Re: linux loopback devices

[Casey Boone <caseyboone@gmail.com>]

i just noticed something i hadnt before, the traditional loopback
device is depreciated in favor of dm-crypt

unfortunately the box is now in production so i can use it for
testing, but i  think i have a machine i can use instead and see how
things compare, maybe it isnt nearly as slow

Casey


On Tue, 8 Feb 2005 20:52:21 -0600, Casey Boone <caseyboone@gmail.com> wrote:
> the different using the loopback with encryption and without is 2MB/s,
> while the difference between using the loopback and not using it is
> 30MB/s.  i even rechecked myself because i just knew i had something
> set up wrong.
> 
> even taking into consideration that using the loopback would mean an
> extra in and out, this all occurs in ram and i just cannot imagine
> that the ram access is that slow.
> 
> after showing the performance differences the customer decided to
> leave it unencrypted (not sure why he wanted it encrypted anyway, the
> way he wanted it would have meant that the key was stored on the
> system...)
> 
> it still puzzles me, so i will probably look into it more later, might
> post to the lkml and ask, didnt see anyone else with the same problem
> (but it might have been lost to me as there are quite a few emails
> containing "loopback")
> 
> Casey
> 
> 
> On Tue,  8 Feb 2005 19:27:49 -0600 (CST),
> mike808@users.sourceforge.net <mike808@users.sourceforge.net> wrote:
> > Casey Boone wrote ..
> > > customer wants his data encrypted, but the performance of the loopback
> > > is utterly horrible
> >
> > Remember that the system is processing the data twice in the loopback.
> > Once in, once out.
> >
> > Also, are you sure it's not the encryption that is slowing you down?
> > Encryption isn't exactly a cheap computation either.
> >
> > mike/
> >
> > -
> > To unsubscribe, send email to majordomo@silug.org with
> > "unsubscribe silug-discuss" in the body.
> >
>

-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.