[Tig] Archiving (was Re: Holographic Drive at NAB)

Jack James jack at surrealroad.com
Tue Apr 22 18:27:38 BST 2008


The good thing about disk drives, especially when configured with a  
massive amount of redundancy, is that you can replace them as they  
fail. The same cannot be said for tape unfortunately. The other  
problem with tape is that they require an intermediate medium-- you  
can't load data to or from tape directly, but have to go via something  
else (typically disks).

Having said that, we use LTO3 backups for all our important data,  
mainly because it is so cost-effective. Over the years (and after  
getting burnt a lot of times), the number one rule I have for data  
backup is not to depend on any one system at a time, in fact right now  
I use off-site tape backup for medium-term, RAID local storage for  
short-term and an Internet-based system for critical data, but  
ultimately I ensure that I always have (at least) two copies of  
anything I don't want to use.

Jack
---
Surreal Road
Emmunicate Insatiably...
www.surrealroad.com





On 22 Apr 2008, at 17:36, Bob Friesenhahn wrote:
> 1826 subscribers as of April 2008
> Gary Adams supports the TIG
> http://colorist.org/wiki3/index.php/TIGNABFS08
> ====
>
>
> On Mon, 21 Apr 2008, Jim Houston wrote:
>
>> This is a poor idea.  As it turns out, hard drives have significant
>> problems from an archival standpoint that make tape a better and
>> safer medium. Among the issues are that hard drives spin on a
>> bearing that is coated with a volatile oil.  Whether you run the
>> drive or not, this oil degrades over time and can cause stickion.
>> Most drives are designed with a 3-5 year lifetime and the failure
>> curve at the beginning of their life and within months of the end of
>> the design life are very high. The built-in heads and unsealed
>
> On its way to the list, this email will pass through a disk drive
> which has been running continuously for 10 years.  The system which
> preceded it had a disk drive which ran continuously for 12 years.  I
> was glad to discard that disk drive (still working fine) since
> full-height disk drives are very loud and the capacity was small.
>
> I have not heard mention of "stiction"
> (http://en.wikipedia.org/wiki/Stiction) for many years since a
> discussion with Ward Christensen (XMODEM inventor) on a forum in the
> early 90's.  As Ward described the issue, "stiction" is actually a
> problem with the heads sticking to the media and not a problem with
> the bearings.  The problem is completely gone now since all modern
> drives park their heads off of the media and don't leave them sitting
> on the media.  By the way, Ward recommends that if you encounter a
> drive with "stiction" and a solid hit on its side may free it up.
>
>> (and as far as the Holographic drives go, I heard a detailed talk on
>> all of the error-correction mechanisms that had to be designed into
>> the InPhase drives.  While it may prove out to have solid
>> engineering, anything that needs as many layers and methods of error
>> correction as this drive does is a bit suspect from an archival
>> standpoint. With holographic storage, if you lose alignment of the
>
> Tapes also use massive error-correction mechanisms. As bit density
> increases, the probability of error for a bit increases so either the
> error-correction encoding needs to increase, or else the user faces
> less reliability.
>
> Regardless, the main question remains if the hardware to make sense of
> the medium still exists (and works) when it finally needs to be
> recovered.  These tape drives you like probably use specific
> interfaces which will require an archaic computer (with those failed
> hard drives) when it comes time to restore them.  The capacitors on
> those tape drives will have degraded and popped and joints will have
> corroded.  After even 20 years, it will be quite difficult to recover
> the content of a tape, or even find a computer which will interface
> with a hard drive.  This becomes more and more of a problem as
> technology continues to accellerate.
>
> Bob
> ======================================
> Bob Friesenhahn
> bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
> GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/
>
>
> _______________________________________________
> No personal abuse; absolutely no advertising or marketing on the  
> main TIG mailinglist.  Emergency
> tig contact address is rob (at) calarts.edu
> Tig mailing list
> Tig at tig.colorist.org
> http://tig.colorist.org/mailman/listinfo/tig





More information about the Tig mailing list