Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Technology

Solid State Drives in Notebooks? 46

Posted by Cliff from the don't-you-wish-you-could-build-your-own-laptop dept.
spenney asks: "It seems like the most problematic part of any notebook is the speed of the hard drive (and they also get noisy). I noticed this site selling 2.5" solid state disks (SSDs). Anybody currently using one of these in a notebook? I can't find pricing anywhere, but they've gotta cost a fortune." How long do you think it will be before the major laptop manufacturers start adopting this technology?
This discussion has been archived. No new comments can be posted.

Solid State Drives in Notebooks? More

Solid State Drives in Notebooks?

Comments Filter:

  • My experience with laptops... (Score:4, Insightful)

    by HaloZero ( 610207 ) <protodeka@@@gmail...com> on Tuesday February 11, 2003 @10:40AM (#5279272) Homepage
    ...as a student, has been that the hard drive is usually the first piece of equipment to fail, with the LCD/TFT or optical drive (if it's a tray) following in a close second. Other concerns are batteries and power supplies, but I digress.

    The constant moving, up and down, left and right, jostling, dropping, the occasional beating-by-classmates (consider laptop being hauled around in a backpack - yes, the Targus ones are damned, good, I have one [If you need a laptop bag, GET ONE!], but the padding doesn't stop the heads from skittering across the platters when the laptop is subjected to smacking, pounding, and even spinning around.) Data is lost, the discs spin down, and it's all just one big bloody mess. Solid state drives, if affordable, could definately revolutionize the way I look at laptops, the way my school looks at laptops as a student solution, and the way the laptop community works.

    But... will it catch on? Please? I hope so. This is one thing that would suck to see it go the way of vaporware.
    • I've run in to the same problem, my college gives every incoming freshman a laptop.
      Over the course of my freshman year, I had 3 hard drives fail, and I know plenty of other people who have lost 1 or 2, and several people who've lost 3 or more. I'm just lucky now that I have a new desktop, the laptop hardly ever gets used.
      I actually made it a point of mine to NOT move the laptop around, if possible, after my second drive died, and for the most part it sat on my desk, with the same result as when I did carry it around with me.

      You know you have HD problems when the service guy can't believe that you were able to get ANYTHING off the disk, that's always a good thing to be hearing.
    • When I was young we had a saying :

      Don't move a machine while it is running.

      The theory behind this was the gyroscopic forces of a four pound (2kg), five platter hard drive spinning at 3600 rpm were incredibly strong and the drive heads were very large (quarter inch by quarter inch, or thereabouts) and were quite a bit more massive than today's itty-bitty drive heads. It was believed that yawing the drive (moving it so the spindle changed the direction it was pointing) would cause insane pressure on the bearings, and that the inertia of even a short quick movement could set the drive heads to enter a harmonic weave or bounce.

      Enter laptop drives spinning at 5400rpm - granted lighter and only a single platter, but still moving a LOT faster and now envision how much movement the laptop gets while it is turned on : you keep it in your lap, you turn it on its side, you flip it around to show your friend, you take it off your lap to put it on the desk so you can get up to get a drink, you pick it up off the desk to put it back on your lap ... all within 5 minutes.

      I would imagine that you could destroy a laptop hard drive in a weekend by vigorously flipping the laptop around while the drive is running - aye? So if you are slowly flopping your laptop around while the drive is spinning, you are merely destroying it slower.

      I would wager that you can't damage a laptop drive with the heads parked (all current drives park the heads when they power down) without cracking the laptop case and screen.

      Granted current generation IDE drives are failing in record numbers, but if you want your drive to live to its potential quit moving it while it is on. Get an external keyboard / mouse (I use the Logitech iTouch keyboard / mouse and love them), this won't affect your hard drive but will let you put the keyboard in your lap and toss it around if you like, while the laptop remains stationary. Put the laptop on a table, turn it on ... when you are ready to bail turn the laptop completely off and when it is off -then- close the lid, flip it all around and put it in your bag.

      I would LOVE a solid state drive, but at a dollar a meg I'm not getting one any time soon. Treat your laptop like it was a delicate, fragile piece of precision hardware and your hard drives are going to last a LOT longer.

