I just bought 2 TB in four 500 MB drives for less than $430 including shipping (that's an improvement from $150 per MB in 1979 to about $0.22 per MB**). With the great tools now available on most motherboards, I arrayed these in a fast and redundant Raid 0+1 setup with 1TB of storage. (Yes, to the total geeks out there, I would have preferred Raid 1+0 but alas the Nvidia chipset on my board did not support it.)
** By the way, this 700x improvement over 30 years actually has little or nothing to do with Moore's law. While some of the materials sciences are related, this improvement has little to do with silicon and nothing to do with transistor density. This is the result of incredible human creativity in the face of brutal competition, both from other hard drive manufacturers as well as from substitutes like static RAM.
Posted on January 23, 2008 at 08:39 AM | Permalink
Out of curiousity, did you consider Raid 5? It would have given you 1.5TB..
Posted by: HTRN | Jan 23, 2008 8:52:06 AM
I still think about the 8 meg Radio Shack Hard Drive I bought in the 80's for about $2500. And then I think about how much that hard drive increased my productivity compared to the increase in productivity you probably got for your recent purchase of a 2TB unit.
Posted by: tmitsss | Jan 23, 2008 9:44:31 AM
I think you're off by a factor of 1,000. Your drives are not 500MB, they're 500,000 MB, meaning they cost $.00022 per MB, or 700,000 times more storage for the buck.
Posted by: Larry | Jan 23, 2008 10:05:12 AM
I think hard drive capacity DOES owe a lot to Moore's Law (and to other factors, but I think 'little or nothing' is probably inaccurate). It's true there are no transistors on the surface of the hard disk. But packing so much data in such small areas decreases the signal-to-noise ratio of the detected data, making it far more difficult to reliably recover the data. Therefore sophisticated digital signal processing algorithms are now used in the integrated circuits that receive the data, enabling hard drives to get closer to theoretical limits; implementing these algorithms was not feasible decades ago, precisely because more transistors can be packed on a chip now.
Posted by: Todd | Jan 23, 2008 1:10:56 PM
You might note that the improvement in CPU power is also "the result of incredible human creativity in the face of brutal competition". Moore's law is just a description of what's happening and has little predictive power. You make it sound like a law of nature or something. The engineers of Intel, AMD, IBM, and other companies work their butts off and would be sad to hear how little you appreciate their efforts.
Posted by: Josh | Jan 23, 2008 1:55:35 PM
Just for clarification, you've got a typo. That should be 500 GB drives. Otherwise you got suckered.
Posted by: Max Lybbert | Jan 23, 2008 10:28:43 PM
However, you might have been better off with this device: Drobo
Posted by: Some Dude | Jan 24, 2008 7:44:59 PM
Back when I was in graduate school in the late 1970s, I used to go up to NCAR (National Center for Atmospheric Research) in Boulder, Colorado to use their computer. It had the "Terabit Mass Storage" system (note: bit). This was a tape drive with 3-inch wide tapes on reels a yard across. It was used like a slow hard drive. They had discovered that if you ever demounted a tape, you'd never be able to read it again.
Now I have a terabyte (note: byte) on the computer I'm typing this on (500Meg internal and 500Meg external for backup). And if you asked me why I need a terabyte, all I'd be able to do is make vague gestures and say "But the 500 Meg drives hardly cost more than the smaller ones!"
Posted by: Bob Hawkins | Jan 26, 2008 3:44:17 PM
The comments to this entry are closed.