CD-R Warning, if ever one were needed.

Totally true.

I only posted this for the hilarious ad ;)
--
--Reinhard
He he, i just wish Mr Cleese would do more computer adverts. I'm thinking specifically of the Michael Jordan IBM advert, (employed to run around the server room, fast enough to fix all the problems, but redundant because of new management technology) and how fun it would have been as a send - up, not the oh - so - gentle know - it - all tones that seem very popular with big corps.

I remember when that ad (linked by Reinhard) launched, and i've not heard anything about the company since. But it's a good ad, and always fun to watch JC do anything.

Why is it though that it reminds me of the spots Peter Sellers did for TWA?

Something melancholy about comedic heroes doing commercial work, late career.

Oh well, much better than Barclays [UK bank] doing their puffed up "You Need a Big Bank" multi - minuters with (amongst others) Donald Sutherland.

oops, drifting OT, so enough already :)
 
LOL! I love it.

Back up Hard Drives for me. I have a quad bay full that I back up to. Also have a drive I fill and store at a family members house with my originals on it just in case.

HDs are cheap nowadays. Why not?

Oh yeah I do DVD-R as well for the originals.

Darrin
 
But i don't doubt at all you have 10 yr old CD-R media that
is just fine. I think the IBM'er is pointing to poor quality dye
substrates.
Ummm - no. "Gerecke said. "Some of the better-quality discs offer
a longer life span, of a maximum of five years." Didn't you read
the article?
I merely wasn't doubting the personal report. Unconnnected comment
with the article.
Well no, you directly connected it to the article by mentioning what you thought the IBM'er was pointing to. And in fact, he was quite explicitely pointing to good quality discs having five year expectancies as a maximum, not at low qualty substrates having that kind of maximum life expectancy. The lower quality discs were given even less than five years. That is very much at odds with the ten year report. So you got it quite wrong, which is odd since that was one of the core points in the article. Based on what the IBM'er said, a 10 year disc would be quite unexpected and casts a pretty bad light on either Gereck's assessment or the article's ability to report accurately.

It seems to me that the reasonble thing is either to doubt the ten year report or to doubt the article's or Gerecke's assessment.
The article raises a very valid point - no
QA measurement standards! That's bad. I keep saying it to people:
one day there'll be a world shattering wail over lost memories.
I searched the text. The word "standard" or "standards" is never
mentioned. Neither is the term "QA" or the phrase "quality
assurance". What article are you referencing?
Well, i read the article, and it was very short on explaining
itself, leaving me to infer what the quoted IBM guy was hinting at.
I think you failed to comprehend what i wrote, and that a point can
be raised by implication. You've read my post as literal grammar,
incorrectly.
Yes, it is an ambiguous statement. But since you previously attributed the article with content that it didn't include, your comment certainly made me wonder.

