's or
ALWAYS make at least 2 x CDR backups of critical data. Store separately.
Offsite if critical. If archiving long term then CRCing the contents of each
eg annually and remastering the data from the best combined data available
MAY save you.
HDDs are getting cheap enough that using a networked drive as backup only
may be better than CDs. An 80 GB IDE drive costs about $US150 here and is
equivalent to about 120+ CDs. That's about the same price as 2 el-cheapo or
one top brand CD per CD image. Worth thinking about. Lifetime may be less
than a CDR BUT as long as there are 2 images available and failure can be
(reasonably) guaranteed to not affect both at once then recovery from a
partial loss may be far simpler.
In this world of high technology rotating mass memories, acid free paper and
known stable ink is still pretty good stuff :-)
http://www.piclist.com hint: To leave the PICList
See also: www.piclist.com/techref/index.htm?key=class+EE%5D
You must be a member of the
piclist mailing list
(not only a www.piclist.com member) to post to the