Write-once Media Validation

Why is write-once media special?

Unlike image files that are stored on hard disk, files on write-once media can't be altered once the disc has been written. This means that any change to even one bit of file data indicates some kind of media degradation. Therefore it's possible to make a checksum of entire discs to determine the exact condition of your backup storage.

Doesn't my disc-burning hardware do this for me?

When you burn a disc, the software does some level of validation of the process. This might be a comparison of the file size which is a basic indication of the burn. The software might even do a bit-for-bit comparison of the original and the copy, which can tell you that the files have been copied perfectly.

These processes however, will only work as long as the original never gets touched. If a keyword is added to the original file, then the copy will no longer match. At that point, you won't know if the mismatch is because the original has been changed or because the media is starting to degrade.

Write-once validation workflow

In order to set up a solid validation workflow, you'll want to make sure to do things in the right order. You'll also need to preserve the checksum somehow in order to check on the status of the disc in the future.

• Make a hash of the original data. Once you are ready to burn your disc, you need to send the data through a program that can create a validation hash. Ideally, this would be done to the source data on the hard drive, rather than the data that has been written to the optical disc.

  Of course, if you have not done this for older discs, and want to implement validation for these discs without having to reburn them, you can make the hash from existing discs. This will tell you if there is any degradation of the storage media at any point after the hash is made.

• Compare the hash to the write-once data. Since we're going through the trouble of making the hash from the original data, we might as well check the write-once disc as soon as its been created. This lets us know the burn has been accomplished with no errors.

• Save the hashes. We'll look at a program called ImageVerifier that can work for Mac and PC. ImageVerifier stores the hash information in a database in the users folder of your computer. This will be valuable information for years to come, so you'll want to make sure you are making good backups of your boot drive, in case of drive failure or other problem.

  A PC-only program called ExactFile lets you hash the original data, and lace the hashes on the optical disc itself, so that the validation information is even more durable and portable.

• Do a periodic validation. You'll want to periodically validate your discs to make sure that everything's fine. If you have a lot of optical discs, it's probably best to do a random sampling, making sure to choose discs from different eras, and to choose discs from different manufacturers if there is variation in your collection.

What to do if you spot any problems

It's likely that, eventually, you'll come across some kind of checksum mismatch. At that point, you'll want to track down the corruption and correct it. ImageVerifier will tell you what files don't match. Try opening them in your image editing application and see if they appear fine (there can be small mismatches that are not visible in the file). If the files don't open, you'll want to make a list and find these files in your primary storage or other backup storage and confirm that these copies still have integrity. If they are fine in other locations, it's time to burn new discs.

If you find any data mismatches, you'll also want to check out any other disks that are made with the same brand of disk. You also might want to check any other disks that might have been burned in the same computer, since substandard disk burning can result in data decay over time.

Disk size and number

As you can see, this process can take a long time to run through. If you've got hundreds of discs, validating each one can be a second job. For this reason, we suggest that you consider using the largest discs that are practical. While DVD discs are quite common, 4.7 GB is not very big, considering the volume today's digital photographers produce. We urge you to consider using Blu-ray disc as an optical disc since it can hold 23 GB per single-layer disc.

How about hard drive validation?

As we said at the top of this page, this technique is best for use with write-once media such as optical disc. People tend to update files on hard drive, even for backup files. If you have a hard drive-only archive, you may want to consider limiting one copy of the backup to an Additive-only workflow. This copy of the data, like your optical disc copy, should be considered as the "disaster recovery" copy. It may not be up-to-date, but at least all the images will be there.

If one backup copy of your images is only added to, and never updated, then you can take advantage of MD-5 checksum validation without having to go to the trouble of burning optical discs. Of course, if you ever update the backups by adding metadata, image adjustments, or any change, you will invalidate the hash and, and no longer have confirmation that the files remain uncorrupted.