  • they say the price is $1-2/mb (Score:3, Informative)

    by pavel_pod ( 626544 ) on Tuesday February 11, 2003 @10:47AM (#5279335)
    ... see this page:
    http://www.bitmicro.com/products_edfeatures .html
    • Hey that's pretty cheap! ( looking at the 50 meg quantum LPS50 from my Amiga that I remember as being a real bargain at about $200 )
      Hey that's pretty expensive! ( looking at the 30 gig IBM laptop drive that I remember as being a real bargain at about $200 )
      Why don't you just load up on ram and set up a software ram drive, and get one of those 500meg compact flash things ( a real bargain at about $160 ) that they use in digital cameras for long term user file storage?
      That would be the cheapest. Booting and program storage is an exercise left to you, but a boot from cd setup would work.

      Lots of ram usually beats a fast disk unless you are sloshing around big files like for instance in video editing or manipulating a 300 meg photograph, and even then an extra gig of ram is probably a better investment than a super fast hard drive.

      Servers are a whole other ball of wax of course.

      Is there some problem you actually trying to solve ?
      • The problem I'd like to solve is the excessive noise and limited speed of mechanical laptop HDDs. I would also imagine a SSD would require less power and produce less heat than a traditional HDD.

        Really though, just wanted to see if anyone is doing it and whether it makes sense to spend $1/mb for it. :)
      • get one of those 500meg compact flash things ( a real bargain at about $160 ) that they use in digital cameras for long term user file storage

        Um, aren't the 500 meg compact flash devices actually miniature hard disks, e.g. the IBM microdrive? (I may just be showing my ignorance though.)
        • No, the 500 meg compact flash devices cost about the same as the 1 gig IBM microdrive miniature hard disks, ( when I bought mine ) but use much less power. I'd guess they don't mind being dropped or run in cold, damp or dirty environments either which is why I use them in my camera. Oddly enough the IBM disks are supposed to be faster though.
        • No. Last I looked you could get FLASH based CompactFlash II format cards all the way up to 1G byte. They were also mentioning plans for larger ones. Nice thing about compact flash is you just strap the right pin and they behave like an IDE disk. All it takes is a passive converter to connect them to and IDE bus.

          • I have done this (Score:3, Informative)

            by Glonoinha ( 587375 )
            I have a PCMCIA adapter for CompactFlash cards and use this as a RAM disk on my laptop and can offer up the following observations :

            It works nicely, but it is a little slow. If you use it like a massive floppy drive for moving massive files between laptops, it is great. Totally reliable. Just a little slow, about on par with IDE drives of a couple of years ago - 1.0MB/s or a little slower is what I remember my rig running. I had visions of running a database app on it (no moving parts! zero latency!) but the read/write speeds and throughput throttled the system pretty bad.

            I just re-benchmarked it, read speed peaked at 875KB/s over the course of 24Megs of data, averaging around 500KB/s - 700KB/s, write speed peaked at 435KB/s averaging maybe 400KB/s over the course of 40Megs total in three large files.

            They come in sizes up to 1G, and the prices on those are dropping FAST (under $150 now for a Gig, maybe $200.) For a removable media they are great. For moving massive files around between computers they are great, esp. if the machines are not networked. For storing a bunch of data while you reinstall your OS and apps - great.

            Hope to replace your hard drive? Sorry but not really fast enough. I guess if you had LOTS of RAM (enough that your machine doesn't swap) and just wanted to boot your computer (OS, apps) from the CompactFlash you could put four 1G cards on adapters and fill the two IDE channels on a computer - one for the OS and the other three to store your programs and data on ... say for use on a small boat (so the constant lurching of the boat didn't destroy the hard drive) it would be great (if a little slow.)

            Would be really quiet though, and if you coupled it with a CPU that was a few generations old (say a mid range Celeron, perhaps) that could use one of those heat-pipe coolers with no cooling fan - totally silent computer.
            • Your working on older technology. The newer Compact Flash (CF) cards are much faster than just a year or two ago. The newer digital camera's have been forcing the CF cards to faster speeds. I definately wouldn't want to wait 6-7 seconds to store a 6 megapixel image. The new CF cards can write about 3-5 MB/sec sustained and read at up to 20MB/sec. Some manufactures claim higher write speeds, but I haven't confirmed this fact myself. Watch the claims closely, some manufactures quote the read speed and pretend it is the write speed. The best place to get hard data is from the digital camera sites that test actual units rigorously and give you the exact part number. The best high speed CF card is still the IBM Microdrives though FYI.

