Does VPNv4 address family requires MPLS?

[Alex Kog]

i am a bit confused about the concept. let say i have P2P connection from R1 to R2 with BGP peering. and lets say i am configuring 2 vrf's on R1:

• VRF1 - RD 1:1 with route-target export 1:1 and route-target import 1:1
• VRF2 - RD 1:2 with route-target export 1:2 and route-target import 1:2

of course when R2 gets the traffic he knows what routes to import to what tables since those statements are carried inside the bgp payload. now the question is whether its possible to implement such design without "mpls ip" command on the egressing interface ?

truly i couldn't find any good related info for my question.

tnx

[MHM Cisco World]

You need sure mpls for vpnv4.

It not work if mpls is disable

[rais]

MPLS is an overlay technology. In the case you mentioned, control information is there but mpls data plane is missing. 

HTH.

[Harold Ritter]

Hi @Alex Kog ,

The route target is only part of the control plane (bgp vpnv4 updates). It is not part of the data plane packets.

At the data plane level, the egress PE knows which VRF the incoming packets belong to based on the mpls label included in the packet. So without mpls enabled on the interface between the ingress and egress PE, this will not work.

Regards,

[CiscoBuddy1]

Hi Harold,

if one of device port in network between PE ad CE have no mpls ip enabled, is it still going to work? or all ports should be part of mpls ip ? because their might be different vendor which don't support mpls feature.

PE -mpls ip - A- mpls ip - B- no mpls - C - mpls - CE

[Harold Ritter]

Hi @CiscoBuddy1 ,

The "mpls ip" configuration statement is only required between the PEs. It is not required between the PE and CE and should not be configured on the interface between these two devices.

Regards,

[CiscoBuddy1]

Harold,

i have RR and all PE routers have below config talking to RR which also includes OSPF point to point between PE.
ip vrf A
rd xxx:xxx
route-target export xxx:xxx

route-target import xxx:xxx

router bgp 655xx
bgp router-id 10.x.x.x
bgp log-neighbor-changes
no bgp default ipv4-unicast
neighbor 10.x.x.x remote-as 655xx
neighbor 10.x.x.x update-source Loopback0
neighbor 10.x.x.x description RR

!
address-family ipv4
exit-address-family
!
address-family vpnv4
neighbor 10.x.x.x activate
neighbor 10.x.x.x send-community extended

exit-address-family
!
address-family ipv4 vrf A
redistribute connected
redistribute static
maximum-paths ibgp 2
exit-address-family

i can see all routes when i do show ip route vrf A and see labels out when i do show ip bgp vpnv4 all labels but i cannot ping any vrf A ip address.

[Harold Ritter]

Hi @CiscoBuddy1 ,

If the VPNv4 control plane is fine, the issue is probably with the MPLS LSP not being signaled properly between the PEs.

To verify that you can run a MPLS from one PE to the other, as follow.

ping mpls ipv4 <remote PE loopback address/32> source <local PE loopback address>

Regards,