Le 23/07/2012 10:33, Zhen Cao a écrit :
On Mon, Jul 23, 2012 at 3:15 PM, Brian E Carpenter
<brian.e.carpen...@gmail.com> wrote:
"It is defined exactly for this purpose to clearly indicate the
next hop address associated with the right RPO, to avoid any
inconsistency arising from sending the two in two different
options."
I asked why there were three options defined and how to relate
them. The above statement said that defining the "Next Hop
Address with Route Prefix option" can avoid inconsistency arising
from sending them in different options, NOT answering why RIO
defined in 4191 cannot solve the problem in the target scenario.
If you want to argue that the original RIO model cannot cause
inconsistency, that's fine and Behcet can answer you. I was simply
pointing out that he did give a reason. Actually that analysis
should be in the draft.
It is important and interesting topic. But I did not find the
analysis in the draft. So Hui suggested that the presentation would
talked about that. I do want to listen to that.
I also want to say that for a normal host, there is no difference
between "next hop" and "default route(r)", for a given prefix. The
draft and RFC 4191 are a bit hard to read together since the RFC
mainly refers to default route(r) and the draft mainly refers to
next hop. If they used the same terminology, the discussion would
be simpler.
We had better also include "specific route"
IMHO, it would be better to separate the documents: specific routes in
one document and default routes another document.
Reading your email exchanges, and the RFC4191, I see you mean different
things by this word 'default'.
Some meaning is: a 'default' router for a particular prefix. As to give
preference to one router when there are two for the same destination prefix.
Another meaning: _the_ default router for ::/0.
Alex
Regards, Zhen _______________________________________________ mif
mailing list mif@ietf.org https://www.ietf.org/mailman/listinfo/mif
_______________________________________________
mif mailing list
mif@ietf.org
https://www.ietf.org/mailman/listinfo/mif
