Re: iSCSI: Last call issues for the Naming and Discovery document

[Mark Bakke <mbakke@cisco.com>]

When we first added B.4, it was intended to be a special case of
the B.3 gateway, with the exception that a B.3 iSCSI-iSCSI gateway
would fully terminate iSCSI targets and initiators on both sides,
while a B.4 proxy would be aware that both sides are iSCSI, and
may pass some PDUs, and the data CRC, through the box without
modification.  It may also handle target naming and other things
more transparently.  The main reason to have this section was to
point out that if a SCSI gateway is aware that iSCSI is used on
both sides, it can do things like keep the CRC all the way through.

Amir writes:
> B.4 is a subset of B.3 and can be dropped. In case we keep it,
> I suggest the following text:
>  
> An iSCSI gateway is a SCSI gateway presenting only virtual iSCSI targets
> and virtual iSCSI initiators. A virtual iSCSI device is also called an iSCSI proxy.
>  
> instead of:
>  
>  "An iSCSI proxy is a SCSI gateway that happens to be terminating the
>    iSCSI protocol on both sides, rather than translate between iSCSI and
>    some other transport."
>  
> An iSCSI proxy by itself is not a gateway making above sentence inaccurate.
>  
> Amir

-- 
Mark A. Bakke
Cisco Systems
mbakke@cisco.com
763.398.1054