Subject: Re: How to use properly ipv6 autoconf over a router interface?
To: Konstantin KABASSANOV <Konstantin.Kabassanov@lip6.fr>
From: JINMEI Tatuya / =?ISO-2022-JP?B?GyRCP0BMQEMjOkgbKEI=?= <jinmei@isl.rdc.toshiba.co.jp>
List: tech-net
Date: 05/16/2006 17:17:47
>>>>> On Mon, 1 May 2006 20:18:49 +0200, 
>>>>> "Konstantin KABASSANOV" <Konstantin.Kabassanov@lip6.fr> said:

> You are right... All specifications talk only about "hosts" and not
> "routers" auto-configuration. Probably because traditional network
> architectures do separate clearly hosts from routers. But what about more
> hybrid architectures where some routers are hosts for other ones...? I agree
> this is an issue to be addressed elsewhere.    

(I've been accidentally a non-subscriber for a while, so this may be
an out-of-date response, but anyway)

Apparently you are talking about a behavior of a half-router/half-host
node, that is, a node that works as a router on some subset of its
interfaces and works as a host on the other set of interfaces.  If so,
in my understanding the protocol standard does not prohibit such a
behavior.  We in fact discussed whether we should explicitly describe
it in revised versions of RFC2461 and RFC2462, but finally decided not
to do so because there did not seem enough operational experiences
(and due to the "conservative" nature of the revision).

I also heard that the (MS) Windows IPv6 stack support this behavior,
but I've not used that feature.

The KAME/BSD currently does not support this behavior on purpose, due
to concerns about possible implementation complexity.  I don't know if
the concern is a FUD or there is any essential issue due to the BSD's
networking architecture that makes it impossible.

					JINMEI, Tatuya
					Communication Platform Lab.
					Corporate R&D Center, Toshiba Corp.
					jinmei@isl.rdc.toshiba.co.jp