--
Jay Turberville
http://www.jayandwanda.com
 
I think you're writing as if i was some kind of nutcase, or at
least deliberately with a hint of sarcasm, as if to imply i am de
facto wrong.
Not at all. I'm just trying to point out that the CD-R can actually be quite convenient and that in our case at least, convenience was one of the specific reasons we went with it. Furthermore, I was giving examples as to why. Further convenience factors I didn't mention are that you don't have to worry about driver or software support in future computers and operating systems for quite some time. I'm not up on current tape software and hardware, but at the time we shifted, that was a significant issue.
When CD-R first came out, there were plenty of issues IIRC with
quality of burning software / buffers / throughput & buffer
problems that could lead your disc to end up a coaster.
Right. Which is why we didn't start archiving on CD-R until 1997. By that time, there were a number of drives that were highly reliable.
FYI i do not need to load a tape to idex it or search file metadata
  • that information exists in a (separately backed up) database
created by Arscserve (similarly for other backup apps) and if i
want to i can expose this to the OS by setting backed-up volumes as
remote media.
But do you have to restore to actually load a file? With a CD-R the data is directly moutable and usable.
I really don't see how any of your operation is easier than leaving
a system set to twice daily incremental backups, and weekly full
backup, running in the background. I prefer zero user intervention,
where it concerns anything that's a important procedural decision.
Now you are discussing system backups, not archives. Those are two different things. I also prefer as little user intervention as possible. So I use drive mirroring and a little program called AutoSave that makes copies of important designated directories and their sub-directories in near realtime. It works very well for our operation with only occasional user intervention needed to purge old backups directories that have been archived.
There's also, IMO, a handling risk involved with optical, if you
accidentally scratch the disc, and in pulling individual discs,
greater manual organisation. Certainly i'm not arguing your system
is inadequate, but other considerations may be more important.
Right. I don't claim our system is ideal. But it is very easy to use and we archive a lot of data.
For
example, i don't like drag and drop because (on NT thru XP) the
creation date is altered, leaving you with only the modified date
to indicate the date origin, which will be invalid if any file sees
another edit.
And for photographs with EXIF data this would be of no importance. Or am I missing something? Creation dates have never been important in our operation, but if they matter, then I agree that this would be a problem.
The final killer is your data creation rate.
Yes. The optical disc is certainly not the solution for every situation. My point is/was that neither is tape.
You say you "went to CD-R". Did you use tape before that? Not all tape
is created equal, and certainly there are some tape systems i would have
strong opinions about, technologically, even before considering other
factors.
Yes. And the better, faster systems back then were quite expensive. Archives and restorals were a major pain.
Basically, i think the options simply got a lot better in recent years, and
the price premiums diminished substantially, too.
I'm sure they have. But I try not to mess with systems that are working fine. And I'm sure that tape can present some excellent solutions. But I do wonder about future hardware, software and driver support. Though perhaps USB and "firewire" have made that less of an issue?

But frankly, I think the "CD-R is unreliable" reports are overstated. I suspect most data losses are from poor storage and handling, cheap media and defective readers. We keep off-site archives at my home and I just popped an 8 year old archive in my cheap $89 DVD-R recorder. It reads just fine. I have personal stuff (old download archives of shareware etc.) that I made on cheap, no-name CD-Rs that reads perfectly as well. There are just too many data points for me to think that I'm just lucky somehow.
best from me,
Likewise,

--
Jay Turberville
http://www.jayandwanda.com
 
Hi again,

i'm not meaning to nitpick, but i do feel i am having to write again long - hand, what i already wrote. This hopefully should dismiss any perceived ambiguity, not that i think i was ambiguous to start with :)
Well no, you directly connected it to the article by mentioning
what you thought the IBM'er was pointing to. And in fact, he was
quite explicitely pointing to good quality discs having five year
expectancies as a maximum, not at low qualty substrates having that
kind of maximum life expectancy. The lower quality discs were
given even less than five years. That is very much at odds with the
ten year report. So you got it quite wrong, which is odd since
that was one of the core points in the article. Based on what the
IBM'er said, a 10 year disc would be quite unexpected and casts a
pretty bad light on either Gereck's assessment or the article's
ability to report accurately.
You make good and fair point. But also you obviously recognised the distinction, and that i wasn't making categorical attributions, so i wonder why you chose your argument as you did. No harm done.

The article is in fact very poor, as if it was really a product of a boozy holiday season for hacks. It doesn't even state whether it is current production media. But one might infer that, when it's posted as a news article. So my "get out of jail" card is the question whether CD-Rs were made of sterner stuff back in the day they were impressive to have.

