Hi, > > I have a number of EX series switches carrying vlans back to a central MX > > 480. > > > > I have remote customers running bgp with their sessions terminating on the > > MX. Currently the whole system operates in layer two. > > > > I would like to move to a partial mesh deployment and use MPLS-CCC to carry > > the EX customer traffic back to the MX. > > > > The problem is, at the MX interface the traffic will be inside the tunnel. > > I ...
> > I spoke with Jay off-list about this, and thought there might be a clever > way of doing this with a logical-tunnel, with the MPLS-CCC tunnel ending on > a logical tunnel interface, then hooking that into a regular logical > interface with an inet address for BGP peering. > I think I have tried something like this. I don't use it though. I still do not have EX'es. You should have an L3 card on MX and you must sacrifice a 10GE port for tunneling or use an internal tunnel port on card with 1GE interfaces. Than try something like this: lt-4/3/0 { description "tunnel interface"; unit 3 { description "PeterP tunnel L2"; encapsulation vlan-ccc; vlan-id 16; peer-unit 4; } unit 4 { description "PeterP tunnel L3"; encapsulation vlan; vlan-id 16; peer-unit 3; family inet { address 195.150.0.129/30; } } L2V-c480pd-c480l { instance-type l2vpn; interface lt-4/3/0.3; route-distinguisher 195.150.0.243:2000; vrf-import L2V-pp-import; vrf-export L2V-pp-export; protocols { l2vpn { encapsulation-type ethernet-vlan; site c480pd { site-identifier 1; interface lt-4/3/0.3 { remote-site-id 2; } } } } } policy-options policy-statement L2V-pp-export term def { then { community add L2V-pp; accept; } } term reject { then reject; } ------------------- Jurek _______________________________________________ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp