RFC4890 question (was: RE: New Version Notification for draft-jiang-v6ops-incremental-cgn)

["Templin, Fred L" <[e-mail]>]

I guess I have to ask again. Can the RFC4890 authors please
comment on the RFC4890, Section A.2 text questions asked in
my previous message (below)?

Fred
[e-mail]

> -----Original Message-----
> From: Templin, Fred L
> Sent: Thursday, June 04, 2009 8:44 AM
> To: Mohacsi Janos; Brian E Carpenter
> Cc: JiangSheng 66104; Seiichi Kawamura; Gert Doering; Fleischman,
Eric; Re'mi Despre's;
> [e-mail]; [e-mail]; Russert, Steven W
> Subject: RE: New Version Notification for
draft-jiang-v6ops-incremental-cgn
> 
> > > Do you think the draft needs to do any more than recommending to
> follow
> > > the recommendations in 4890? It seems as if that should be
> sufficient.
> >
> > This should be enough.
> 
> Regarding the 4890, there is advice in that document that I
> would like to understand better. In Appendix A.2, it says:
> 
>    "If a network chooses to generate packets that are no larger than
the
>    Guaranteed Minimum MTU (1280 octets) and the site's links to the
>    wider Internet have corresponding MTUs, Packet Too Big messages
>    should not be expected at the firewall and could be dropped if they
>    arrive."
> 
> But, that would seem to be in conflict with the text of
> Section 5 of RFC2460, where the reader is informed that
> the network can return PTB messages reporting MTU values
> smaller than 1280 if a protocol translator is in the path.
> The host should then react to these PTBs by inserting a
> fragment header with (MF=0; Offset=0) in subsequent packets.
> With the (RFC4890, Appendix A.2) text however, this behavior
> is suppressed.
> 
> Where did the RFC4890 text come from? And, why is it there?
> 
> Thanks - Fred
> [e-mail]