XOR Commands Meeting - response to comment

Jay Elrod Jay_Elrod at notes.seagate.com
Mon Dec 12 13:13:44 PST 1994


This is in response to a recent message on the reflector regarding
the 11/7/94 XOR meeting minutes.
 
George Penokie wrote:
 
>I just read through the minutes of the 11/7/94 XOR meeting (yes some of
>us do actually read minutes) and I have some major concerns.
>
>It looks to me like there are proposals to define things (like redundancy
>goups and RAID 5 mode pages, addressing etc.) that are already defined within
>the SCC Standard. The SCC model covers all levels of the system and this
>includes drive that have SACL like functions built into them.  If a drive
>wants to be a storage array (which it can) then it must follow the rules
>for SCC devices. And like it or not putting XOR functions in a drive that
>controls the flow of data to other drives to do RAID functions is an SCC
>device.
>
>I do not have a problem with the XOR group defining new drive commands to
>deal with the movment of data between drives, as defined in Seagate's XOR
>Commands proposal.  But it appears there is a movement to expand beyond
>that original charter.  This expansion can only bring confusion to the
>industry if terms and concepts used in the XOR proposals are not consistant
>with the defination of the those same terms in the SCC standard
>(ie Redundancy group, addressing, etc.).
 
In response:
 
The primary purpose of the recently added RAID Group Address mode
page is to provide a mechanism for allowing a greater-than-3-byte physical
address to be used in addressing secondary targets during XOR
commands. This is necessary (for some interfaces) since the CDB only
has room for 3 bytes of secondary address information. For those
interfaces which require more than 3 bytes for the physical address the
initiator would send the RAID Group Address mode page containing the
desired address translation information to the target. The initiator would then
use a 1 byte "logical" address in the XDWRITE CDB, for example, and the
target would use the RAID Group Address mode page information to
translate that 1 byte address to a "physical" (up to 8 byte) address. Note
that the initiator is entirely in control of the address of the secondary
target.
 
Assuming there is agreement within the XOR meetings (and there has been so
far) our intent is to integrate the RAID Group Address mode page into the Direct
Access Device section of the SCSI standard, limiting its use to only those
physical interfaces which require more than 3 bytes for a physical address.
 
As with the entire XOR Command proposal, the RAID Group Address mode
page is the product of much thought and discussion. It certainly in not above
scrutiny or change, however, (nor is any other part of the XOR proposal) and
if there is a reasonable alternate method recommended to accomplish the
intended purpose, within the context of direct access devices, it will be
considered.
 
Jay Elrod
--
Jay Elrod -- Jay_Elrod at notes.seagate.com
-------------------------------------------------------------------------
Seagate Technology   -   920 Disc Drive   -   Scotts Valley, CA 95066 USA
Main Phone 408-438-6550   -   Email Problems postmaster at notes.seagate.com
Technical Support: BBS 408-438-8771  Fax 408-438-8137  Voice 408-438-8222  
-------------------------------------------------------------------------

### OGATE Version 8 message trace and attachment information:
### MsgFileName: m:\mgate\outbound\2059.MSG
### Org Date:    12-12-94 08:37:03 AM
### From:        Jay Elrod at SEAGATE
### To:          SCSI @ wichitaks.ncr.com @ internet
### Subject:     XOR Commands Meeting - response to comment
### Attachments: none




More information about the T10 mailing list