This is why i didn't hold a line and doubt a 10yr old disc being fine. Not only because of that, but because 1) there's always a possible exception, by dint of chance 2) arguing publicly over one person's anecdotal evidence is a sure way to drag the discussion downhill 3) as you may have noticed, my biggest point is not "CD-R is so useless, you must go to tape" but the relative benefits of tape once you scale into bigger data sets and that i see no point in the optical game at the same time as many potential pitfalls.
It seems to me that the reasonble thing is either to doubt the ten
year report or to doubt the article's or Gerecke's assessment.
You are right, in principle, from your perspective. Please try to comprehend, i did not post the article with a comment "this guy is totally right, i believe everything he says as God Gievn Truth". I merely wasn't towing that particular line that you have fed me above, nor am i. I'm skeptical both of Gerecke's arguments, because they're almost off the cuff, as well as i'm skeptical optical media makes any sense for long term archival. I hope still that someone who really knows about optical writable discs can add some more concrete information as to the state of things. But i digress . . .
Well, i read the article, and it was very short on explaining
itself, leaving me to infer what the quoted IBM guy was hinting at.
I think you failed to comprehend what i wrote, and that a point can
be raised by implication. You've read my post as literal grammar,
incorrectly.
Yes, it is an ambiguous statement. But since you previously
attributed the article with content that it didn't include, your
comment certainly made me wonder.
Where was it i categorically attributed anything to the article? I've just re-read the entire thread, and nowhere do i make claims about what the article says. I did speculate about what the article might import. You have inferred intent from my words, which assuming rather a lot i think, when you take what you see in my words, as opposed to what was written, as argumentative "fact".

I don't feel like debating comprehension, but please also don't take my comments personally.

I think the article conclusions are startling, if you consider how little context is given. But like you misread my intent above (that somehow i should either dispute the poster's longevity claim or else dispute the article), i was making a wholly other point - that top tape drives are surprisingly affordable to even a serious amateur, and that owning one circumvents a host of nuisances associated with optical writable media, of which the linked article is just a reminder.

One last time, with a caveat this isn't intended as personal comment, but i feel you're making a semantic mountain out of a molehill, and are skilfully reinterpreting my words contrary to both fact and intent. This results in false and useless argument, and if you display too, as above, little understanding of how a tape system is actually operated or what it's capabilities are, then i think your effort wholly specious. N.B. i am not claiming particular knowledge about optical media, just they have in my experience long been rejected for long - term high capacity archival in commercial and Govt. applications. There are two parts to that argument - handling and capacity inconvenience, and studies indicating the failure or bit rot rate doesn't give any advantage over tape, if you factor in recopy necessity for tape as well as optical. That equation may change however with holographic storage, or much higher density writeable optical. Other factors such as data transfer rates may very well give the final edge to optical, as tape is just about maxxed at 20 to 30 MB/s, compressed, unless you use multiple drives. We do need much much faster transfers.

cheers!
  • kirbs
--
====================================
Proof, it it ever were needed, that Mr. Rockwell is not a Brit
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

'This may be a blessing in disguise for fast-shooting fudge packers, since you'll start having the D200 lock up on you before you're really full, and it will free up again for another few shots.' [ http://www.kenrockwell.com/nikon/d200/d200-high-speed.htm]
= = = = = = = = = = = = = = = = = = = = = = = = =
 
may well be behind me, but not for everyone.

so here's an attempt at balance:
Not at all. I'm just trying to point out that the CD-R can
actually be quite convenient and that in our case at least,
convenience was one of the specific reasons we went with it.
Furthermore, I was giving examples as to why. Further convenience
factors I didn't mention are that you don't have to worry about
driver or software support in future computers and operating
systems for quite some time. I'm not up on current tape software
and hardware, but at the time we shifted, that was a significant
issue.
Driver and software support! Don't get me started. Try importing legacy Veritas NetBackup media into CA Arcserve. (it's doable, just about, but you start to value CA's 24/7 and genuinely good support)

What you'll find is just as most big databases write direct to disc, bypassing the OS FS, most backup sioftware of serious persuasion, ships with custom drivers.

You'll also find that in this field, legacy support is not ignored like it is in the consumer world. If a business somewhere relies on it, a company will provide support. I could think of many examples, but for some apps., a outmoded ISAM database performs just fine, thank you, and that obsoleted tech is still actively sold.

