June 25, 2004

CDs & Long-term Data Storage Tips

Courtesy of Dave Rakowski, who posted this link today on the ABA-LawTech listserv:

"Long Term Data Storage on CD-R Discs (how to store your data for a long, long time)", while containing some dated references, covers the controversy surrounding just how long one can expect data stored on burned CDs to survive. In this regard, I liked how it colorfully mentioned "Don't Screw Up the Dye Layer". Most people take great care to protect the clear plastic bottom of the CD, the one that the laser reads, and there's even handy devices like the Skip Doctor to help buff out scratches on that side to bring even badly scratched CDs and DVDs back from the dead.

However, what some people may not realize is that your disc's data is stored literally right behind the label, in what's known as the dye layer. So while there's a lot of clear plastic between the laser side and the data, your data is only a hair's breadth away from being rendered partially or completely unreadable by even a small scratch on the label side. The good news is that this makes it very easy to dispose of these discs securely by taking a sharp object like a flat head screwdriver to the label side with a minimum and quick effort (which I've done in well under a minute). The bad news is that once the dye layer is damaged, say hasta la vista to the affected data.

Perhaps even more helpful, the article discusses why some file formats are better than others -- not for storage, but for long term accessibility due to the fact that technology and formats change so frequently. For example, the included file type table suggested various MPEG versions as preferred over more proprietary formats such as AVI, Quicktime, and RealVideo, which may or may not be around in say, five, ten, or more years. Another great suggestion: "To maintain maximum flexibility for your archived data, you might want to store two copies on each disc, one copy in an industry-standard format, and another in the application-specific format of your choice." Lastly, it pays to check your data collection every few years, to make sure they're still readable and to transfer them to new media types as the old ones become obsolete (not to mention the hardware required to read them).

Topic(s):   Electronic Discovery  |  Other Musings
Posted by Jeff Beard