Understanding P2MPs LSP for the EVPN Inclusive Provider Tunnel

An EVPN instance comprises Customer Edge devices (CEs) that are connected to Provider Edge devices (PEs) to form the edge of the MPLS infrastructure. A CE may be a host, a router, or a switch. The PEs provide virtual Layer 2 bridged connectivity between the CEs. There may be multiple EVPN instances in the provider's network. The PEs may be connected by an MPLS Label Switched Path (LSP) infrastructure, which provides the benefits of MPLS technology, such as fast reroute, resiliency, etc.

The PEs may also be connected by an IP infrastructure, where IP/GRE (Generic Routing Encapsulation) tunneling or other IP tunneling can be used between the PEs. Here we understand the MPLS LSPs as the tunneling technology designed to be extensible to IP tunneling as the Packet Switched Network (PSN) tunneling technology.

Starting in Junos OS Release 18.2R1 onwards, Junos OS provides the ability to configure and signal a P2MP LSP for the EVPN Inclusive Provider Tunnel for BUM traffic. P2MP LSPs can provide efficient core bandwidth utilization by using the multicast replication only at the required nodes instead of ingress replication at the ingress PE.

You can configure and signal a P2MP LSP for the EVPN Inclusive Provider Tunnel in the PMSI Attributes of the Inclusive Multicast Ethernet Tag Route. The P2MP tunnel is used for forwarding Broadcast, unknown Unicast, and Multicast (BUM) packets at the ingress PE. When representing the PMSI attributes in the Inclusive Multicast Ethernet Tag Route, a transport label is not represented for P2MP Provider Tunnels, except in the case where Aggregation is used. The transport label is used to identify the EVI at the egress PE.

When the P2MP Provider Tunnels are used, the ESI labels for Split Horizon are assigned upstream instead of downstream. The label that the egress PE receives as a Split Horizon label is allocated by the ingress PE. Since each upstream PE cannot allocate the same label, the label does not identify the ESI and must be referred in the context label space of the ingress PE. The transport label uniquely identifies the ingress PE and the split horizon label is identified by tranport-label or SH-label tuple.

At the ingress PE, an upstream assigned ESI label is allocated and signaled for Split Horizon function. This label is added to the BUM traffic that originates from the given Ethernet Segment. By default, an egress may receive traffic with an ESI label that is not configured on this PE because of P2MP tunnels and upstream assigned labels.

For E-tree, the leaf label is assigned upstream and its function is similar to the SH label. The ingress PE allocates the leaf label and represents it in the E-tree extended community for the Ethernet A-D per ES route. When the Leaf PE acts as an ingress, it adds the leaf label for BUM traffic. An egress PE acting as a leaf discards the traffic which contains the leaf label. Egress PEs acting as a root pops the leaf label and forwards the traffic. Since leaf labels are upstream assigned, they may not be unique because multiple Leaf PEs may have allocated the same leaf label. Therefore, the leaf label must be identified by the (tranport-label, Leaf-label) tuple.