The corrolary, is what happens when everything is blue - ray, and you can't buy a CD drive any more? (guess that won't happen whilst HD - DVD is around, but it's possible)

But as you i think pointed out, after all, it's just data, so migrate the stuff. Think you'll find it a darn sight easier with tapes and not hundreds or even thousands of 5 inch silver or gold discs :-)
Right. Which is why we didn't start archiving on CD-R until 1997.
By that time, there were a number of drives that were highly
reliable.
Fair cop!
But do you have to restore to actually load a file? With a CD-R
the data is directly moutable and usable.
Yes, you run a restore. That much is i hope obvious with tape - that you don't, unless you earnestly need to, like in a bare metal restore, even mount a tape as an extension of the FS directly. But you have to understand that it doesn't take more than a few minutes, and can be quicker than identifying a disc on a shelf. I don't think this is a point you score, unless you qualify that we're stuck off - LAN with a laptop, which case you'd be dead right on every count. Basically, you can't think in terms of "runs everywhere" with tape, and it may be very important to you to have a standard format. I'll burn DVDs too, for posting purposes, but inside the shop, we strictly do not use DVD or CD-R for anything else. Sadly, i have experience to consider any removable writeable media to be a security risk.
Now you are discussing system backups, not archives. Those are
two different things. I also prefer as little user intervention as
possible. So I use drive mirroring and a little program called
AutoSave that makes copies of important designated directories and
their sub-directories in near realtime. It works very well for our
operation with only occasional user intervention needed to purge
old backups directories that have been archived.
Well, our backups become archives, when stored. Anything current is on or near - line for access (plus a whole lot of nearly legacy stuff that we keep around for reasons that include having searchable access to what we did for who). Once the tape is no longe rpart of a working set, it is archived, and whole images are taken, time to time, at bitcopy level, in addition. All that gets archived, even if potentially it is a working backup, if restored. I don't see the difference that you find here. I mean, to me, archive is just multi - year storage. Saying i'm talking backups not archive makes no sense to me.
Right. I don't claim our system is ideal. But it is very easy to
use and we archive a lot of data.
Personally, i consider all data to be potentially useful, and i try to make the difference between online capacity and total data stored as minimal as possible. But then discs are 50% or more of our budget, and i've stuck to that. How much is a lot, BTW? There's also differences in how backup and archive suit working styles - if you turn around a job quickly, it may make sense to just burn a disc and put it on the shelf afterwards. But that makes no sense at all if you need data accessible to lots of people or work long project cycles, when if you really have masses of data and no disc space for them, you can add HSM options to the kit, and let users pull tape files just as if they were mounted.

[hit the char limit so have to split this one]
 
