Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!mailrus!umich!yale!cs.utexas.edu!usc!rutgers!bellcore-2!bellcore!messy!mo
From: mo@messy.bellcore.com (Michael O'Dell)
Newsgroups: comp.arch
Subject: Re: why SCSI is good
Keywords: SCSI
Message-ID: <26132@bellcore.bellcore.com>
Date: 18 Aug 90 14:33:11 GMT
References: <1990Aug17.134957.735@math.lsa.umich.edu> <105075@convex.convex.com>
Sender: news@bellcore.bellcore.com
Reply-To: mo@bellcore.com (Michael O'Dell)
Organization: Center for Chaotic Repeatabilty
Lines: 21

SCSI spans a much larger performance space than all it's competitors.
By building a basic double-wide double-fast SCSI-II controller,
and then using only one cable at regular speed, you get an interface
which can provide a CHEAP interconnect at modest speeds for small
numbers of disks, printers, scanners, tapes, etc, all the way up to
a theoretical 40 Megabytes/second (probably 30-35 is readily realizable)
for the double-speed-double-wide interface for things like disk
arrays, all under the imperfect-but-better-than-anything-else-around
umbrella of the Common Command Set.  And it doesn't take half an
acre of custom VLSI to do it (a clear distinction from IPI-3).
(Oh yes, the speeds assume that your bus and memory controller
are hoss enough to handle it.)

So, if you have only a finite amount of engineering time to build device
interconnects for a family of computers, doing an adaptable SCSI-II
double-wide double-fast controller leverages your engineering talent
like nothing else you can build.

This is why I predict that SCSI-II will squeeze IPI-3 right out of a niche.

	-Mike