Using USB Memory Stick to Improve Performance

[Siggy Brentrup]

On Mon, Sep 07, 2009 at 16:49 -0300, Derek Broughton wrote:
> David Fox wrote:
> 
> > On Sun, Sep 6, 2009 at 2:42 PM, Siggy Brentrup<ubuntu at psycho.i21k.de>
> > wrote:
> > 
> > I'm sceptical of using a USB stick for the purpose you describe.
> > (There was another thread that was related to this, which I almost
> > commented on: using a usb stick in an elaborate "cache" setup,
> > presumably because USB was faster.)
> 
> The threads I recall were different because they focused, as you say, on 
> "cache" - read/write applications.  Siggy wants to put read-only data there.  
> Not such a bad idea, on the whole.

Thanks for the flowers :)

> > While USB sticks have no moving
> > parts, I'm pretty sure that disk accesses (reads) are slower than
> > using the internal HD.drive.
> 
> _Transfer_ rates are certainly faster, because you can't reach the same 
> speeds over USB.  Access speeds could be comparable.  _Seek_ times will be 
> much slower on disk.

Some impressions after one day of use:

  - Annoying noises from heads moving to and fro at high speed are
    gone.
  - Everything feels faster.
  - I already posted a 30% performance gain on booting.

More to come, still have to work out the runlevel stuff for upgrades.

There's also an obvious security issue I came over only now: always
make sure you know exactly what's on the stick before booting.

On a Sony Vaio laptop I'm using this on the better approach is to use
the magicgate slot but for now I only have a 32MiB flash sitting there
that holds my secret keyrings.  4GiB cards are a bit more expensive
than USB sticks.

Regs
  Siggy
-- 
O< ascii ribbon campaign - stop html mail - www.asciiribbon.org+
|39 days until|bsb-at-psycho-dot-informationsanarchistik-dot-de|
|www.Ubucon.de|or:                bsb-at-psycho-dot-i21k-dot-de|
+-------> ceterum censeo javascriptum esse restrictam <--------+