And for photographs with EXIF data this would be of no importance.
Or am I missing something?
EXIF often gets stripped. EXIF isn't maintained in a PDF-X file for pre-press, nor in a complex layered edit. I think you're assuming i only ever touch RAW files, which i've hinted before isn't the case :)
Creation dates have never been
important in our operation, but if they matter, then I agree that
this would be a problem.
Depends on how you bill your clients, how big they are (in terms of burocracy and asking time to time your billing policy) and the sad fact that law suits are just a fact of life in certain fields. If you're certain you'll never need all that meta data intact, then fine. But i've never treated my business with the approach we'll never need something. Overkill, i guess can only be judged in hindsight.
Yes. The optical disc is certainly not the solution for every
situation. My point is/was that neither is tape.
Well, we have equations, rather than certainties, for sure. But let's say you're a working pro, in a small shop i.e. you have colleagues take care of production or someone else is the dedicated photog. Not unusual at all. Someone like this can be generating terabytes of new work a year, and yet i see and hear reports of many who think DVD is just fine, as their shelves groan under the weight. Your DVD-R came with your computer, most likely. My tape drive and backup software most likely cost more than the typical computer budget. But it was comparable in price to many a lens, and takes away one heck of a lot of work.
You say you "went to CD-R". Did you use tape before that? Not all tape
is created equal, and certainly there are some tape systems i would have
strong opinions about, technologically, even before considering other
factors.
Yes. And the better, faster systems back then were quite
expensive. Archives and restorals were a major pain.
LOL! "quite expensive". I think you jest with your understatement. Aynhow, Sony drives like i use are now about $2K and utterly robust and backwards compatible with older media. Basically, you have AIT, Ultrium and good old DAT. Ultrium is too slow and plagued by access times (linear 32 track, IIRC, though they may have uprated that) DAT ain't got the capacity, and AIT to my mind rocks - enough capacity, good transfer rate, media with inbuilt indexing RAM chip to keep spooling lead time down, and an advertised and adhered - to roadmap.
Basically, i think the options simply got a lot better in recent years, and
the price premiums diminished substantially, too.
I'm sure they have. But I try not to mess with systems that are
working fine. And I'm sure that tape can present some excellent
solutions. But I do wonder about future hardware, software and
driver support. Though perhaps USB and "firewire" have made that
less of an issue?
Life is still largely SCSI out there, migrating to Fiber Channel, or iSCSI, because you don't see FW or USB in the datacenter, which is what drives the market. Sony do offer a FW drive i believe, and i'm sure others do similar, but only at the low end of capacity. That's dumb - datacenter sized working sets are becoming all too commonplace.
But frankly, I think the "CD-R is unreliable" reports are
overstated. I suspect most data losses are from poor storage and
handling, cheap media and defective readers. We keep off-site
archives at my home and I just popped an 8 year old archive in my
cheap $89 DVD-R recorder. It reads just fine. I have personal
stuff (old download archives of shareware etc.) that I made on
cheap, no-name CD-Rs that reads perfectly as well. There are just
too many data points for me to think that I'm just lucky somehow.
With me, and no, this ain't the reason for my love of tape (!) being a klutz keeps me from evaluating any optical media longevity properly :-) I will also add that a brief but really dumb forway into using Travan drives for home office is not something I remember kindly. Try getting those POS drives in working order today. Or even finding the drivers, which of course, could just have been left on the manufacturer's site, but were taken down as if the drive space to host them was bakrupting HP . . .If you used one of those, you will know what i mean, and if that's what you used, dang it, i will completely understand you'll use anything but tape today :-)

Actually, i think it would be fair to say that most data loss with optical is handling damage, and in handling, i would also consider backup procedure, about which you're obviously professional in your application to the job. I wish sometimes i could finish jobs and just store them away discretely, but i find everything forever interrelated. There's nothing like having a top end machine (at least dual proc), tons of ram, a serious disc subsystem and a little tool like X1 (add .com for product site) that makes light of intense amounts of data. I cannot stress enough that it's great if you can afford enough on-line or near-line [i.e. slower cheaper drives] space, and that i think it the exact opposite point of digital data to archive stuff off-line on a non-human readable disc. But when you have oceans of data kicking around your LAN, then you sure need a more capable system than manually burning DVDs. I genuinely think that it's a travesty that any data is even temporarily unreadable, and so the though that, to boot, the dang disc is going to "rot" on me, beyond convenient observation, gets my sense of irony piqued.

sorry about the split post - better to keep as much context as possible i think.

take good care now,
  • kirbs
 
The corrolary, is what happens when everything is blue - ray, and
you can't buy a CD drive any more? (guess that won't happen whilst
HD - DVD is around, but it's possible)
Right. Dealing with migration once every ten or twenty years isn't so bad. For my personal stuff, I'll be migrating it all to DVD soon. For important stuff, you should be spot-testing for data integrity anyway.
But as you i think pointed out, after all, it's just data, so
migrate the stuff.
I don't think I pointed that out, but I agree.
Think you'll find it a darn sight easier with
tapes and not hundreds or even thousands of 5 inch silver or gold
discs :-)
Depends on the tape formats I guess. I have some DAT tapes I'd hate to have to try to read. And some proprietary (iomega I think) stuff that would be a real treat. But I moved that stuff to CD-R at the time, so I don't have to worry about it.

