Interworking of EVPN-IFL and IPVPN in the same VPRN

This section describes the SR OS interworking details for BGP owners in the same VPRN. The behavior is compliant with draft-ietf-bess-evpn-ipvpn-interworking.

A VPRN service can be configured to support EVPN-IFL and IPVPN simultaneously. For example, the following MD CLI excerpt shows a VPRN service configured for EVPN-IFL (vprn>bgp-evpn context) and IPVPN (vprn>bgp-ipvpn context):

[ex:/configure service vprn "vprn-ipvpn-evpnifl-AL-80"]
A:admin@PE-2# info
    admin-state enable
    service-id 80
    customer "1"
    bgp-evpn {
        mpls 1 {
            admin-state enable
            route-distinguisher "192.0.2.2:80"
            vrf-target {
                community "target:64500:80"
            }
            auto-bind-tunnel {
                resolution any
            }
        }
    }
    bgp-ipvpn {
        mpls {
            admin-state enable
            route-distinguisher "192.0.2.2:80"
            vrf-target {
                community "target:64500:80"
            }
            auto-bind-tunnel {
                resolution any
            }
        }
    }
    interface "lo0" {
        loopback true
        ipv4 {
            primary {
                address 2.2.2.2
                prefix-length 32
            }
        }
    }

When EVPN-IFL and IPVPN are both enabled on the same VPRN, the following rules apply:

• IPVPN and EVPN-IFL routes are treated by BGP as separate routes; that is, the selection is done at route table level and not at the BGP level.

• At the route table level, IPVPN and EVPN-IFL routes may have the same route table preference (by default, this is 170 for both routes), route selection between IPVPN and EVPN-IFL routes is based on regular BGP path selection.

• ECMP across IPVPN and EVPN-IFL routes for the same prefix is not supported. When vprn>ecmp is configured to 2 or greater, installing multiple equal cost next hops for the same prefix in the VPRN route table is only supported within the same route owner, IPVPN or EVPN IFL.

• When EVPN-IFL and IPVPN are both enabled in the same VPRN, by default, EVPN-IFL routes are exported into IPVPN and the other way aroud (CLI configuration is not required).

• The configure>service>vprn>allow-export-bgp-vpn command is relevant within the same owner (either IPVPN or EVPN-IFL) and works as follows:

  • The command re-exports a received EVPN-IFL route into an EVPN-IFL route to a different peer.

  • The command also re-exports a received IPVPN route into an IPVPN route.

  • If EVPN-IFL and IPVPN are both configured in the same VPRN, an EVPN-IFL route is automatically re-exported into an IPVPN route. Conversely, an IPVPN route is re-exported into an EVPN-IFL. This is true unless export policies prevent the automatic re-export function.