RE: IPv6 Home Use to stimulate deployment over IPv4-NAT

["Christian Huitema" <huitema@windows.microsoft.com>]

> > In the IETF, we used to deal with arguments of complexity by
> > considering running code, rather than by theoretical debates.
> > We have running code for Teredo. I have used the code myself.
> > I guess this meet at least 1/2 of the requirements -- the
> > other 1/2 would be to prove Interop with an independent
> > implementation.
> 
> I agree this is a very very good argument.  I want to do the same but
I
> need find DSL Router folks to make it work.  We both have done manual
> config and 6to4.  But I am wondering if some kind of very very simple
> code patch can make it so the DSL router can assist with the process.

Teredo is not meant to be implemented in the DSL router: it is a single
host solution. The only requirement of the DSL router is that it
provides "reasonable" support for UDP-IPv4; our tests show that this is
the case for the overwhelming majority of the small routers available on
the US market. The correction needed by the other routers mostly amount
to bug fixes: a slight modification of the "5-tuple" mapping algorithm,
and a requirement to not do anything stupid after receiving an ICMP
"unreachable" error message. The same fixes are also needed to enable
the STUN scenarios that are used by IPv4/VoIP, so I expect market
pressure will lead the home router manufacturers to "do the right
thing."

If these home router manufacturers want to help deploy IPv6, then the
obvious requirement is to support 6to4, to allow configured tunnels, and
to also allow relaying of an ISP native IPv6 service. This is very much
what we are describing in the "unmanaged networks" scenarios.

-- Christian Huitema