How likely is dvd data loss over time?

vbimport

#1

Hi

Has someone any statistical data on how likely a burned dvd will loose data over time?

I have heard and read several stories about dvds not being readable after a few months. But has any magazine made some statistical relevant test or something similar?

I know that this actually depends on the producer of the dvd, the quality of the specific batch of dvds, the burner, the burning speed, the weather conditions, probably even the moon phases.

And furthermore, dvd-rs are not around for that long, so long-time tests are rare.

I know that I cannot expect very reliable information. So I am looking for unreliable data :slight_smile:

I just want to make a very rough cost/risk estimation. I want to archive certain data and evaluate how likely a loss of that data is. This is not irreplacable data, so the cost for archival is important too.

I also include par2 parity files to increase redundancy. Has anyone else any experience with this? or does it not matter since the whole dvd becomes unreadable at once?

As a concrete example, I am verifying all my DVDs just after burning on another pc, and i burn roughly 5-10% par2 parity data on each dvd. I use a nec2500a and burn at 4x or 2x. How sure can I be that in 2 years 95% of my data is still readable? (by that time I can easily burn all my current data on one of the new x-ray discs :slight_smile: )


#2

I don’t know of any rigorous statistical data that would help you on this. I don’t think that even a rough cost/risk estimation can be made.

The only things we know for sure is that redundancy is good. DVD-R discs only cost a dollar or two now so you should try backing up your data to say four different discs and then keep the discs in different physical buildings in case there is a fire.

And then every 6 months or so you can make new copies of the data just to be safe. But I can’t give you any kind of probability distribution function for the reliability of a DVD-R disc.


#3

Here’s what I’m doing. I burned my first seven disks only a month ago. I left all the original files on my hard disk (some 29 GB). Every month, I will run a comparison and count any bits and bytes that have gone bad. I’m doing this on two different media: TDK DVD+R (RICOHJPNR01) and Ritek DVD-R (RIT3KG04).

If others would do the same, we could build a results library, perhaps even start a Longevity subforum.

Jon


#4

Originally posted by Jucius_Maximus
I don’t know of any rigorous statistical data that would help you on this. I don’t think that even a rough cost/risk estimation can be made.

Ah well, I was afraid that would be the case. Thanks nevertheless

I guess I just have to play save.
Does anyone know what would actually be saver:
2 very cheap media with identical content, or 1 expensive one?

Originally posted by jonabsoul
Here’s what I’m doing. I burned my first seven disks only a month ago. I left all the original files on my hard disk (some 29 GB). Every month, I will run a comparison and count any bits and bytes that have gone bad. I’m doing this on two different media: TDK DVD+R (RICOHJPNR01) and Ritek DVD-R (RIT3KG04).

I dont have nearly enough harddisk space to try that :slight_smile:
Here is what I do: I burn a par2 checksum file on every disk.
That way I can verify quickly if the data is still readable and correct. I cannot actually tell how many bytes are damaged, but I can tell how many “blocks” are damaged (blocksize is userdefined upon creating the checksum file)
It has the added bonus that I can verify the files on any pc

Maybe we can really start such a longevity database here.


#5

Originally posted by Zem
[b]I guess I just have to play save.
Does anyone know what would actually be saver:
2 very cheap media with identical content, or 1 expensive one?

Maybe we can really start such a longevity database here. [/B]

If I had to choose, I’d burn only on single Taiyo-Yuden discs as opposed to multiples of low quality discs.

As to a longevity database … I don’t think it’s going to happen. Look how long CD-Rs have been around and there still is no longevity database for those.

And here’s some food for thought. I was just going through some of my old CD-R discs. I have CMC Magnetics CD-Rs that I burned in 1997 on a SCSI Matsushita 4X burner and they are still 100% readable. (Note: CMC CD-Rs are among the worst CD-R discs available.) Naturally my Taiyo-Yuden discs from 1997 are also 100% readable. That’s just something for you to think about.


#6

You may want to have a read of this thread before you decide on your back-up strategy:

http://club.cdfreaks.com/showthread.php?s=&threadid=90921

A worst case scenario i think but you can see such things can happen even when you think your being careful.


#7

Originally posted by Jucius_Maximus
As to a longevity database … I don’t think it’s going to happen. Look how long CD-Rs have been around and there still is no longevity database for those.

Yes, but CDs currently dont seem to have much problems with longevity.
But I can see another problem: Any data we might collect might soon be outdated.
Well, maybe C’t will start a big test with a simulated speed aging process. it would fit them :slight_smile:

Originally posted by TL0
You may want to have a read of this thread before you decide on your back-up strategy:

http://club.cdfreaks.com/showthread.php?s=&threadid=90921

A worst case scenario i think but you can see such things can happen even when you think your being careful.

Yes, threads like that one were the reason I started this thread.

I just cannot estimate what the chances of this happening to me are. And if I simply assume the worst, I arrive at the conclusion that I should not use dvds at all :bigsmile:


#8

Originally posted by Jucius_Maximus
[B]If I had to choose, I’d burn only on single Taiyo-Yuden discs as opposed to multiples of low quality discs.

As to a longevity database … I don’t think it’s going to happen. Look how long CD-Rs have been around and there still is no longevity database for those.

And here’s some food for thought. I was just going through some of my old CD-R discs. I have CMC Magnetics CD-Rs that I burned in 1997 on a SCSI Matsushita 4X burner and they are still 100% readable. (Note: CMC CD-Rs are among the worst CD-R discs available.) Naturally my Taiyo-Yuden discs from 1997 are also 100% readable. That’s just something for you to think about. [/B]

The CMC discs I burned 2 years ago are still readable. BUT, they are already starting to have reading problems. As in, halfway through the disc, I hear my drive slowing, spinning up, slowing down, etc. The data is retreived just fine. Did a md5sum, and the checksums are all okay. But there are signs of degradation, as they didn’t have these read problems back when they were first burned.