Can't come soon enough: flash drives

Discussion in 'NZ Computing' started by steve, Jan 4, 2006.

  1. steve

    steve Guest

    steve, Jan 4, 2006
    #1
    1. Advertising

  2. "steve" <> wrote in message
    news:...
    > I'm very much looking forward to the day when spinning metal platters are
    > retired forever and storage is based on flash/solid-state media.
    >
    > - Access times in nanoseconds instead of milliseconds
    >
    > - Quiet
    >
    > - No mechanical failures
    >
    > - robust physically
    >
    > - low power usage
    >
    > Just a few more years to go.
    >
    > http://news.com.com/Bye-bye hard drive, hello flash/2100-1006-6005849.html?part=dht&tag=nl.e433


    It would be like the invention of the petrol engine - all the old
    technologies would just disappear.

    You'd see huge companies such as Seagate, WD , Maxtor etc simply disappear.
     
    news.xtra.co.nz, Jan 5, 2006
    #2
    1. Advertising

  3. At that very moment, steve turned to nz.comp and said
    > - Access times in nanoseconds instead of milliseconds
    > - Quiet
    > - No mechanical failures
    > - robust physically
    > - low power usage


    - Reduced throughput.
    (and therefore boot will usually be slower)
    - Always smaller size for the same money as a hard drive.




    --
    aaronl at consultant dot com
    For every expert, there is an equal and
    opposite expert. - Arthur C. Clarke
     
    Aaron Lawrence, Jan 5, 2006
    #3
  4. steve

    steve Guest

    Aaron Lawrence wrote:

    > At that very moment, steve turned to nz.comp and said
    >> - Access times in nanoseconds instead of milliseconds
    >> - Quiet
    >> - No mechanical failures
    >> - robust physically
    >> - low power usage

    >
    > - Reduced throughput.


    Why? There is no reason for this to be true.

    A spinning platter's access time is 9ms on average.

    Accessing a chip 1 million times faster is unlikely to see throughput
    decline.

    > (and therefore boot will usually be slower)
    > - Always smaller size for the same money as a hard drive.


    There will be a crossover point...and that's what I'm talking about.
     
    steve, Jan 5, 2006
    #4
  5. steve

    Rob J Guest

    In article <Jy_uf.12485$>,
    says...
    >
    > "steve" <> wrote in message
    > news:...
    > > I'm very much looking forward to the day when spinning metal platters are
    > > retired forever and storage is based on flash/solid-state media.
    > >
    > > - Access times in nanoseconds instead of milliseconds
    > >
    > > - Quiet
    > >
    > > - No mechanical failures
    > >
    > > - robust physically
    > >
    > > - low power usage
    > >
    > > Just a few more years to go.
    > >
    > > http://news.com.com/Bye-bye hard drive, hello flash/2100-1006-6005849.html?part=dht&tag=nl.e433

    >
    > It would be like the invention of the petrol engine - all the old
    > technologies would just disappear.
    >
    > You'd see huge companies such as Seagate, WD , Maxtor etc simply disappear.


    Not likely, they would evolve. Quantum used to make big expensive solid
    state HDDs
     
    Rob J, Jan 5, 2006
    #5
  6. steve

    AD. Guest

    Aaron Lawrence wrote:

    > - Always smaller size for the same money as a hard drive.


    But there seems to be a constant minimum price for disks based on
    manufacturing costs - they are complex, precise and mechanical.

    So disks don't really get cheaper, they just get dramatically bigger as
    densities increase.

    Meanwhile flash drives have small but quickly growing capacities.
    Economies of scale for flash memory have kicked in with the growth of
    USB keys and digital cameras etc, and flash memory units can now be made
    cheaply. The minimum cost is much much lower than that for a HD.

    There will come a cross over point soon where the size of cheap flash
    drives is big enough for non demanding computer storage needs (eg less
    than 40GB). When that happens, the better price per GB of the cheapest
    HD won't matter because the HD is so much bigger than what the low end
    of the market needs.

    eg at that point (in say 3-5yrs) a 40GB flash disk might be a cheaper
    unit than the minimum sized HD (say 200GB then) even if the disk still
    has a better cost per GB.

