Re: draft-ietf-multi6-multihoming-requirements-06.txt

[Jay Ford <jay-ford@uiowa.edu>]

On Thu, 19 Jun 2003, Pekka Savola wrote:
> What is the number of the destinations you typically want to influence the
> selection process?
>
> Note that if it's a rather small one, this one:
>
> Default Router Preferences, More-Specific Routes, and Load Sharing
> draft-ietf-ipv6-router-selection-02.txt
>
> might help you a bit.

That might work from some cases, but not my current situation.  My net has a
single core router for each user subnet, with the path divergence occurring
in the core or border routers.  Therefore, the hosts can't choose the path
because they only have one.

> > Also, what about the interaction between (srcaddr,dstaddr) selection &
> > anti-spoofing-type filtering?  The routers do the filtering, but the host
> > does the address selection.  How does that work?  I don't want to have more
> > spoofed attacks with IPv6 than we have now with IPv4.
>
> At the moment, there is no interaction.  The routers have to allow both.
> But that's one field that in particular should be worked at.

This is operationally burdensome, especially because the burden propagates
upstream.

________________________________________________________________________
Jay Ford, Network Engineering Group, Information Technology Services
University of Iowa, Iowa City, IA 52242
email: jay-ford@uiowa.edu, phone: 319-335-5555, fax: 319-335-2951