RE: iSCSI & iFCP Overlapping (Was iFCP as an IP Storage Work Item)

["Douglas Otis" <dotis@sanlight.net>]

Y.P.

A company's success with ST-506 leveraged off an inappropriately simple
interface, floppy, and has now monopolized drive interfaces where SCSI
enjoyed a 10 year lead.  iFCP has potential in that it attempts to make
fewer changes to a complex but stable interface (some may argue not stable
enough).  I encourage those working on both FCIP and iFCP to share a common
encapsulation and define node handling in a separate proposal.  There is
significant differences between iSCSI and iFCP.  Problems in iSCSI proposals
have already been solved with FCP so unless there is a problem, what is
being fixed?  What limitations with FCP merit iSCSI?  Let the marketplace
decide as SCTP may obsolete choices made in iSCSI.  The IPS WG should
attempt to minimize variants of a common theme.  In that respect, iFCP seems
to be more in line than does iSCSI but, there are many wishing to make a
career out of iSCSI.  One wonders if they will have time should iFCP provide
a stable solution first.

Doug

> > iFCP and FCIP have nothing to do with OSPF.  I think your statements
> > show that you do not understand iFCP and its objectives.  I would
> > have hoped that those critical of iFCP would have at least a minimal
> > understanding of it.
>
> Josh,
>
> Had I stayed with the topic of "saving the customers' investment in FCP
> stack" I would have done better and not invited the above
> statement.  :-) I
> have no doubt that you are a very good and wonderful network
> engineer.  But,
> I am afraid, iFCP simply is a better mouse trap solving the same
> problem as
> that by iSCSI.
>
> After reading the iFCP draft carefully, I can't help but conclude that by
> connecting N-Port to IP directly iFCP usurps the fibre channel
> Extended Link
> Services and replaces class F traffic by all the wonderful stuff like EGP,
> BGP, OSPF, etc., offered by the Big Internet.  I found the iFCP effort
> overlaps with iSCSI and it provides no additional benefit.  As you have
> indicated that iFCP provides a transition into the iSCSI world of the
> future.  We must agree that initially even iSCSI storage devices will have
> FCP or SCSI devices "inside the box".  This is because the
> storage industry
> will take some time to change its infrastructure to produce iSCSI drives.
> On the other hand, the iSCSI HBAs will be available quickly -- I
> hope you do
> give me more credit in knowing the HBA world. :-)  Therefore, it is very
> easy for the Network Storage Industry to have an iSCSI target
> adapter inside
> its box to communicate with the iSCSI hosts and clients.  If iFCP is a
> viable solution, the HBA industry can produce iFCP target and
> host adapters
> just as quickly as the iSCSI adapters.  Trust me, if we can implement the
> Fibre Channel and InfiniBand specifications, we can do iFCP.  I
> do hope you
> see my logic now.
>
> Therefore, while iFCP may be a better mouse trap than FCIP, IMHO, it is
> providing the same solution as that of iSCSI.
>
> By the way, lets hear from other folks on this topic, please!
>
> Y.P. Cheng, ConnectCom Solutions.
>
>
>