Re: MPLS behavior in multipoint interfaces

Just because the ints run MPLS doesn't mean it won't accept IP
packets. Untagged packets are accepted and processed as usual.

How you get to an untagged packet is beyond the scope of this reply
and assumes the traffic originates or was routed to the MPLS core
table (global table).

Is this VPN traffic or something else that runs over a tunnel LSP?

On 5/16/10, Jorge Cortes <jorge.cortes.cano_at_gmail.com> wrote:
> Hi team,
>
> Today I ran into a "weird" behavior when enabling MPLS in a multipoint
> environment. The summarized topology is as follows:
>
>
> R1------------| |-------------------R3
> | FRAME-RELAY |
> R2------------| |-------------------R4
>
> Physical interfaces (multipoint) are used on all routers and there is
> a full mesh of PVCs among all routers. I enabled MPLS on the
> interfaces of R4 and R2, but not on R3 and R1, thus traffic sent from
> behind R1 or R3 that needs to reach a destination behind R2 or R4 will
> be sent untagged. Since R2 and R4 interfaces are MPLS enabled I would
> expect them to drop any untagged packets arriving at them; however
> traffic traverses successfully and is able to get to its destination,
> so this keeps thinking what is the expected behavior of MPLS in
> multipoint interfaces. Any pointers are really appreciated.
>
> Thanks,
> Jorge
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
>
>
>
>
>
>

-- 
Sent from my mobile device
William McCall, CCIE #25044
Blogs and organic groups at http://www.ccie.net