  • What about the limit on number of rewrite? (Score:4, Insightful)

    by linuxghoul ( 16059 ) on Tuesday February 11, 2003 @10:55AM (#5279398) Journal
    Doesnt Flash memory have a really low number of rewrites, like 10,000 after which the chip goes bad? To me, this means tht one just cant use a flash chip as primary storage with regular consumer operating systems...think /tmp and /var/log and their equivalents under win32. Or look at yesterdays story about the sector which holds the FAT, which is written/rewritten every time a file on the filesystem is modified. 10,000 total modifications, and ur FAT sector (and probably the physical chip its located on? i am not sure...) craps out. Heck...that means, a new device might not even last through the installation of a linux distro.

    or is this a different kind of flash from an alternate universe that i dont know about. I noticed on the webpage, they mention a very high MTBF, which is logical, but dont say anything about the number of rewrite cycles...

    Ghoul2

    • Re:What about the limit on number of rewrite? (Score:4, Informative)

      by ikeleib ( 125180 ) on Tuesday February 11, 2003 @12:15PM (#5280086) Homepage
      Flash disks have a layer between what the computer sees as it's "blocks" and what are really it's blocks. It uses a system that evenly distributes writes around the memory and marks off bad blocks. Unless you frequently write data onto your entire disk (like a video recorder), it isn't much of a problem.

      Also, many flash parts have a 1e6 writes rated life span. That is, they will survive a *minimum* of 1e6 writes or you can have your money back.

    • Re:What about the limit on number of rewrite? (Score:4, Informative)

      by mr3038 ( 121693 ) on Tuesday February 11, 2003 @02:17PM (#5281287)
      Doesnt Flash memory have a really low number of rewrites, like 10,000 after which the chip goes bad?

      Actually, they say that typical endurance is 27 years for a drive that gets written 100GB a day [bitmicro.com] and 28000 years if the drive gets written only 100MB a day. And those are just for 1GB model. 4.6GB model can take 100GB a day and still survice 123 years. I'd call that damn reliable. No details how they do that but I guess there's some hardware layer that remaps new data to least used areas.

      The only thing I don't like is the read and write speeds. And the price, probably.

      • I talked to one of the reps on the phone, oh say an hour and a half ago. He claims that the write failure is 1 million writes per sector. They have never deployed anyting larger than a TB so needless to say they were excited to hear from me .. ( see the lower post about my specific application)
      • Yes, SSD disk are very reliability. This would be great for the consumer. But do the manufacturers want to make drives that are that reliable? That kind of reliability could cause sales to reduce, due to the devices longer life span. Why would anyone consider replacing a drive that has the penitential to work reliably for 20 + years, if it meet their needs. I'm still using my old 6 GB drive. I will continue to use it until it dies.

  • Laptops? They dont think so... (Score:3, Insightful)

    by GeekWithGuns ( 466361 ) on Tuesday February 11, 2003 @11:05AM (#5279484) Homepage

    From their own Applications page [bitmicro.com] you can see that their not even looking for the laptop market:

    Portable Computer Applications

    Notebook and laptop computers will use Solid State Disks as the main external memory storage because of their low power consumption and resilience to mechanical stress. There is no need for the external memory storage to withstand environmental conditions that are better than those required by the LCD screens. LCD environmental requirements are generally more restrictive than those for mechanical disk drives.

    The majority of portable computers will continue to be equipped with magnetic and optical mechanical drives. The competition in this market will be challenging because the customers in this market will be price sensitive. We still believe that some customers who need the higher performance and reliability of our E-Disks® will be willing to pay more.

    Not that it isn't a good idea, but they are just not going to price them to compete with the standard Magnetic disks. But looking at the performace [bitmicro.com] these would kick butt in any server application!

    • Your comments about the uses of such a device in servers is perceptive. I see stories about Solid-State Hard Disks(SSHD) a few times a year. Like many other technology topics, it's either cyclic, or brought out in slow news days.

      Just about everywhere I've seen them discussed in any real depth though, server applications are the ones most commonly brought up. There are 2 main reasons for this. The first is that businesses are much less sensitive to price overall than your average Joe looking for a nice system to do email with. For the forseeable future (or until a fundamental change of technology), these SSHDs are going to be really expensive, especially when you compare them to magnetic media like hard disks. ($1/meg as opposed to $1/gig).

