Volatile Mediums

I’ve recently been thinking a lot about storage mediums [1] – especially in the long-term.

Technology has made a lot of progress. Digital storage mediums started out only being able to store 224KB on a tape drive for an average lifetime of up to 30 years. Now, we can store terrabytes of data on hard drives and solid-state drives. However, no one ever really answered the question about long-term storage.

(Note: the following is based off an assumption that the storage medium is only being used to make backups or archive data. The device itself could be unplugged and stored when no backup is in progress.)

Even though theoretically hard drives could store data for 20+ years, random bit flips, drive failure, etc. all make hard drives too volatile of an option. As always, of course redundancy takes away some of these issues.

SSDs are in an even worse position: they cost significantly more than hard drives per TB right now, and last I heard, there were still issues with bit fade when unpowered.

CD/DVD is sounding a lot better, but there are some serious issues here too. Variable quality directly impacts the storage lifetime. Physically storing the discs is a lot more risky since the disc itself doesn’t have as much built-in protection as a hard drive or SSD has. You’ll need a much larger quantity to store the terrabytes of data that you can easily dump on one hard drive. And finally, life expectancy is still fairly low – while manufacturers of recordable discs (the ‘R’ in CD-R, DVD-R, etc.) claim life expectancies of 100-200 (!) years under optimal conditions, others are slightly more conservative, giving an estimate of 30 years. Oh, and remember how I mentioned this is for recordable discs? That means they’re single write. The random access (RW - CD-RW, DVD-RW, etc.) discs have even lower life expectancies.

All in all, humanity has not gotten very far with the digital storage medium. All of these life expectancies have an inconsequential variance when we zoom out to the century view of history.

[1] And no, I’m not talking about the kind you pay to see your dead great-great-aunt to figure out if you’re actually related to George Washington.

This is intended to be the beginning of a learning series/personal study on the issues surrounding information preservation, digital permanence, and their related issues.