[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Comments on draft-wbeebee-ipv6-cpe-router-01.txt
Sorry, I had two typos in the email below. Here are the fixed sentences
for typos shown in quotes.
[Asymmetrical routing may be likely for the CPE Router because besides
the default "route", ...]
and
[for the direction of "SP" to the CPE Router]
Hemant
-----Original Message-----
From: owner-v6ops@ops.ietf.org [mailto:owner-v6ops@ops.ietf.org] On
Behalf Of Hemant Singh (shemant)
Sent: Wednesday, July 16, 2008 10:14 AM
To: Antonio Querubin
Cc: Ole Troan; Stark, Barbara; v6ops@ops.ietf.org
Subject: RE: Comments on draft-wbeebee-ipv6-cpe-router-01.txt
Antonio,
Please see in line below.
-----Original Message-----
From: Antonio Querubin [mailto:tony@lava.net]
Sent: Wednesday, July 16, 2008 9:19 AM
To: Hemant Singh (shemant)
Cc: Ole Troan; Stark, Barbara; v6ops@ops.ietf.org
Subject: RE: Comments on draft-wbeebee-ipv6-cpe-router-01.txt
On Wed, 16 Jul 2008, Hemant Singh (shemant) wrote:
> Even if the CPE Router has a default router out the WAN interface to
> the SP router, RPF may support a knob where "Allow default route to
> match when checking source address" is not allowed. One can configure
> a router to disable checking the default route to match src-addr by
RPF.
Aren't we talking about a single-homed router?
<hs
Yes.
</hs>
On a mass-market consumer router with one logical WAN interface
connected to a service provider's edge, how likely would such a knob be
used?
Asymmetrical routing is proably not a significant factor.
<hs>
Folks may think the knob is less likely to be used. Asymmetrical routing
may be likely for the CPE Router because besides the default router, as
our draft says, RIPv6 may optionally run between the CPE Router and the
SP aggregation router. You see, all folks in this discussion have said,
RIPv6 is not needed. But when they said it's not needed they are talking
about the upstream direction from the CPE Router to the SP. It's the
downstream direction that is interesting. I already said yesterday, for
the direction of RP to the CPE Router, how does the SP aggregation
router know the routes to the PD's assigned to the CPE Router? In the
March IETF, folks presented one solution (to v6ops) on how does the SP
aggregation router learn routes to PD's in the downstream. IETF has
still not standardized one solution for this problem nor am I sure if
IETF would like to standardize such a solution. In the meantime, we are
saying one may use RIPv6 to solve the problem. The other alternative is
to have the aggregation router support DHCPv6 snooping for the IA_PD and
then inject a route for the PD.
</hs>
Hemant
Antonio Querubin
whois: AQ7-ARIN