I agree that migrating hundreds of CDs would be a chore. I doubt we'll migrate the studio stuff. It is already surprising we don't have file format software issues that are greater now. In another ten years, the vast majority of that data will be trivia.

But from a photographer's standpoint, of course you'd have to migrate. And if I was a professional I might approach archiving differently - or I might not. But I'd have to investigate other options either way.
But
you have to understand that it doesn't take more than a few
minutes, and can be quicker than identifying a disc on a shelf.
But don't you have to identify the tape on the shelf too. Anyway, my point isn't whether tape or CD-R is some small amount faster. My point is that CD-R is very convenient. For us, it is nice to be able to quickly browse image, dataset or media files directly from the disc to verify that we even have the data that we are looking for. Random access is nice. I know tapes are very fast, but I have my doubts that they are as fast. Though, the larger the archive set, the more having it all on one piece of media can be an advantage.
I
don't think this is a point you score, unless you qualify that
we're stuck off - LAN with a laptop, which case you'd be dead right
on every count.
No. We're seldom concerned about being off the LAN. I can't say that disc is faster than tape since I'm not up to speed on current tape software and technology. But I can say that disc is so quick that it would be tough to be much faster. The only thing that would be faster would be to just keep all or a lot more of the data online. Something that we are seriously considering doing with drives being so inexpensive.
Basically, you can't think in terms of "runs
everywhere" with tape, and it may be very important to you to have
a standard format.
Yes. Everyone knows how to use a disc drive. Everyone has one. No extra burden is put on the network. No special steps or training.
I'll burn DVDs too, for posting purposes, but
inside the shop, we strictly do not use DVD or CD-R for anything
else. Sadly, i have experience to consider any removable writeable
media to be a security risk.
Yes, discs are a security risk. And it is a lot easier to kill a CD than it is to kill a DVD since the DVD has a tough polycarbonate layer on both sides. But I suspect USB ports and the internet are bigger security risk than discs.
Well, our backups become archives, when stored. Anything current is
on or near - line for access (plus a whole lot of nearly legacy
stuff that we keep around for reasons that include having
searchable access to what we did for who). Once the tape is no
longe rpart of a working set, it is archived, and whole images are
taken, time to time, at bitcopy level, in addition. All that gets
archived, even if potentially it is a working backup, if restored.
I don't see the difference that you find here. I mean, to me,
archive is just multi - year storage. Saying i'm talking backups
not archive makes no sense to me.
We clearly have different paradigms. Projects are begun, finished and archived. We don't care about disk images or months and years. We are project based.

At home, I deal with photography chronologically and also by "keepers" or other more important images I've also toyed with the idea of putting a database together to help identify images and when the image was taken etc.
Right. I don't claim our system is ideal. But it is very easy to
use and we archive a lot of data.
Personally, i consider all data to be potentially useful, and i try
to make the difference between online capacity and total data
stored as minimal as possible.
Yes, but you need to try to balance that with the cost of storing and keeping that data.
How much is a lot, BTW?
A typical project is between 1 and 10GB. We only use CD-R for the smaller projects these days.
There's
also differences in how backup and archive suit working styles - if
you turn around a job quickly, it may make sense to just burn a
disc and put it on the shelf afterwards.
Which is pretty much our situation.
But that makes no sense at
all if you need data accessible to lots of people or work long
project cycles, when if you really have masses of data and no disc
space for them, you can add HSM options to the kit, and let users
pull tape files just as if they were mounted.
Sure, but that is hardly the typical situation for the typical photographer. Maybe a stock agency has those issues.