      That's where the performance comes in. For sheer performance, you can't beat memory speeds. Even the 10k RPM drives are pokey by comparison to the access and transfer you can achieve in properly engineered SS hardware. Historically, we've seen, in general, a 1000x difference between disks and memory. (I'll probably get slammed on my numbers)

      So, if you have an application that really needs a bunch of speed in randomly accessing a great deal of data, you might be willing to pay for it if you need it badly enough. You get 1000 times the speed at about 1000 times the price.

      Personally, I'd think there are better solutions for this though. Rather than having a SSHD, you could just use more main memory and cache the hell out of your data. I used to work at a place where we had a multi-GB database, that was read into memory on boot, and then accessed from there. This was necessary due to the extreme time-sensitive nature of responses to queries necessary on the device. Disk reads of any kind would have pushed us beyond the required response times, so we just didn't have them. Sure made the system slow to boot up though :-)

  • Server Market (Score:3, Interesting)

    by Hungus ( 585181 ) on Tuesday February 11, 2003 @12:43PM (#5280393) Journal
    I know in the application I am currently developing these would be of a signifigant help. Right now I am looking at sustained wriiting in the order of 14MBs-1 then another 20MBs-1 in reads. As this is a medical application and every transaction has to be recorded potentially forever ( or 120 years whichever comes first :) ) ( also means we cant use ram to cache the database)
    One issue wil be total cost though. Currently we estimate the need for 4 clusters of drives.

    1 X 42TB cluster and

    3 X 28TB clusters.

    At $1 per MB those are some signifigant numbers.
    126 million dollars in arrays. vs something like an X Raid [apple.com] at $6038 TB-1 [apple.com] or a total of 761 thousand. There is a cost factor difference of 165.

    • Rather than attempt to store your entire 126TB of data on solid state disks, consider implementing existing technical solutions to your cause.

      1. Multi-tier the data. Early computers went CPU - Core Memory - Paper Tape for storage. This was wicked slow so they added a faster medium in the middle : CPU - Memory - Hard Drive - Paper output. Still slow, added something fast here : CPU - really fast cache memory - RAM - hard drive - paper output. Consider something between the memory and the hard drive as sort of a non-volatile cache ... perhaps a 16G solid state drive here would make a massive difference in performance without driving up the price too much.

      2. Tweak your data model. If database performance is the bottleneck, consider denormalizing your data some to make it more 'write friendly.' Also if you can trim the size of your recordsize by 10% and throughput was an issue, all of a sudden you increase performance by 10%.

      3. Perhaps something that large and importance would be best designed and implemented on something other than Apple [apple.com] hardware. With a server room full of Blade servers, three racks could hold 1024 servers, each with two 60G drives and you would be at your 126TB with the processing power of a super computer (1024 CPUs of whatever flavor you want ... 1 terahertz using 1GHz CPUs, 2THz using 2GHz CPUs, and (no clue if they are using these yet, but if they are ..) 3THz of hyperthreading badness if you can get em to build a rack using the new 3.06GHz P4s.

      4. If you are going to dream, dream big. By the time you are done writing the software Moore's Law will have caught up with you to provide you with the hardware necessary to get decent performance. If a computer simulation will take 5 years to run on today's hardware you will finish faster by waiting a year and running it on whatever hardware is top of the line then.

      Just curious about the nomenclature you used - 14MBs-l - could you break this down for me? I am guessing megabytes per second, but the -l throws me. Unless that is seconds to the -1 (which is MB/s, which is what I was guessing anyways ... D'oh)
      • I appreciate your feedsback so let me respond with the following:

        1. Medical application thus with 5000 individual databases that need redundandcy .. well cant happen in the manner you are suggesting

        2. Unfortunately because of Hippa we are limited in a number of ways. However its teh issue of trackinghe movements of 5000 users constantly through the system with Database writes that causes the bigger issue .. who looked at what for how long and whos record and what did they do to it. - every record looked at has to have this information logged .. thank you washington.

