Storage

by Henry Farrell on January 28, 2007

According to this “article”:http://privacy.cs.cmu.edu/people/sweeney/explosion2.pdf by Latanya Sweeney, one million hard disks with a total storage space of 90 terabytes were sold in 1983. My computer alone, purchased in late 2006, has one terabyte of storage space mounted in 2 RAID drives (I use them to back each other up). If my understanding that storage technology has yet to hit a brick wall is correct, it seems likely that a not-especially rich US consumer will have as much information storage capacity available to her as was available to the entire world in the early 1980s a few years down the line, if she wants it. (Sweeney’s article is really about the broader privacy issues that arise because of this expansion in the ability to store, and indeed to gather, personal information, but this figure hit me between the eyes when I read it).

{ 29 comments }

1

ogged 01.28.07 at 10:42 pm

What are you doing with a terabyte of storage, monitoring the NSA?

2

John Quiggin 01.28.07 at 10:55 pm

Taking a slightly longer time interval, the Kennedy Space Center exhibit of the mission control room for the Apollo program advises visitors that “your watch probably contains more computing power than all the machines in this room” (my emphasis added).

3

Henry 01.28.07 at 11:13 pm

It’s in a redundant array, so it is effectively 500 gig backed up very safely (I’ve had hard disk crashes in the past and would prefer not to experience them in the future if I can avoid it; the crucial stuff I am backing up elsewhere too). But I’m also going to be doing some video work for my class next fall, which gobbles up a lot of space far more quickly than you might imagine …

The Apollo thing I had heard before, and didn’t find quite as striking; perhaps because I have clear memories of using a computer in 1983, and don’t have memories of so doing in the late 1960s, by reason of not having been around. All that said, I’m not sure that my PC is _that much_ more wonderful than my old ZX Spectrum 48k-plus-tape-recorder combo.

4

robotslave 01.28.07 at 11:39 pm

“it seems likely that a not-especially rich US consumer will have as much information storage capacity available to her as was available to the entire world in the early 1980s”

This is true in a literal sense, but ignores the fact that consumers in the early 80s were using video tapes, audio cassettes, arcades, and boxes of photos for the sort of “information storage” that tends to comprise the better part of the “data” on hard drives containing more than 100GB or so of the stuff these days. And I suspect I’m being rather generous with that arbitrary estimate, too…

5

Chuchundra 01.29.07 at 12:03 am

Actually, I expect that in 1983 the majority of digitally stored data was kept on 9 track tapes. Hard drives were very expensive and only the most well off PC owners could afford a giant, 5MB, external drive for their PC. Most of that 1983 hard drive storage was probably used in data centers and such, not connected to individual user machines.

6

nick s 01.29.07 at 12:05 am

obGeek: RAID is not a backup solution. It’s a protection against isolated hardware failure.

On the more general point, we’ve reached a stage where it’s more fuss than it’s worth to throw away bits of our digital lives (apart from those times when there’s a crash or a virus attack or you’re migrating OS.) The problem then becomes one of organisation.

My first PC (1997) had a 2Gb hard drive. I’ve now got about 700Gb of stuff in one box, backup external drives, piles of burned CDs and DVDs. It’s all a little scary.

On the less-is-more thing: it scares me even more that you can fit the entirety of software developed for the ZX Spectrum on a DVD, and that the scanned images of various magazines and manuals take up more space in the online archive. The restrictions of the Spectrum environment — relatively crappy graphics and sound — seem to have encouraged an efficiency and creativity that served programmers well when they moved on to other platforms.

Lastly, robotslave is right to point out that storage in 1983 was, for the most part, in other media (not least the repurposed media used for computer storage). The ‘Winchesters’, as hard drives were called back then, were LP-wide platters that cost a ridiculous amount of money.

7

John Quiggin 01.29.07 at 12:27 am

I can remember buying my first Mac in 1984 and thinking that I wouldn’t need a whole box (10) of 400k floppy disks to store my files. For a week or two, I didn’t, either.

8

srv 01.29.07 at 1:01 am

A problem looking for solution, or a solution looking for a problem?

