Re: ULTRA ATA/100 announced

From: Dr. Kelsey Hudson (kernel@blackhole.compendium-tech.com)
Date: Tue Jun 06 2000 - 12:52:50 EST


SCSI is still a better bus

I heard somewhere (although Andre would probably know for sure) that even
though ATA is deemed 'busmastering,' meaning you can access (command) both
drives at the same time, you can still only recieve data from one drive at
a time. Hence the reason SCSI is better, being that you can access and
recieve output from every drive on the chain at the same time. I also
heard somewhere that SCSI releases CPU strain, because a lot of the I/O is
done on the card and not by the CPU. But then again this could be a
controller-specific issue too...
 
SCSI also allows greater flexibility. At max on a single ATA channel, you
can have 2 drives. SCSI allows for 7 with narrow controllers, and 15 with
wide and ultrawide controllers. Mind you, this is taking 1 slot in the
machine and one interrupt line. To come anywhere close to this using ATA,
you would need 2 cards, taking 2 slots and 2 IRQs (but usually 4), gaining
the use of 8 drives. Make that 4 cards if you want up to 16 devices.

SCSI also has the flexibility of external devices, which to my knowledge
ATA still has yet to support (except in paralell port versions, which are
slow).

If none of these things appeal to you, then ATA is fine. I, for one, use
SCSI for all 3 of those reasons above (saying the first is a true reason).
But, in other machines where it's not necessary (and of course to save me
some MONEY), I use ATA, because it's available, performs well enough, and
its cheap.

Besides, Ultra160/M SCSI is faster anyways. :)

 Kelsey Hudson khudson@ctica.com
 Software Engineer
 Compendium Technologies, Inc (619) 725-0771
---------------------------------------------------------------------------

On Mon, 5 Jun 2000, Stephen Torri wrote:

> If we get ATA/100 then why buy SCSI?
>
> Stephen
>
> --
> Stephen Torri
> s.torri@lancaster.ac.uk
> Lancaster Univeristy
>
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
> Please read the FAQ at http://www.tux.org/lkml/
>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:25 EST