Digital media is a double-edged sword. Digital data itself can be duplicated an unlimited number of times without any generational loss – meaning it can theoretically last forever. But digital storage on physical media is subject to failure – and that failure can render the data inaccessible. In other words, archivists (including you) have to transfer data before the media fails.
And we’re already entering an age when one of the most popular formats is reaching the start point for common failures.
A report by Tedium (republished by Motherboard) demonstrates one of the most alarming failures. Some media, evidently using faulty dyes, can fail in under ten years, via something unpleasantly dubbed “disc rot.”
At issue is the fact that optical media uses a combination of different chemicals and manufacturing processes. That means that while the data storage and basic manufacturing of a disc are standardized, the particulars of how it was fabricated aren’t. Particular makes and particular batches are subject to different aging characteristics. And with some of these failures occurring in less than ten years, we’re finding out just how susceptible discs are outside of lab test conditions.
In short, these flaws appear to be fairly widespread.
That just deals with a particular early failure, however. In general, CD formats start to fail in significant numbers inside 20 years – on average, not just including these rot-prone flawed media.
What’s tough about this is that the lifespan can be really unpredictable. Before you dismiss the CD as a flawed storage format, many discs do reach a ridiculously long lifespan. The problem is really the variability.
To get an accurate picture, you need to study a big collection of different discs from a lot of different sources. Enter the United States of America’s Library of Congress, who have just that. In 2009, they did an exhaustive study of disc life in their collection – and found at least some discs will be usable in the 28th Century (seriously). The research is pretty scientific, but here’s an important conclusion:
The mean lifetime for the disc population as a whole was calculated to be 776 years for the discs used in this study. As demonstrated in the histograms in Figures 18 and 19, that lifetime could be less than 25 years for some discs, up to 500 years for others, and even longer.
COMPACT DISC SERVICE LIFE: AN INVESTIGATION OF THE ESTIMATED SERVICE LIFE OF PRERECORDED COMPACT DISCS (CD-ROM) [PDF, Preservation Directorate, Library of Congress]
Other research found failures around 20-25 years. That explains why we’re hearing about this problem round about now – the CD format was unveiled in 1982, and by the 90s we all had a variety of optical disc storage to deal with.
There are two takeaways – one is obviously duplicating vital information on a regular basis. The other, perhaps more important solution, is better storage. The Library of Congress found that even CDs at the low end of life expectancy (like 25 years) could improve that lifespan by twenty five times if stored at 5 degrees C (41 degrees F) and 30% relative humidity. So, better put that vital collectors’ DVD in the fridge, it seems. That means instead of your year-2000 disc failing in 2025, it fails in the 27th Century. (I hear we have warp-capable starships long before then.)
But anyone using discs for backup and storage on their own should take this even more seriously, because numerous studies find that writeable CD media – as we purchased with optical drives in the 90s – are even more susceptible to failure.
There are many other issues around CDs, including scratch and wear. See this nice overview, with some do’s and don’ts:
CD and DVD Lifetime and Maintenance [wow, 2007 Blogger!]
CDs Are Not Forever: The Truth About CD/DVD Longevity, “Mold” & “Rot” [makeuseof]
I’ve seen some people comment that this is a reason to use vinyl. But that misses the point. For music, analog storage media still are at a disadvantage. They still suffer from physical degradation, and reasonably quickly. For digital media, hard disc failures are even more frequent than CDs (think under three years in many cases), and network-based storage with backups more or less eliminates the problems of aging generally, in that data is always kept in at least two places.
The failure of CDs seems to be more of a case of marketing getting divorced from science. We’re never free of the constraints of the physical world. As an archivist will tell you, we have to simple adapt – from duplication to climate control.
But I’d say generally, with network-connected storage and automation, digital preservation is now better than ever. The failure point is humans; if you think about this stuff, you can solve it.
But ideally digital preservation should not use optical discs because of how unpredictable they are, and because these failures can prevent data access entirely. In particular, writeable discs are very much prone to failure, and if you have data on them, you probably want to consider retrieving that data and putting it on another medium sooner rather than later.
And it’s about time to start heeding these calls, given that they’ve been widespread since the turn of this century. For instance:
IBM expert warns of short life span for burned CDs [PC World, 2006]
The solution then? Tape.