--
Jay Turberville
http://www.jayandwanda.com
 
And for photographs with EXIF data this would be of no importance.
Or am I missing something?
EXIF often gets stripped. EXIF isn't maintained in a PDF-X file for
pre-press, nor in a complex layered edit. I think you're assuming i
only ever touch RAW files, which i've hinted before isn't the case
Actually, I've given little thought to what you personally are doing. But I know that I maintain a trail to the original image that does contain the EXIF data. Like I said, if maintaining creation date info is important, then CD-R and DVD-R may be insufficient. But I expect that is not the usual case.
Creation dates have never been
important in our operation, but if they matter, then I agree that
this would be a problem.
Depends on how you bill your clients, how big they are (in terms of
burocracy and asking time to time your billing policy) and the sad
fact that law suits are just a fact of life in certain fields.
Yes, for instance if we are doing work for lawyer (forensic animations for instance), then we do a lot more project documentation.
If
you're certain you'll never need all that meta data intact, then
fine. But i've never treated my business with the approach we'll
never need something. Overkill, i guess can only be judged in
hindsight.
In our business, our clients very seldom have any claim to anything other than a finished product. So the risk is very small. And the reality of life is that you can't cover all risks. Trying to do so introduces new risks. The key is trying to be objective in picking a reasonable balance of priorities.
Yes. The optical disc is certainly not the solution for every
situation. My point is/was that neither is tape.
Well, we have equations, rather than certainties, for sure. But
let's say you're a working pro, in a small shop i.e. you have
colleagues take care of production or someone else is the dedicated
photog. Not unusual at all. Someone like this can be generating
terabytes of new work a year, and yet i see and hear reports of
many who think DVD is just fine, as their shelves groan under the
weight. Your DVD-R came with your computer, most likely. My tape
drive and backup software most likely cost more than the typical
computer budget. But it was comparable in price to many a lens, and
takes away one heck of a lot of work.
In business, it is all about the bottom line. If you can save tens or hundreds of hours by spending a few thousand dollars, then you spend it. Period.
LOL! "quite expensive". I think you jest with your understatement.
Expensive is relative. At the time, starting a new business, everything seemed expensive. And some thing just weren't budgetable.
Life is still largely SCSI out there, migrating to Fiber Channel,
or iSCSI, because you don't see FW or USB in the datacenter, which
is what drives the market. Sony do offer a FW drive i believe, and
i'm sure others do similar, but only at the low end of capacity.
That's dumb - datacenter sized working sets are becoming all too
commonplace.
Yes. Especially with media content creation companies.
With me, and no, this ain't the reason for my love of tape (!)
being a klutz keeps me from evaluating any optical media longevity
properly.
Well I just haven't found them to be particularly fragile.
If you used
one of those, you will know what i mean, and if that's what you
used, dang it, i will completely understand you'll use anything but
tape today
We were using a Sony DAT drive. Read times were not particullary quick. The software was fairly primitive (there was much more advanced software on the Amiga OS.) It was workable, but cumbersome.
There's nothing like having a top end machine (at
least dual proc), tons of ram, a serious disc subsystem and a
little tool like X1 (add .com for product site) that makes light of
intense amounts of data. I cannot stress enough that it's great if
you can afford enough on-line or near-line [i.e. slower cheaper
drives] space, and that i think it the exact opposite point of
digital data to archive stuff off-line on a non-human readable
disc.
With SATA RAID controllers being relatively cheap and SATA drive also inexpensive, I have been seriously looking at that option. Upgrade the LAN and put together a fault tolerant data repository. When stuff gets very old, then move it offline. Now in that situation, I'd probably look for a large tape drive. Even if retrieval was harder, it would be worth it for the simplifying the archival of very large amounts of data in one operation.
But when you have oceans of data kicking around your LAN,
then you sure need a more capable system than manually burning
DVDs.
Right. We have data assets that stay online and regualry available. But these assets don't grow quickly. Project assets become less useful extremely quickly after project completion.
I genuinely think that it's a travesty that any data is even
temporarily unreadable, and so the though that, to boot, the dang
disc is going to "rot" on me, beyond convenient observation, gets
my sense of irony piqued.
A couple minutes to get access isn't such a travesty. And like I said, I haven't seen any disc rot yet. :)

g'nite.