    I think flash disks in laptops would be great for durability, battery
    life, heat etc etc. Does the average laptop user really need 60GB+ of
    storage?

    Of course users wanting lots of space will still want HDs for quite a
    while yet.

    --
    Cheers
    Anton
     
    AD., Jan 5, 2006
    #6
  7. T'was the Thu, 5 Jan 2006 14:23:39 +1300 when I remembered
    "news.xtra.co.nz" <> saying something like this:

    >You'd see huge companies such as Seagate, WD , Maxtor etc simply disappear.


    I'd say they'd migrate to flash technologies themselves over going
    "Well that was a good run boys, I'm off to Pukekohe in New Zealand to
    retire".
    --
    Cheers,

    Waylon Kenning.
     
    Waylon Kenning, Jan 5, 2006
    #7
  8. steve

    Don Hills Guest

    You won't see flash based drives replacing conventional drives any time, let
    alone any time soon. Flash memory has an inherent characteristic that makes
    it unsuitable for general read/write use: limited number of write cycles.
    You're more likely to see MRAM (Google it) or similar technology as a hard
    disk replacement.

    (MRAM is faster and denser than DRAM and is non-volatile.)

    --
    Don Hills (dmhills at attglobaldotnet) Wellington, New Zealand
    "New interface closely resembles Presentation Manager,
    preparing you for the wonders of OS/2!"
    -- Advertisement on the box for Microsoft Windows 2.11 for 286
     
    Don Hills, Jan 5, 2006
    #8
  9. steve

    AD. Guest

    Don Hills wrote:
    > You won't see flash based drives replacing conventional drives any time, let
    > alone any time soon. Flash memory has an inherent characteristic that makes
    > it unsuitable for general read/write use: limited number of write cycles.


    Which also seem to be growing. Apparently some newer stuff can take 1M
    cycles. And there are dynamic remapping schemes that offset that a bit.

    > You're more likely to see MRAM (Google it) or similar technology as a hard
    > disk replacement.
    >
    > (MRAM is faster and denser than DRAM and is non-volatile.)


    Possibly, but I don't see where the economies of scale will come from in
    the near future. MRAM will have to displace Flash in consumer
    electronics or DRAM in computers before that will happen. Flash has the
    current economics advantage (due to all the consumer device demand) that
    tends to trump technical advantages.

    Flash is not perfect, but good enough seems to beat near perfect all the
    time when economics is on its side.

    Flash currently seems to have far greater capacities than MRAM - ie
    gigabits vs megabits and quickly rising. Also Flash is now cheaper than
    DRAM and MRAM is unlikely to be cheaper than DRAM until more MRAM is
    produced than DRAM.

    I suppose the question is will the Flash vs HD crossover happen before
    the possible MRAM (or similar) vs Flash one? My guess is that Flash will
    take over low end HDs (maybe 3-5 yrs) before whatever replaces Flash is
    ready for the larger capacities (maybe 5-8 yrs).


    Although it is interesting wondering about a future MRAM equipped
    computer (for both working memory and storage). Where would the working
    memory live? In a virtual memory file in the filesystem maybe? ;) Or
    more likely the filesystem would live in a ramdisk. It would presumably
    also need a 64bit OS for addressing it all :)

    --
    Cheers
    Anton
     
    AD., Jan 5, 2006
    #9
  10. steve

    Don Hills Guest

    In article <V6ivf.12790$>,
    "AD." <> wrote:

    (a differing opinion)

    Well, we'll just have to wait and see who will be right. Probably neither of
    us, something will come out of left field that no-one thought of. :)

    >Although it is interesting wondering about a future MRAM equipped
    >computer (for both working memory and storage). Where would the working
    >memory live? In a virtual memory file in the filesystem maybe? ;) Or
    >more likely the filesystem would live in a ramdisk. It would presumably
    >also need a 64bit OS for addressing it all :)


    IBM introduced a working system using that architecture 30 years ago and is
    still using it today. It's currently known as the AS/400(*). It has no
    concept of a filesystem as you know it. Everything (programs and data) is in
    memory at all times, memory mapped in a 64-bit address space. Since the
    machine doesn't have that much RAM, it uses a VMM (Virtual Memory Manager)
    to swap out LRU (Least Recently Used) pages to make room for pages that are
    currently being accessed. The swap medium is currently hard disks, and the
    disk formatting includes a 64-bit field at the beginning of each sector
    which contains the 64-bit memory address of the first byte of data in the
    sector.

