Will your data survive as long as you will? A look at the best media and methods for long-term backup
You're desperate for a certain piece of information. Perhaps you need a customer's records from 1977, or your boss has asked you to review the company's profit-and-loss statements for the past 20 years. OK, the information was all properly stored, but it's on paper tapes. How are you going to read those long spools of paper with holes in them? And how is the next generation going to read the floppy disks, digital tapes, and CD-ROMs you're using for storage today?
Every company knows that for protection against computer crashes, disk failures, viruses, fire, and theft, it needs a daily backup system--though many wait until they've had a disaster before they implement one. But what about information you might want to retrieve in 5, 10, or even 50 years, like tax records, personnel data, customer lists, or historical profit-and-loss statements? Few companies have a strategy for long-term backup that takes into account not only the reliability of the medium but also the problem of obsolescence.
The first question is which medium to use: digital tapes, magnetic cartridges, magnetic disks, optical discs, or recordable CD-ROMs (designated by the somewhat backward abbreviation CD-Rs). What do we know about them? First, digital tape--any kind of tape--is unreliable for long-term storage. Tape stretches, sticks to itself, and tangles. And the more you store on it, the more problematic it is. If you've dutifully been using a tape backup system (which used to be recommended for long-term storage), it's time to upgrade. I would also rule out erasable backup media--magnetic disks and rewritable optical discs. If your data are important enough to put into long-term storage, you don't want to risk their being erased.
As the price of CD recorders has dropped (it's now below $600), the long-term backup medium of choice has become the CD-R. Today most computers come with CD-ROM players, and a recordable disc costs just $6 for 650-MB capacity, enough to store the entire Encyclopaedia Britannica. CD-Rs can be written on only once, and they can't be erased, an excellent feature for long-term backup.
How long do CD-Rs last? No one can really say for sure. CD-R makers, basing their estimates on accelerated tests, talk about 70 to 100 years. Makers of all kinds of backup media generally agree that CD-Rs have the longest life span available today.
Long before 70 years pass, however, you're going to have to deal with a critical issue in long-term storage: obsolescence. Your backup strategy has to include access to the software and hardware you're going to need to retrieve the data. Whenever you significantly upgrade software or an operating system, think about converting your long-term backup files. Because most programs can read only the previous version or two (for instance, a Version 6 program may read Version 5 and 6 files but not Version 4 or earlier), you may have to open and save each file on several successive versions--a tedious process. And you may need auxiliary files--fonts, color-calibration tables, and so on--to fully recover the main file. That means your backup will have to store not only your documents but also the complete system files.
Or if appearance isn't important, you can save the text as an ASCII file. There will always be programs that can read ASCII. Hypertext markup language (HTML), used for World Wide Web pages, is probably a better long-term choice, but most programs can't save a file as HTML...yet.
Then there's the matter of hardware. Some kind of optical drive that can read CD-Rs should be common for at least 10 years, but the computer of 50 years from now is unlikely to use any storage medium with moving parts. So whenever you upgrade your hardware, also think about transferring your backup data to an upgraded medium. In about three years, you might convert all your CD-R archives to the new, higher-density digital-video-disc (DVD) format. And you might keep one old computer with old software around simply to read old files. Otherwise, you may have to haunt the flea markets to find a computer that can open them.
There's only one storage medium--often overlooked--that I can guarantee you'll be able to read 50 years from now. It takes up a lot of space, and you can't search it easily, but it doesn't require any specific version of software or hardware--or, for that matter, any electronic equipment. It's acid-free paper. Antiquated as paper may seem, it's worth considering for tax returns and other short annual summaries. Someday one of those may be just the document you desperately need.
Cary Lu was formerly technology editor of Inc. He backs up his documents every day and makes a CD-R every six months.