Here’s just the video content a kid born today can look forward to:
1) sonargrams
2) 24×7 monitoring of maternity ward
3) family/baby videos
4) baby-sitter, back yard, pool cam(s)
5) 8×5 day-care cam
6) 8×5 elementary school cam

A child born this year in a metropolitan area can probably expect to be under various forms of direct survellience for 20% of their lives. That’s 5-75 GigaBytes per day, depending on the refresh rate.

9

bad Jim 01.29.07 at 3:39 am

My camera has more storage than the computer I used ten years ago. It may well be that my phone does, too. I don’t care. The source for the carefully crafted microcontroller code I wrote back then fit comfortably onto a floppy, and the machines for whom I crafted them woke up instantly, whereas my phone and camera and stereo take an eternity (nearly a minute) to boot themselves.

Whether starting Windows or Linux, my computer takes approximately as long to get started as my old twin-floppy Osborne. The latest machines are all much more capable, and oh so elegant in presentation, but in too many ways not as quick as they might be.

10

Moz 01.29.07 at 5:09 am

Yeah, the camera thing is pretty bad. Even not taking many photos I’m using nearly 1GB/mo just for stills, and if I let the footage from my helmet cam pile up that uses nearly 1GB/day (I video my ride to and from work just in case I need the footage after a crash).

I just upgraded my disks – from 4x300GB (RAID5) to 4x750GB (~2TB usable). I regret not buying the 8 port RAID card now, it would let me use cheaper disks. 8x320GB disks costs about 50c/GB, the 750GB disks are more like 70c/GB.

But backing that up is a nightmare. Incremental backups to DVD are not so hard to do (I have a system), but the thought of restoring it gives me nightmares. I have ~300 DVDs, so assuming I can copy and change one every 10 minutes (per DVD drive), that’s 50 hours of copying. Not fun.

11

dave heasman 01.29.07 at 5:59 am

I worked for a data cantre (or “bureau”) in 1973 and we used a rack of IBM 2311 disc drives. We had 6 drives. Each held 7.25 Mb. When we upgraded to the 2314 drives we had to change the code for all the jobs.
The “Winchester” drive was so-called because its actual model number was “3220”.

Discs were only used for transient data, sort work areas etc. Permanent data was on 9-track tape. And there wasn’t that much of it – you don’t need much if you’re just doing accounts and payrolls, it’s when you start modelling, using relational databases with huge indexes etc that the space requirements ramp up.

I get into trouble at home because I only back up monthly, and that to an external hard disc. My wife works for a small housing association and they back up daily, and I tell her that they have a lot less data, and it’s a lot more volatile, but I can’t get her to believe me.

12

abb1 01.29.07 at 7:22 am

Ah yes, IBM mainframe with 128Kb RAM, a couple of 7.25Mb drives, tapes, boxes of punch-cards and two dozen people to service it all. UNIT=SYSDA. Those were the days, my friend.

13

K R Hasan 01.29.07 at 7:50 am

I got my first electronic “programmable” calculator (a Sinclair) in the early seventies. Around the same time our class in college was proudly introduced to a Hewlett Packard model that was rolled in on a trolley, which could actually invert a 4 by 4 matrix! That was still the punch card era so far as computers were concerned.
We’ve made a lot of progress, but not quite as much as the increase in available RAM or storage capacity might suggest.

14

Cranky Observer 01.29.07 at 8:05 am

Of course, in 1983 a one-page memo in Professional Write format took up about 12k. In 2006 the exact same memo in Microsoft Word 2003 format takes up 1-2 MB.

Hardware giveth, Microsoft taketh away.

Cranky

15

John Quiggin 01.29.07 at 9:00 am

Winchester was 3030, wasn’t it?

16

tom brandt 01.29.07 at 9:14 am

Yes, the Winchester’s model number was 3030.

I remember specing out a minicomputer (remember them?) in 1978. Main memory cost $1/byte. That means my PC has $1 billion worth of memory on it in 1978 terms.

17

Slocum 01.29.07 at 9:16 am

This is true in a literal sense, but ignores the fact that consumers in the early 80s were using video tapes, audio cassettes, arcades, and boxes of photos for the sort of “information storage” that tends to comprise the better part of the “data” on hard drives containing more than 100GB or so of the stuff these days.

Yes, exactly — the vast majority of that storage will be used for digital image, audio, and video data that used to be stored in analog form. And the amount of information stored in the old forms was not necessarily much lower (when you consider analog video tapes vs DV and 35mm film negatives vs digital camera images).

