Re: iFCP as an IP Storage Work Item

["Mark A. Carlson" <mark.carlson@sun.com>]

In my mind, the issue of whether this should be a work item should
only revolve around two issues: 

	1) Is there a customer need for this kind of protocol that
	   is not being addressed by this or other working groups?

	2) Is there enough interest in the working group (from 
	   multiple companies) to support the development of the 
	   protocol?

On 1), I suspect that we will need to address some sort of "gateway"
protocol for customers to use in transitioning from FC Fabrics to IP
Networks. Today's reality is that we have FC host OS drivers and HBAs
as well as FC devices that will be with us for some time during the
transition. If you assume that you need a gateway protocol that does
not require the upgrade of either end, I think that 1) is satisfied.
There may be other gateway protocols needed as well for when you have
an iSCSI HBA and OS Driver, but existing FC devices. Look at the market
for the FC/Parallel SCSI routers/bridges to see a similar transitioning
market. Whether we standardize it or not, these type of devices will
exist due to the market need.

On 2), I'd like to see more diversity in the support of iFCP among 
individuals from more than just the handful of companies that have
spoken so far, but the real proof is in the interest in helping to
author the related documents.

One possibility is to task a subgroup with FC to IP storage transition
protocols, and leave the FCIP subgroup to focus just on the bridging
and tunneling issues.

My thoughts.

-- mark