OT: Archiving data (LONG)

From: McFadden, Mike <mmcfadden_at_cmh.edu>
Date: Wed May 31 09:14:37 2000

Archiving data has as many definitions as there are people interested in it.
At work we talk about archiving x-ray images on film for 7 years for adults
and 25 years for children. Physical loss of the data is more of a problem
than deterioration of the media. I have data on 12" optical packs that are
5 years old and there are conversion/data retrieval services that will
retrieve my data but the cost is prohibitive. If I tell the powers in
charge (PIC) that it will cost them $5 to retrieve an image that was created
for $1 they think I am crazy. However if there is a lawyer making the
request and the risk is $10,000,000 then the PIC are happy to pay.

Currently we may have for a single patient's cardiac angiography study about
1,000 35mm images, about 1 GB. The system now uses CDROM-RW and puts the
images all on 1 CD using JPEG compression, they swear it will last 25 years.
I guess we will wait and see.

The big issue for magnetic media archival is that it is an ongoing process
with very few retrievals ever actually requested. If every 5 years I need
to refresh my magnetic data then I have a reoccurring cost. This seems to
be what many DP operations are planning. If I ask for lots of dollars for
saving old data most organizations will laugh. I'll bet when reality or a
lawsuit hits they will hope to beg forgiveness for loosing the data.
 
Most media storage lifetimes are extrapolated from accelerated testing and
assume optimal humidity and temperature. If you want proven lifetimes then
there are only several known/proven methods. I know the government and the
Mormons, two separate entities, are storing data etched on iridium
substrates in human readable form. They talk about lifetimes of 1,000's of
years. Other examples are stone tablets, photographic film, and paper. I
think that we can eliminate stone tablets for storing magnetic data. Paper
seems to be cumbersome and is fairly fragile. That leaves photographic
methods. Photographic images from the Civil War still exist. Source code or
HEX dumps on photographic media may be the best way.

Maybe photographic film quality "paper tape" is the answer. The other
solution may be photographic film floppy disks. I'd better get right off to
my patent attorney and file.

I've started babbling.
Mike
Received on Wed May 31 2000 - 09:14:37 BST