From:    "Tony B. Shepherd" <[log in to unmask]>
Subject: Re: *PLUG* - HP compat RAID
 
>On Tue, 17 Jan 1995, Richard Gambrell <[log in to unmask]> wrote:
>> Subject:      Re: *PLUG* - HP compat RAID
>> Open hardware is different than open software - or so it seems to me.
>> Posix, so far as I know, defines open software. SCSI is supposed to
>> be an open standard for hardware, but MPE appears to be "special". The
>
>But "open systems" (IMHO) should cover hardware as well as software.  If we
>can't explain why HP's SCSI drive costs so much more than the drive at the
>computer store (which HP says we can't use), we may not be able to get the
>business solution robustness of a good operating system.
 
  I don't know the cost differential, so I won't comment their;  but I do
  know that each drive that MPE supports is certified with various tests.
  Talk to Jim Knight@[log in to unmask] for supporting non-HP SCSI
  devices.
 
>So (with my jaw out a bit) I say either don't call it a SCSI drive, or show
>me how to hook up the one from the computer store.
 
Calling it a SCSI drive is correct, it follows the SCSI protocol.  As with
any protocol, there are grey areas.  Different vendors implement these grey
areas differently.  Remember the days on PC hardware where some disc
interface cards would not work with some video cards?  Both cards plugged
into the backplace.  Both cards said they followed the backplane protocol,
but did they play together well?  Some did, some didn't.  The same thing
could happen with off-the-self SCSI devices.  I not saying it would, I'm
saying I don't know.
 
>And let's not try to limit hardware alternatives either.  The days of a
>computer just sitting in a back room, and all communication with other
>systems are weekly on magnetic tape are coming to a close.  More and more
>the 3000 is part of a global peer-to-peer network.  Data interchange needs
>will continue to grow - why shouldn't a RAID or DAT start looking less like
>a peripheral and more like another computer system?
 
This is a good vision of the future.  When more network backbones get faster
and we solve sharing disks without a lot of overhead, it will be there.
 
Rick Ehrhart
My thoughts are my own.
 
 
>Just my .02  --  regards  --  Tony B. Shepherd  --  [log in to unmask]