I work for a small SP, we have an existing plain IP core, and I am trying to implement MPLS.
My approach has been to enable LDP and MP-BGP on a few links and watch what happens first, and implement VRF's later.
When I enable LDP and MP-BGP, some applications/customers on that link perform badly or not at all. Some customers report their IPSec trafic stops working, where other customers with IPSec on those links have no problem. Some customers just using port 80 with no IPSec have poor web response times.
When I remove the MPLS interface commands, everything is ok again. MP-BGP is still configured and causes no problem.
The config I am using is this:
ip cef
mpls ip
!
router bgp 1
no synchronization
bgp log-neighbor-changes
neighbor 1.1.1.1 remote-as 1
neighbor 1.1.1.1 update-source Loopback1
neighbor 1.1.1.1 route-reflector-client
no auto-summary
!
address-family vpnv4
neighbor 1.1.1.1 activate
neighbor 1.1.1.1 send-community extended
neighbor 1.1.1.1 route-reflector-client
exit-address-family
int FastEthernet0/0.500
mpls label protocol ldp
tag-switching mtu 1512
tag-switching ip
This config is pretty basic. IOS is 12.3(12).
I also tried changing the "tag-switching mtu 1512" to a MTU of 1800. It made no difference.
Some questions:
a) Shouldn't existing traffic not be affected by MPLS, unless I put it in a VRF first? (which I am not)
b) Am I correct in assuming existing IP traffic will not be labelled with MPLS with this config?
c) Do I need MP-BGP as well as LDP?
I need to get MPLS working in parallel with our existing IP customers. I only want to use MPLS for customers who want VPN.
What am I doing wrong? I need to get a VRF working soon!
