MPLS Configuration Guidelines

 

When configuring MPLS on QFX Series devices or on EX4600, note that the number of IP prefixes supported depends on the specific platform being used. See the scale specifications in the data sheet of your device for additional information.

  • We recommend the following:

    • If your ingress provider edge (PE) switch needs to support more than 8000 external IP prefixes, use a larger capacity device as an ingress PE switch.

    • If you use a switch as a route reflector for BGP labeled routes, use it as a dedicated route reflector (that is, the switch must not participate in managing data traffic).

    • If you use a switch as a PE switch or as a route reflector for BGP labeled routes, configure routing policies on the PE switch and the route reflector to filter external IP routes from the routing table.

      The configuration example for a routing policy named fib_policy (at the [edit policy-options and [edit routing-options hierarchy levels) to filter BGP labeled routes from the inet.0 routing table is given below:

  • Packet fragmentation using the allow-fragmentation statement at the [edit protocols mpls path-mtu] hierarchy level is not supported on QFX Series devices or on the EX4600 switch. Therefore, you must ensure that the maximum transmission unit (MTU) values configured on every MPLS interface is sufficient to handle MPLS packets. The packets whose size exceeds the MTU value of an interface will be dropped.