This is prolly because the IP2 cannot perform two lookups on the same
"packet" in one forwarding pass. Hence the need for tunnel PICs on GRE
tunnels, MSDP peers, multicast enabled LANs, etc. This is a guess, though,
but seems logical to me.
chris
> -----Original Message-----
> From: Jesper Skriver [mailto:jesper@skriver.dk]
> Sent: Monday, April 02, 2001 4:49 PM
> To: Lane Patterson
> Cc: 'juniper-nsp@puck.nether.net'; 'mpls-ops@mplsrc.com'
> Subject: Re: Juniper as egress LSR: to penultimate hop pop, or not?
>
>
> On Mon, Apr 02, 2001 at 01:43:22PM -0700, Lane Patterson wrote:
> >
> > It appears that if I want to use a Juniper as an egress LSR, I MUST
> > use penultimate hop popping? For example, using signalled LSPs,
> > juniper will only send RESV's with a "3" label (implicit NULL), and
> > I don't see knobs for it to use an IPv4 explicit NULL "0" label.
>
> Just like Cisco ...
>
> > Is this because the juniper forwarding path is incapable of doing
> > both a label pop and an IPv4 forwarding decision at line rate on
> > the egress LSR?
>
> Don't think so, how it was explained to me was that in
> earlier versions
> of JunOS it did send a explicit NULL "0" label, as it couldn't handle
> both MPLS and IPv4 on the same physical interface.
>
> We have a feature request open to add the knob you want, we
> want to use
> it in QoS setup's, where we carry the MPLS exp bit's to the very edge,
> and queue based on them, and leave the IP TOS field untouched.
>
> If you want it too, please push it via your account team ...
>
> /Jesper
>
> --
> Jesper Skriver, jesper(at)skriver(dot)dk - CCIE #5456
> Work: Network manager @ AS3292 (Tele Danmark DataNetworks)
> Private: FreeBSD committer @ AS2109 (A much smaller network ;-)
>
> One Unix to rule them all, One Resolver to find them,
> One IP to bring them all and in the zone to bind them.
>
This archive was generated by hypermail 2b29 : Mon Aug 05 2002 - 10:42:41 EDT