        3. Sorry sir, as a 20 year network guy with in excess of 70 industry certs in M$, Apple and Unix systems I am sticking with Apple in this. Besides you should know better than to try and change a macHeads mind on systems .. shame on you. :)

        4. Who is dreaming. quite honestly this is not that big a project. We are at final beta stage now as it is. Moores law really doesnt apply as it is data storage that is the issue not processing power. As it stands the model involves 4 clusters of machines as it is.

        fianally as you guessed it is MBs-1 not MBs-L so yes I could also write it MB/s or even MB but as i prefer standard scientific notation I shall remain with MBs-1 :)
        • R&D time. Heck, I dunno. Granted that you are in final Beta it will be a LOT easier to propose hardware improvements than software changes :) Purely from a hypothetical standpoint I might suggest :

          5000 databases? Sounds like it is massively Normalized to me, and then some. All well and good until the relationship engine is driving databases in which the relationships between databases are four or more layers deep (in a row - meaning that changing the record on the master table drives the records on all the secondary tables, and these drive the records in 3'iary tables, which in turn drive 4'iary tables ... which is how most database design mandates suggested doing it during the last 20 years) Flatten out the structure and see performance zoom - if the same data exists in multiple tables and the system is 2x as fast ... it is worth the coding necessary to keep all the data synched.

          Watch the resource allocation during production. If the system is swapping data in and out to the swapfile and the box will hold more RAM, add more RAM.

          How about 15 full size racks of Blade Servers (340 per rack), thus 5000 completely autonomous 1.3GHz machines, each with two 20G drives set up in a RAID 1 configuration ... each CPU could be processing data as an independent SQL engine with a Gig or two of RAM.

          I wasn't saying you were dreaming when you designed it, I was saying that you weren't dreaming when you designed it. If you start designing a system and all of the hardware is commercially available at design time, you aren't pushing it hard enough :p

          Is fibre channel an option? How about fiber optics between all the servers, and between the servers and the last switch before the end users? Anything faster than Gigabit networking available for the Apple servers (I don't know, I'm asking)?

          I'm not quite as long in the tooth (13 years professional, plus the 5 I spent in college) but I would love to get a look at some of the system documentation, design spec's ... make suggestions from there. Best case scenario is I will learn something and maybe even understand the rationale behind using Apples and use em myself some day in a production environment, other best case scenario being I make a good suggestion that helps the project.
          • Rather than take the time to put a bunch of info out here i will just point you to the XRaid page [apple.com]

            There is a reason there are 5000 databases :). Remember in the medical field the priority is record keeping, record keeping and record keeping especially with HIPPA.

            Processing power isnt an issue the 22 servers planned are way more than enough to handle anything we want to throw at them from this application. and if it isnt the system is completely scalable.

            Sorry, I can't go into specifics for confidentiality reasons.

  • Here's some pricing (Score:3, Informative)

    by EatHam ( 597465 ) on Tuesday February 11, 2003 @12:59PM (#5280546)
    Looking for pricing? Here you go. [tri-m.com] Of course, you will probably want to look at exactly which drive [tri-m.com] is which first.
    • What's interesting about those prices is that the solid state disks are about $1 per megabyte. This is where mechanical hard drives were a decade ago, so I'd like to see if solid state disks become affordable in the next five or ten years. When that occurs, we will finally be free of the current bottleneck in hard drive performance.
  • At work we have around a dozen web servers. Maby a couple of them actualy need more then 10gb of storage, but most of them have 20gb drives that are 75% empty.

    Right now, its as easy to put $170 drives in each machine rather then go with a centeral RAID/NFS server, but even with such a beast, it would be nice to have a solid state, realy fast, drive ~2gb. Im also thinking things like nodes in a (computational) cluster, thin clients, lab machines, etc etc.

    Does such a thing exist?

    • Exist? (Score:3, Informative)

      by Glonoinha ( 587375 )
      Sure thing, there are a few links in this thread already. About a dollar a meg, so figure $2,000 for a two gig solid state drive. Of course if your data was pretty static I would suggest a 2G RAM drive from http://www.superspeed.com - they have a free 30 day trial to check it out. If you want to play with Solid State Drives to see the kinds of performance gains you are going to get from going that direction, put an extra gig or two of memory in your favorite machine and turn it into a RAM drive - if you want to keep on using it, great, if not split up the memory and share it between your other machines because it will make them run better regardless.

      This doesn't cover all your uses, but for the ones it does cover ... it covers nicely :)

  • Power consumption. (Score:2, Informative)

    by nsrbrake ( 233425 )
    Asside from the maximum number of writes (erases), how about this:

    From the linked page the 30GB drive is listed at using
    write 3.3W, read 3.1W,idle 2.4W.

    From a fijitsu web page http://hdd.fujitsu.com/global/drive/mhs2xxx/catalo g.html
    wead/write 2.30W, idle 0.65W, standby 0.25W, sleep 0.10W.

