SCSI Surrogate Mode?

Ralph Weber ralphoweber at compuserve.com
Fri Oct 15 05:30:37 PDT 1999


* From the T10 Reflector (t10 at t10.org), posted by:
* Ralph Weber <ralphoweber at compuserve.com>
*
For reference, the complete source message is duplicated at the end of
this message.

Paul,

Regarding:

} The latest SAM-2 and SPC-2 seem to contain no instances of "surrogate,"
} so I'm not inclined to believe that it is defined there.

Beyond the example cited by Charles Monia, SPC-2 provides for three specific
instances of the function you call surrogate, but it calls them embedded
devices.  You will find these in the description of the Standard INQUIRY
data (PDF pages 109 to 112 in ftp://ftp.t10.org/t10/drafts/sam2/sam2r12.pdf ).

If the Standard INQUIRY data has the SCCS bit set, then it has an embedded
array controller device, e.g., a disk device (peripheral device type 01h)
would set this bit to indicate that it is fronting for a RAID array controller.

If the Standard INQUIRY data has the EncServ bit set, then it has an embedded
enclosure services processor.  Further details of this configuration can be
found in the SES (SCSI Enclosure Services) standard.

If the Standard INQUIRY data has the MChngr bit set, then it has an embedded
medium changer and offers access to that medium changers via a limited set
of the commands found in the SMC (SCSI-3 Medium Changer Commands) standard.
Additional details of this relationship should be present in SMC, but I've
not personally verified this.  I should also note that this might be exactly
the configuration of interest to you.

As Charles noted, the method by which the embedded device communicates with
it's SCSI front device is beyond the scope of the standards in most cases.
You will note, however, that SES provides some suggestions for embedded
enclosure services devices.

I believe that T10 would be open to defining additional embedded devices,
whenever a request with suitable justification is received.

Thanks.

Ralph...

============
Subject: SCSI Surrogate Mode?
Date: Wed, 13 Oct 1999 15:17:34 -0700
From: Paul_A_Suhler at notes.seagate.com
To: t10 at t10.org


* From the T10 Reflector (t10 at t10.org), posted by:
* Paul_A_Suhler at notes.seagate.com
*
Hi, everyone.

It has been suggested to me that a "SCSI Surrogate Mode" is defined by a
SCSI-3 spec.  In this mode of operation, a removable-media drive (e.g.,
tape) has a serial connection to the media changer in which the drive is
installed.  The changer has no SCSI (or Fibre Channel) interface itself.
An initiator sends a media changer command to the drive, but at LUN 1, and
the drive passes it over the serial connection to the changer.  When the
command is finished, the changer sends status to the drive, which sends it
to the initiator.  To the initiator, the target looks like a drive with an
independent media changer, i.e., a loader.

The point of this would be to allow a low-cost changer by saving on a
SCSI/FC interface.

The latest SAM-2 and SPC-2 seem to contain no instances of "surrogate," so
I'm not inclined to believe that it is defined there.

Does anyone know whether this concept has been considered for
standardization?

Thanks,

Paul



*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org






More information about the T10 mailing list