[j-nsp] (SOLVED) L3VPN and VPLS units on same physical IQ2 PIC port

[David Ball]

  Turns out I didn't need to do pushing and popping after all.  I was
operating under false assumptions with respect to the way our Nortel
switches and Juniper routers would interact.

David


On 09/11/2007, David Ball <davidtball at gmail.com> wrote:
>   Hi Peter.   The customer doesn't plug directly into an IP interface.
>  They must first traverse a layer 2 network, and eventually cross a
> shared q-in-q trunk where it's imperative that VLANs are unique per
> customer (which we have to manage, to ensure VLANs are unique).
> Return traffic must be tagged with the correct VLAN prior to sending
> back across the Q-in-Q trunk to ensure it ends up going to the correct
> endpoint via the right VLAN.  I suppose the 'pop' on input to the
> Juniper might not be absolutely necessary, but the 'push' is required
> when leaving the Juniper to traverse the Q-in-Q trunk.
>
> David
>
>
> On 08/11/2007, Peter Van Oene <pvanoene at juniper.net> wrote:
> > Hi David,
> >
> > VLAN operations are not designed for IP interfaces.  I'm not sure
> > exactly why you need to manipulate the layer two header for a layer
> > three interface?
> >
> > Pete
> >
> > > -----Original Message-----
> > > From: juniper-nsp-bounces at puck.nether.net [mailto:juniper-nsp-
> > > bounces at puck.nether.net] On Behalf Of David Ball
> > > Sent: Thursday, November 08, 2007 11:51 AM
> > > To: juniper-nsp at puck.nether.net
> > > Subject: [j-nsp] L3VPN and VPLS units on same physical IQ2 PIC port
> > >
> > >   I have a T640 wit an 8-port IQ2 Gig PIC running 8.2R1.7.  One port
> > > on this PIC is a Q-in-Q trunk to a Nortel switch transporting traffic
> > > for various customers (hence the Q-in-Q).  Thus far, all customers on
> > > said trunk have been VPLS customers.
> > >   I'm now trying to add an L3VPN/VRF customer interface to the trunk
> > > but am running into problems. Config is shown below.  The Nortel
> > > switch has to q-in-q the traffic for transmission across the trunk,
> > > but we them immediately want to pop the outer tag upon receipt at the
> > > Juniper interface, at which point they're dumped into their VRF (it's
> > > layer2 across the Nortel network).  Similarly, when sending from
> > > Juniper to Nortel, we'll need to push a tag.  I can't seem to use the
> > > (input|output)-vlan-map statements though, as they require vlan-ccc,
> > > extended-vlan-ccc, vlan-vpls or extended-vlan-vpls encapsulations
> > > (none of which will work for L3VPN afaik).
> > >   I really feel like I'm missing something, or doing something dumb.
> > > I'm going to blame it on being sleep deprived due to having a newborn,
> > > but any help you can provide would be stellar.
> > >
> > >
> > > david at router# show interfaces ge-7/0/5
> > > description "Q-in-Q trunk to Nortel switch"
> > > flexible-vlan-tagging;
> > > encapsulation flexible-ethernet-services;
> > > unit 100 {
> > >     encapsulation vlan-vpls;
> > >     vlan-id 100;
> > >     input-vlan-map pop;
> > >     output-vlan-map push;
> > >     family vpls;
> > > }
> > > unit 789 {
> > >     vlan-id 789;
> > >     input-vlan-map pop;
> > >     output-vlan-map push;
> > >     family inet {
> > >         address 172.16.1.1/24;
> > >     }
> > > }
> > > _______________________________________________
> > > juniper-nsp mailing list juniper-nsp at puck.nether.net
> > > https://puck.nether.net/mailman/listinfo/juniper-nsp
> >
>