    It's fully object oriented, program and data objects are encapsulated and
    can only be accessed by the defined methods. The security model is enforced
    by hardware so there's no way to trick it into doing something above your
    privilege level. A pretty boring machine by computer enthusiasts standards,
    but beloved by businesses who want a reliable business system that requires
    a minimum of skilled administration.

    (*) You may never have heard of the AS/400, but if the AS/400 division of
    IBM were to be split off from the rest of IBM it would be the second largest
    computer company in the world behind the rest of IBM.)

    --
    Don Hills (dmhills at attglobaldotnet) Wellington, New Zealand
    "New interface closely resembles Presentation Manager,
    preparing you for the wonders of OS/2!"
    -- Advertisement on the box for Microsoft Windows 2.11 for 286
     
    Don Hills, Jan 6, 2006
    #10
  11. steve

    steve Guest

    Don Hills wrote:

    >
    > You won't see flash based drives replacing conventional drives any time,
    > let alone any time soon. Flash memory has an inherent characteristic that
    > makes it unsuitable for general read/write use: limited number of write
    > cycles. You're more likely to see MRAM (Google it) or similar technology
    > as a hard disk replacement.
    >
    > (MRAM is faster and denser than DRAM and is non-volatile.)


    I wasn't being hung up on the access method or technology available today.

    I'll settle for any solid state memory that is cheap, fast, capacious and
    MUCH faster than spinning platters.

    :)
     
    steve, Jan 6, 2006
    #11
  12. steve

    steve Guest

    news.xtra.co.nz wrote:

    > You'd see huge companies such as Seagate, WD , Maxtor etc simply
    > disappear.


    I think Maxtor is disappearing anyway as Seagate just bought them.
     
    steve, Jan 6, 2006
    #12
  13. steve

    MarkH Guest

    (Don Hills) wrote in
    news::

    >
    > You won't see flash based drives replacing conventional drives any
    > time, let alone any time soon.


    You are a bold one!

    Personally, I would never dare to state that something will never happen -
    I don't see that you could ever know the future for certain.

    I can't see that the number of write cycles would be that limiting -
    especially as the limits increase. There are ways to reduce the amount of
    writes anyway, like working more within the normal RAM and writing to flash
    RAM less.

    We already have OSs that will load completely into RAM and run from there,
    how much easier will that be when a standard machine has 4GB RAM, or even
    64GB RAM?



    --
    Mark Heyes (New Zealand)
    See my pics at www.gigatech.co.nz (last updated 5-September-05)
    "The person on the other side was a young woman. Very obviously a
    young woman. There was no possible way she could have been mistaken
    for a young man in any language, especially Braille."
    Maskerade
     
    MarkH, Jan 6, 2006
    #13
  14. At that very moment, steve turned to nz.comp and said
    > > - Reduced throughput.

    >
    > Why? There is no reason for this to be true.
    >
    > A spinning platter's access time is 9ms on average.
    >
    > Accessing a chip 1 million times faster is unlikely to see throughput
    > decline.


    I'm talking about the bandwidth for transferring data (ie. MB/s). Flash
    is generally some distance behind hard drives, which isn't what you'd
    expect but it's true...

    --
    aaronl at consultant dot com
    For every expert, there is an equal and
    opposite expert. - Arthur C. Clarke
     
    Aaron Lawrence, Jan 6, 2006
    #14
  15. steve

    steve Guest

    Aaron Lawrence wrote:

    > At that very moment, steve turned to nz.comp and said
    >> > - Reduced throughput.

