RE: iFCP as an IP Storage Work Item

[Wayland Jeong <wayland@troikanetworks.com>]

Yes. I am a bit curious as well about the merit of having a common
encapsulation method. If the thought is to standardize on a common TCP
encapsulation method (i.e. hardware assist mechanisms such as TCP options
for PDU alignment or a periodic marker in the byte stream) then I can see
the value. But if the thought is to have a common encapsulation layer below
TCP (i.e. iFCP headers) then this doesn't make any sense to me. The FCIP
draft as it stands doesn't even require such a layer for tunneling FC
frames. Such a layer would be pure overhead for FCIP.


-----Original Message-----
From: Ken Hirata [mailto:Ken.Hirata@Vixel.com]
Sent: Thursday, January 04, 2001 6:47 PM
To: Ips@Ece. Cmu. Edu
Subject: Re: iFCP as an IP Storage Work Item


David,

Why do you want to standardize a common encapsulation protocol
for FCIP and iFCP if their semantics and behavior are completely
different?  Would you want tunneling protocol implementations to
also augment certain ELSs even though it isn't necessary for tunneling
protocol operation?

If a common encapsulation protocol was defined, I believe a
negotiation protocol would be necessary to distinguish between
usage as a gateway or tunneling protocol.  After that behavior
would be completely different; the tunneling protocol doesn't
need to augment ELSs, resolve Fibre Channel Address Identifiers
to IP/N_Port IDs, or do anything else the gateway protocol
would do.

                                         Ken

David Robinson wrote:

> I am throwing in my hat to have the WG support both iFCP and
> FCIP.  From a business/customer perspective I find believable
> markets for both approaches (both of which are still speculation).
> >From a technical perspective they are similar enough that having
> one standard mechanism and one different defacto mechanism will
> cause more problems than it solves.
>
> It is clear that the semantic meaning of the two proposed protocols
> cannot be merged as they do not operate in the same plane of
> traditional stacks. However, from my reading of the two proposals
> the encapsulation mechanisms are remarkably similar even though
> their semantics are diverge significantly.  What I have started
> (before my two weeks away from work, ahhhh...) but haven't
> yet finished, is an investigation on standardizing a common
> encapsulation protocol for FC over TCP/IP. The the FCIP vs iFCP
> becomes a higher level interpretation of the semantics of the
> bits instead of also having two completely different stacks.
>
>         -David

--
Kenneth Hirata
Vixel Corporation
Irvine, CA 92618
Phone: (949) 450-6100
Email: khirata@vixel.com