Fw: [Mipshop] Possible Problem with IANA Considerations in FMIPv6 Draft (was: Re: Candidate 2 for Draft 8 for CARD

----- Original Message ----- 
From: "gabriel montenegro" <[email protected]>
To: "James Kempf" <[email protected]>
Cc: "Thomas Narten" <[email protected]>; <[email protected]>;
<[email protected]>
Sent: Friday, September 10, 2004 10:51 AM
Subject: Re: [Mipshop] Possible Problem with IANA Considerations in FMIPv6
Draft (was: Re: [Seamoby] Candidate 2 for Draft 8 for CARD


> James Kempf wrote:
> > The idea was that the CARD options would be included on the FMIPv6
SolPrxyRt
> > and PrxyRtAdv rather than in the CARD ICMP messages. There is some
protocol
> > for determining whether the AR supports this, see Section 4.4 of the
CARD
> > draft for how, but there is nothing in the CARD draft about piggybacking
on
> > ICMP messages other than these, though it was discussed (RtAdv was a
> > candidate).
>
> Ok, piggybacking is not specified to work on anything but FMIP ICMP
messages.
> Section 4.4 is appropriately specific about this, but 5.1.1 is not.
>
> I suggest tightening the language in 5.1.1 as follows (should be possible
> in AUTH48):
>
> OLD
>
>     Hosts and Access Routers use the Experimental ICMP type main header
>     [Ke04] when CARD protocol messages cannot be conveyed via another
>     outgoing ICMP-type message. The MN-AR interface MUST implement and
>     SHOULD use ICMP for transport, with either the CARD ICMP header or
>     piggybacked on another ICMP message.
>
> NEW
>
>     Hosts and Access Routers use the Experimental ICMP type main header
>     [Ke04] when CARD protocol messages cannot be conveyed via ICMP-type
Fast Mobile
>     IPv6 [Kood03]. The MN-AR interface MUST implement and
>     SHOULD use the CARD ICMP Type header for transport.
>     If available, the MN-AR interface MAY use the ICMP-type Fast Mobile
>     IPv6 [Kood03] for transport (section 4.4).
>
>
> Notice that the above does two things:
>
> - tighten the language about piggybacking to only that which is specified
>
> - specify the MUST implement more clearly. OLD text implied that one
>    could choose to carry CARD messages on either CARD ICMP or FMIPv6 ICMP.
>    But the latter is not a requirement for CARD (it's an informative
reference),
>    so it cannot be depended upon.
>
>
> [Hmmm... perhaps that should've been cc-ed to seamoby. Feel free to do so
if
> you think that's necessary.]
>
> -gabriel
>