    >>
    >> Why? There is no reason for this to be true.
    >>
    >> A spinning platter's access time is 9ms on average.
    >>
    >> Accessing a chip 1 million times faster is unlikely to see throughput
    >> decline.

    >
    > I'm talking about the bandwidth for transferring data (ie. MB/s). Flash
    > is generally some distance behind hard drives, which isn't what you'd
    > expect but it's true...


    Yep.....today.

    I was talking about when it's cheap, fast and ready for primetime.

    It might not be "flash".....but it will be solid state storage and operate
    at memory speeds.....not spinning platter speeds.

    I hope. That's what I want. :)
     
    steve, Jan 6, 2006
    #15
  16. steve

    Rob J Guest

    In article <>,
    says...
    >
    > You won't see flash based drives replacing conventional drives any time, let
    > alone any time soon. Flash memory has an inherent characteristic that makes
    > it unsuitable for general read/write use: limited number of write cycles.
    > You're more likely to see MRAM (Google it) or similar technology as a hard
    > disk replacement.
    >
    > (MRAM is faster and denser than DRAM and is non-volatile.)


    Core makes a comeback...
     
    Rob J, Jan 6, 2006
    #16
  17. "Rob J" <> wrote in message
    news:...
    > In article <Jy_uf.12485$>,
    > says...
    >>
    >> "steve" <> wrote in message
    >> news:...
    >> > I'm very much looking forward to the day when spinning metal platters
    >> > are
    >> > retired forever and storage is based on flash/solid-state media.
    >> >
    >> > - Access times in nanoseconds instead of milliseconds
    >> >
    >> > - Quiet
    >> >
    >> > - No mechanical failures
    >> >
    >> > - robust physically
    >> >
    >> > - low power usage
    >> >
    >> > Just a few more years to go.
    >> >
    >> > http://news.com.com/Bye-bye hard drive, hello flash/2100-1006-6005849.html?part=dht&tag=nl.e433

    >>
    >> It would be like the invention of the petrol engine - all the old
    >> technologies would just disappear.
    >>
    >> You'd see huge companies such as Seagate, WD , Maxtor etc simply
    >> disappear.

    >
    > Not likely, they would evolve. Quantum used to make big expensive solid
    > state HDDs


    I'm not sure - the expertise of these companies is in harddrives.

    They would be competing with companies whose expertise is in flash mem.
     
    news.xtra.co.nz, Jan 6, 2006
    #17
  18. steve

    AD. Guest

    Don Hills wrote:

    >>Although it is interesting wondering about a future MRAM equipped
    >>computer (for both working memory and storage). Where would the working
    >>memory live? In a virtual memory file in the filesystem maybe? ;) Or
    >>more likely the filesystem would live in a ramdisk. It would presumably
    >>also need a 64bit OS for addressing it all :)

    >
    >
    > IBM introduced a working system using that architecture 30 years ago and is
    > still using it today. It's currently known as the AS/400(*).


    Yep, AS/400s are 'different' to how the rest of us see PCs and more
    reliable than anything short of a mainframe. I did some short term tech
    support tech support at a company with 3 AS/400s - they were in their
    own separate server room and nobody there knew much about managing them.
    But they didn't really need to - issues were so rare, and the servers
    would phone home to IBM if any parts were failing etc.

    I suspect new comsumer level PCs with solid state storage would not use
    some radically different architecture though.

    --
    Cheers
    Anton
     
    AD., Jan 6, 2006
    #18
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Philip Colmer

    USB wireless not starting soon enough?

    Philip Colmer, Sep 1, 2005, in forum: Wireless Networking
    Replies:
    4
    Views:
    488
    Philip Colmer
    Sep 12, 2005
  2. Replies:
    2
    Views:
    755
  3. Mr Scebe

    Can't come soon enough

    Mr Scebe, Dec 11, 2004, in forum: NZ Computing
    Replies:
    4
    Views:
    297
    steve
    Dec 12, 2004
  4. Enkidu
    Replies:
    12
    Views:
    712
  5. AD.
    Replies:
    1
    Views:
    407
    Gordon
    Oct 8, 2008
Loading...

Share This Page