--
Jay Turberville
http://www.jayandwanda.com
 
For around $15 I can buy an IDE/USB2.0 cable and AC to DC powersupply to connect a hard drive to my pc for backing up. Since it only needs to run for backup purposes, this would practically nullify any concerns about the life and MTBF of the mechanical parts which are designed to run for years. Normally, the drive would just be sitting on my bookshelf. At 50 cents per gig, it's difficult to beat the cost and convenience of hard drives.
 
"These are gold-plated (no telltale blue dye cast) and are supposed to have lifetimes of 75 or more years".

Where is the proof that they last that long? These figures are estimates based on laboratory tests. Cds have not been around for 75 years.

Magnetic media has always been the best proven way to store data long-term - and I suspect that it always will be.
 
Fire would be a problem. Theft too. User error (deleting a directory by accident!). That's why offsite is nice.
 
Fire would be a problem. Theft too. User error (deleting a
directory by accident!). That's why offsite is nice.
So you have multiple cheap disks and keep some off site, just like tape. A $15 USB cable/IDE adapter is a lot cheaper than a high capacity tape drive.

I'm sure that user error can louse tapes up also, without using much imagination.

Wayne Larmon
 
If memory serves me correctly, Ken Rockwell (on his web site) asserted error correction on DVDs is not as reliable as error correction on CDs.

Have no idea if this is true.
--
thezero
 
Store images on good quality media. TDK, verbatim, etc

Keep media in a cool place away from sunlight...

Also back up onto another hard drive.....

Not had a single problem!
 
I dunno, but over the past five years I lost several CDs and DVDs, but all which went bad either had stick on lables or were chewed up by a puppy. I also once lost many years of B&W negs and chromes in a flood. 'Feces occurs' regardless of storage technique. I'll stick with the CDs and DVDs.

--
Old fart, retired pro shooter
 
I think the main problem with offline hard drives is shipping them. Think how to pronounce UPS now :-)

Also the package will be attractive to theft, just because of it being a high capacity disc drive.
Fire would be a problem.
Erm, safety deposit companies?
Theft too.
See above
User error (deleting a
directory by accident!).
Now several companies sell read - only hardware SCSI and FW adapters for forensic duties. They're afforable enough you could hack up one in the patch of your connexion and epoxy any write enable switch. Inelegant, but effective.

Alternatively, deny user write permissions, but this still is not being manitained at the FS level adequately. However something like EFS plus hardware keys to encrypted data would be pretty effective on a NT network.

Still feels like a kludge though.

Better yet, run VMS as your file server, which will solve the above. But that's asking a bit much of a small shop with no experience in VMS. Asking an awful lot, actually, but i digress.
That's why offsite is nice.
Yup! And ideally if this were economically feasible, offsite backup over WAN. Little beats the physical security or environmental control of a carrier hotel. This is what bugs me about ISPs, expecially now ADSL2 and SDSL are widely practicable - the lack of billing and bandwidth flexibility. If my ISP could provision me uncontended windows of upstream once a week, instead of charging rental rates for the whole time (at considerable cost) i'd run a WAN backup for various data. I just don't think the DSLAMs installed or to be installed support this, even though MPLS (which feels like old hat now) should be capable of assigning QOS end to end . . .
 
bwhahaha.

No, I won't comment on Ken R here but DVD error correction (as the format is more recent and the density is much higher) is way more advanced than CD error correction. (Lots of technical info on the internet. Just google)

--
--Reinhard
 

Keyboard shortcuts

Back
Top