And I question that this explosion in capacity has many implications for privacy. The kind of data we worry about with respect to privacy (names, credit card numbers, financial transactions, medical histories, etc) are simple text and, therefore, very compact to store.

18

tps12 01.29.07 at 9:48 am

Keep up, all you grandpas…we measure storage in “songs” now.

19

Tim McG 01.29.07 at 12:29 pm

And “song” of course, means what those of us born in the seventies call “music video.”

20

Cranky Observer 01.29.07 at 12:40 pm

> Keep up, all you grandpas…we measure
> storage in “songs” now.

Good point – I haven’t actually seen a 1-page memo in 10 years.

Cranky

21

nick s 01.29.07 at 3:11 pm

And I question that this explosion in capacity has many implications for privacy. The kind of data we worry about with respect to privacy (names, credit card numbers, financial transactions, medical histories, etc) are simple text and, therefore, very compact to store.

That’s a very selective definition. Discrete low-level personal information has fairly discrete privacy value, but it’s the stuff that doesn’t come with a keep-secret warning that’s much more troubling, particularly given the blurring of public and private online.

If someone gets hold of my credit card number, it’s a very real pain in the arse, but if someone nabs my photos, that’s a personal intrusion. (Consider the emotional impact of robberies: the violation of sentimental items hits much harder than the loss of cash and prizes.)

[On the one hand, digital photography makes it less likely that a mother will get a visit from the local plod because someone at Boots took offence at the bathtime-playtime photos of her children. That’s a gain of privacy. On the other hand… etc.]

22

tom s. 01.29.07 at 3:13 pm

“If my understanding that storage technology has yet to hit a brick wall is correct”

It is correct, if Fujitsu are to be trusted.

“Fujitsu has developed a technique it claims will allow the company to produce hard drives with an areal data density of 1Tb per square inch – almost seven times the density of today’s latest perpendicular-recording hard drives.”

http://www.reghardware.co.uk/2007/01/25/fujitsu_1tbpsi_hdd_breakthrough/

23

Slocum 01.29.07 at 5:16 pm

If someone gets hold of my credit card number, it’s a very real pain in the arse, but if someone nabs my photos, that’s a personal intrusion.

I suppose — but there’s really very little motive for somebody to nab your family snapshots. Unlike account numbers, credit histories, and medical records, personal photographs generally have zero value to anybody other than the owners.

(Consider the emotional impact of robberies: the violation of sentimental items hits much harder than the loss of cash and prizes.)

Yes — losing irreplaceable items has a special impact, but when we’re talking about digital images, there generally is no single copy that can be stolen. If thieves made off with copies of my vacation photos, I wouldn’t much care, because I’d still have copies myself.

24

nick s 01.29.07 at 11:33 pm

personal photographs generally have zero value to anybody other than the owners.

That’s my point: when they do, they do in big ways that don’t fit the discrete private/public model. There’s a reason why the redtops in Britain employed the services of someone famous for searching through bins for scraps of personal information.

Consider the swarm effect when a group of people on the internets want to find out about someone based upon a few scraps of public data, which can then lead to online spoors.

Even without the bits to which you ascribe value — in essence, the stuff of financial life — my guess is that the contents of one’s hard drive can be used in very uncomfortable ways.

25

MQ 01.30.07 at 1:36 am

#10, you’ve got to be kidding.

26

Bill Gardner 01.30.07 at 9:57 am

I’m stunned that some of you have multiple terabytes of disc based storage. That’s more or less the size of our hospital’s data warehouse (not counting the image data).

27

Bill Gardner 01.30.07 at 10:07 am

“I video my ride to and from work just in case I need the footage after a crash.”

God rest your soul.

28

Ajax 01.30.07 at 12:29 pm

“has one terabyte of storage space mounted in 2 RAID drives (I use them to back each other up).”

Precisely how can two drives “back each other up”? Does each drive contain a sub-drive which is a copy of the other drive? If so, does each sub-drive contain a copy of the copy of itself which is contained on the other drive’s sub-drive? And so on, ad infinitum?

Do you really mean you back-up some third drive twice, once on each of these two drives?

29

tom s. 01.30.07 at 7:12 pm

And so on, ad infinitum?

The ultimate storage technology.

Comments on this entry are closed.