    When I first saw the post, I was hoping for quite the opposite. The last thing I need is my laptop to run dry faster, I don't need the fast read/write in a laptop.

    • Re:Power consumption. (Score:1, Informative)

      by Anonymous Coward
      Thats for a full 36 gigs..

      Most people could get by with 2 Gigs.

      Assuming its proportional: (which I doubt, but I bet its pretty close)

      Your Wead/write would be 0.18W
      and your idle at around 0.12W

      Sounds sweet to me.

  • My company looked into the Bitmicro drives about a year ago.

    If I remember correctly, the cost was on the order of $1,000 a Gigabyte.
    Quite a prohibitive price for most applications.
    • This is a dollar a Meg - expensive by today's standard but there was a day that $1/M was cheap for hard drives so everything is relative.

      For most applications today yes, I would agree with you this is way more expensive than we have become accustomed to paying, but it is also way more performance than we are accustomed to getting - if adding a $2,000 solid state drive to a web server doubled the number of simultaneous connections it could handle that would be very cheap : $2,000 is about two days of custom development and there is no way you could get a programmer to double the performance of a system in two days.

      For commercial apps, I see these things possibly making a difference IF they can keep the performance substantially faster than regular hard drives.
  • I tested a solid state 500mb HD in 1990, while working in Japan. >Format..done. I've yet to see these in wide (consumer) application. I wouldn't hold my breath on them showing up any time soon.

  • Have 1GB Sandisk FlashDrive. Silence is wonderful! (Score:3, Interesting)

    by oakwood ( 602181 ) on Wednesday February 12, 2003 @08:38AM (#5286717)
    I have a 1GB Sandisk FlashDrive in this notebook. It a type3 PCCard in a metal frame that has a 3.5" IDE form factor, so it fits instead of the hard drive. It is wonderful.

    I do have to be careful about space and it is a little slow. Very important to defrag regularly, speed drops greatly with fragmentation. I'm using Win98 to save space. Unfortunately, it will not run with Win's Virtual Memory set low or to zero. It can be tricky to format the drive.

    Love the silence.

    http://www.sandisk.com/oem/flashdrive.asp
    • Looks like they have a 2GB version now, but I'm wondering if that would be enough. Given that WinXP + Office takes up about 1.75GB all by themselves I'd think 4GB is probably the sweet spot. Even Linux is getting pretty beefy on disk, especially with all the kernel dev packages, KDE/GNOME, etc.

      I suppose you could get a 2GB SanDisk for your boot/swap drive and then put in one of their 2GB PC Cards (Type II slot) for apps and such. Add a wireless adapter and you're all set. Anyone know if the PC Cards will work under Linux?
  • i saw few months ago, toshiba 2 gig pcmcia drive for $13 after $100 mail in rebate. Currently, 5 gig versions are on sale for about $150. that would make it about 3 cents/mb which is about 30-60 times cheaper than speculated price of $1-2 for BitMicro.
    • The whole idea is going solid state in leiu of spinning media.

      Take a 1G CompactFlash card and a 2G PCMCIA Toshiba hard drive. Get in your car. Driving down the street at 30 miles an hour, toss both out the window. Find them, take them home, put them in a laptop.

      Actually the original idea for the thread was using solid state drives (not the SanDisk CompactFlash like I am using - I think the thread got hijacked to the tune of flash disks as alternatives to hard drives) to get zero noise, incredibly fast access, and less fragile - the PCMCIA drive by Toshiba may be quiet, but not silent, and fast, but not as fast as the BitMicro stuff wants us to believe the SSDs are.

  • get these drives cheap on ebay. (Score:1, Informative)

    by Anonymous Coward
    I wanted to use one of these drives on a laptop for a project I was working on. I found a company that sold them, then paid $250 for a 128 Meg one. But then I checked on ebay, and paid $100 for a 1 gig one on ebay, this would have normally cost around $3000.
    • If `these drives' are Flash drives, be very careful buying used materials -- Flash drives wear out much more quickly than memory, or spinning drives, or whatnot (generally guaranteed between 100,000 and 1,000,000 writes).

      If you're not talking about Flash, then ignore the above, and accept my apologies for the confusion.

Slashdot Top Deals

A morsel of genuine history is a thing so rare as to be always valuable. -- Thomas Jefferson

Close