> On Oct 2, 2017, at 8:22 AM, Jules Richardson via cctalk
> <cctalk@classiccmp.org> wrote:
>
> Does anyone know why IDE/ATA even came about? I mean, why SCSI wasn't used?
> It would have been an established standard by then, the drive complexity
> seems comparable to IDE/ATA (i.e. intelligent commands over a parallel bus),
> and SCSI controllers can be extremely simple - just a handful of LS logic ICs
> - unless you want to add loads of command queuing and such (again, comparable
> to IDE)
>
> Did it simply come down to pressure from vendors, wanting to distinguish
> between expensive workstation-class drives and something cheaper which could
> be associated with the lowly PC?
Comparable complexity, I don't think so. ATA, until about 2002, had no command
queueing. You could issue exactly one command which the drive would then
execute, and after completing that it was willing to accept another command.
Semantically ATA looks like the RK05 or RP04 disk API.
SCSI is an entirely different beast. It has command queueing, and logical
addressing. It feels more like MSCP.
Over time, especially once "native command queueing" was added to ATA, it
started to look more like SCSI and closer in complexity, but that clearly was
not the case early on.
As for differentiation, that's plausible too. This has been done for a long
time, consider that fibre channel and SCSI continued to be used for high
performance drives even after SATA started to appear in enterprise grade high
density lower speed drives. There wasn't any technical reason for associating
SCSI or FC with 15k RPM drives, or SATA with 7200 RPM drives.
paul
