RE: NOTE: WG Last Call: draft-ietf-v6ops-mech-v2-01.txt

["Pascal Thubert (pthubert)" <pthubert@cisco.com>]

Itojun:

The example that Dave proposed, a MIPv6 (Nemo) router, has exactly that
behavior when not at Home: It acts as a Host at ND level on the roaming
interface to discover an attachment router and to autoconf a CareOf. 

This gives you at least one case where it's an interesting thing to do.

Fred: in our initial implementation, we've been through this but tried
to save in the IPv6inv4 tunnel, using some automatic tunneling based on
6to4. In that case, there's no tunnel to send RS over.

What we did fits with what Dave suggested (if I understand correctly).
Rely on a correct IPv4 or IPv6 administrative MTU, and send PMTU ICMPs
back based on RFC 2473.

For more details about how that works (but nothing much about MTU since
its RFC 2473) you may want to have a look at Doors (but I know Pekka
won't). Goes through all sorts of PAT/NATs.
http://www.ietf.org/internet-drafts/draft-thubert-nemo-ipv4-traversal-01
.txt


Pascal

> -----Original Message-----
> From: owner-v6ops@ops.ietf.org [mailto:owner-v6ops@ops.ietf.org] On
Behalf Of Jun-ichiro
> itojun Hagino
> Sent: mardi 25 novembre 2003 02:11
> To: ftemplin@iprg.nokia.com
> Cc: dthaler@windows.microsoft.com; v6ops@ops.ietf.org
> Subject: Re: NOTE: WG Last Call: draft-ietf-v6ops-mech-v2-01.txt
> 
> > Nodes may act as hosts on some interfaces and routers on other
> > interfaces. They may also act in different turns as routers and
hosts
> > on the same interface. I see nothing in RFC 2461 that precludes such
> > "hybrid" nodes from sending RSs and processing RAs.
> >
> > In fact, the second-to-last paragraph of RFC 2461, section 6.2.7
> > ("Router Advertisement Consistency") says:
> >
> >    "Note that it is not an error for different routers to advertise
> >    different sets of prefixes.  Also, some routers might leave some
> >    fields as unspecified, i.e., with the value zero, while other
routers
> >    specify values.  The logging of errors SHOULD be restricted to
> >    conflicting information that causes hosts to switch from one
value to
> >    another with each received advertisement."
> >
> >
> > So, if router A has good reason to believe that router B advertises
> > a different set of prefixes, I see nothing unusual about A sending
> > RSs to B (and getting RAs back from B) to discover the different
> > prefixes.
> 
> 	i guess i mentioned it in the past, if we allow routers to
advertise
> 	different set of prefixes (which is the case now with RFC2462)
and
> 	allow routers to autoconfigure themselves with RA, we go into
> 	bootstraap problem when router reboots.  i consider
autoconfiguring
> 	router is a dangerous idea, and "hybrid router" idea is too.
> 
> itojun