Re: Proposed change to draft-ietf-6man-multi-homed-host after IESG review

[Philip Homburg <[e-mail]>]

>While I don't have a strong objection to the proposed change, I 
>am uncomfortable with what I read is a request to not even mention 
>source-destation routing.  I agree it should not be a requirement in 
>this document, but I don't see anything wrong with a forward 
>looking pointer.  Perhaps something like this (based on the NEW text):

My preference is to keep host-router interactions separate and decoupled from
routing protocols.

So I think it improves the document to avoid mentioning how routers coordinate
among each other. As such not mentioning [I-D.ietf-rtgwg-dst-src-routing]
is an improvement.

There is one thing that confuses me in this document. I think the protocol
description in this draft is fine, it is some of the surrounding wording
the I find confusing.

In my experience, if routers implement some form of dst-src routing then
existing redirects are more than enough. Linux, OSX, FreeBSD work fine in
this context.

So where hosts need to change is in the situation where routers don't use any
kind of routing protocol, are unware of each other and therefor not sending
redirects. This is the main area where this draft introduces a protocol
change.

But in this case, mentioning dst-src routing is irrelevant, because the
assumption is that the routers are not talking to each other anyhow.

So I find the new text for Section 2.2 confusing. Selecting a default router
based on advertised PIOs should make routing work even if routers don't 
speak any kind of routing protocol.