Steve Timm (SYS)
stevetim at microsoft.com
Fri Mar 3 14:39:24 PST 1995
Note that SCAM intolerant devices are those that are not
compatible with the PnP SCAM protocol in such a manner
that they will not work when on the same bus. Clearly, this
is catastrophic and thus such configurations are not permitted
as they are the antithisus of Plug and Play. But ...
There is a general assumption here that most (more than
95% of) legacy devices are SCAM tolerant. Limited sample
testing has not raised any issues to date either by the device
vendors or during a recent Plug and Play SCSI test fest. At the
test fest, testing with a mix of PnP and Legacy devices was
encouraged. Another test fest is being planned for SCSI sometime
in April/May timeframe. A larger test sample is expected to include
a matrix of legacy and PnP devices with results publicly distributed.
The goal of PnP SCSI is to be fully backward compatible with
the installed base. The message is simple, legacy devices
may coexist on a PnP SCSI bus. Legacy host controllers may
function with PnP peripherals on the bus. If this is not true, then
SCAM has not met the goals set for PnP. But as I stated, there
is no data to indicate that SCAM intolerance is a problem.
From: Thomas Tewell <thomas.tewell at seqadvtech.com>
To: <scsi at wichitaks.hmpd.com>
Cc: <tore.slotfeldt at seqadvtech.com>
Date: Thursday, March 02, 1995 1:08PM
I've been looking over the SPI document and it would appear that it
states that legacy and SCAM devices cannot be mixed on the SCSI bus.
Is this correct? I've been to several 'plug and play' meetings where
much discussion has taken place concerning the proper actions for
mixing the device types on the bus.
Specifically B.3.1 Configuration requirements states:
"a) SCAM intolerant devices (i.e. legacy SCSI devices which are not
SCAM tolerant) are not permitted on the bus."
Does "SCAM tolerant" imply that devices that do not misbehave during
the SCAM protocol are allowed on the bus or that all legacy SCSI
devices are precluded from playing on the bus? The definition states
that "SCAM tolerant" means "An SCSI device which does not implement
the SCAM protocol but which complies with certain requirements
specified by this annex."
How likely is this? Is there anyone testing devices for SCAM
tolerance? It seems to me that there is going to be a GREAT deal of
confusion from the end-users about SCAM. I'm a bit worried that the
computer media pundits will rip SCAM to shreds if we are not very,
very specific about this issue...
Thomas "Rick" Tewell
Sequoia Advanced Technologies, Inc. TEL: (415) 459-7978
55 Shaver Street, Suite 240 FAX: (415) 459-7988
San Rafael, CA 94901 BBS: (415) 459-7977
U.S.A. EMAIL: thomas.tewell at seqadvtech.com
More information about the T10