A friend recently called to tell me she was unable to attend a meeting we were scheduled for. A nasty virus (biological) had her flattened.
No big deal, said I. I asked her to give me a copy of the information she was supposed to present, and I would handle it.
“”Oops,”” said she “”it’s in
my e-mail, and the server is down. I don’t have a hard copy.””
Sound familiar?
Downed servers are increasingly replacing hungry dogs as excuses for missing “”homework.”” Data exists only in digital form, and if the system it lives on is inaccessible or incapacitated, we’re out of luck.
Think back five years, and remember the software you were using to create those critical documents that could mean life or death to your business. Does it still exist? Do you still have a copy? Come to think of it, if a regulator demanded the files, or you needed to retrieve some old but crucial information, would they still even be readable?
A packrat like me would cheerfully say yes, then closely follow the statement with a heartfelt “”darn.”” I certainly have a copy of the data and software in question – but it’s on a five-inch floppy disk, and there isn’t a five-inch drive in the building. Or, it’s on a tape medium for which the drive has long since become landfill.
Even recent files, if written to optical media, are in doubt. According to a recent Associated Press story, so-called “”CD rot”” plagues some compact disks and DVDs, oxidizing the aluminum disk and destroying its contents unless the disk is treated with kid gloves.
The story quoted a specialist at the National Institute of Standards and Technology, who noted that manufacturers change materials and manufacturing techniques regularly enough that it’s impossible to know what you’re getting — a disk that was great six months ago may be anything but today. And there’s no standardized testing that can give users a clue of what to expect in disk life.
How comforting. Now we’re faced with two problems in our archives: are the bits even readable, and do we have the software to interpret them.
It makes one long for the days of quill pens and parchment. At least you can easily assess when ink begins to fade.
At a recent IBM storage briefing, IBM Fellow Jai Menon presented research that partly addresses the issues. The company is developing a way to include a reader in data files (they call it the Universal Virtual Computer), so regardless of the vintage of the software, files can be interpreted. Its goal is to allow today’s data to be readable in 100 years (assuming the media holds up that long, and I’m sure they’re working on that too. In fact, IBM has “”self-preserving”” magnetic tape today.).
And just so we’ll have enough space available to stash 100 years’ worth of data, he also showed off a sinister-looking black cube that strongly resembled a Borg ship from Star Trek Voyager.
That cube, made up of 27 smaller cubes called bricks, each containing a full disk drive with controller, contained no cables — drives were connected by capacitive coupling — and offered 32TB of storage (enough to store the contents of the U.S. Library of Congress), with sufficient redundancy that the cube could be unmaintained for five years with no ill effects.
Menon said that one petabyte of data, which would take up about 800 square feet of traditional disk storage, could potentially be stored in a 120 square foot cube. Instead of adding standalone drives to increase storage capacity, we’d be able to pop another brick or two onto the cube, with zero incremental floor space — just the thing for corporate data centres where a “”small”” SAN is a mere 30TB and storage requirements grow faster than this spring’s dandelions.
Now, if the server would just stay up, we’d be all set.
