EVPN
-
Support for blocking asymmetric EVPN Type 5 routes (MX960, QFX5110, and QFX10002)—Starting in Junos OS Release 22.2R1, you can configure the local node to reject asymmetric EVPN Type 5 routes on EVPN-VXLAN networks. The local node examines the incoming EVPN Type 5 route packets and rejects the route when the virtual network identifier (VNI) in the ingress route differs from the locally configured VNI.
To block asymmetric EVPN Type 5 routes, include the
reject-asymmetric-vni
statement at the[edit routing-instance routing-instance-name protocols evpn ip-prefix-routes]
hierarchy level.[See EVPN Type 5 Route with VXLAN encapsulation for EVPN-VXLAN and ip-prefix-routes.]
-
Automatically derived ESI configuration (MX Series, QFX5100, QFX5110, QFX5120-32C, QFX5120-48T, QFX5120-48Y, QFX10002, QFX10002-60C, QFX10008, and QFX10016)—In the current implementation, Junos OS derives the Ethernet segment identifier (ESI) from the system ID and the administrative key on the local multihomed provider edge (PE) device that is a part of the LACP link (actor). Starting in Junos OS Release 22.2R1, you can also configure the multihomed devices on an EVPN-VXLAN network to automatically generate the ESI from:
-
The system ID and administrative key on the remote customer edge (CE) device (partner).
-
The locally configured
mac
and local discriminator values.
To automatically derive the ESI using the system ID and administrative key on the remote CE device, include
type-1-lacp
at the[edit interfaces aeX aggregated-ether-options lacp auto-derive]
hierarchy level.To automatically derive the ESI using locally configured values, configure
mac
andlocal-discriminator
at the[edit interfaces aeX aggregated-ether-options lacp auto-derive type-3-system-mac]
hierarchy level.[See Understanding Automatically Generated ESIs in EVPN Networks.]
-
-
EVPN active/active redundancy, aliasing, and mass MAC withdrawal (MX Series and vMX)—Starting in Junos OS Release 22.2R1, the listed devices support EVPN active/active redundancy, aliasing, and mass MAC withdrawal, integrated with VXLAN in the data plane. These features provide resilient inter-data center connectivity to the established Data Center Interconnect (DCI) technologies. This new support builds an end-to-end DCI solution by integrating EVPN active/active multicast with DP VXLAN.
Use existing configuration statements to configure active/active redundancy at the ESI level on the loopback (lo0) interface. Include lo0 as the virtual tunnel endpoint (VTEP) interface in the routing instance.
-
Support for BGP domain path attribute in EVPN Type 5 and IPVPN routes on gateway provider edge (PE) devices (MX480, MX960, and vMX)—Starting in Junos OS Release 22.2R1, you can configure the BGP D-PATH attribute on your gateway PE device to add a domain ID to BGP routes. The BGP D-PATH attribute enables gateway PE devices to identify the domains through which EVPN IP prefix routes and IP-VPN routes have traversed. Additionally, the BGP D-PATH attribute uses its path selection algorithm to install the best routes in gateway PE device IP virtual routing and forwarding (VRF) tables, which prevents routes from looping.
To configure the BGP D-PATH attribute on all of your configured virtual routing and forwarding instances on the gateway PE device, enable the
uniform-propagation-mode
statement with the domain-id option in the[edit routing-instances]
hierarchy. When you configure the statement, it is also enabled at the global[edit routing-options uniform-propagation-mode domain-id type]
hierarchy level. Use the <type> variable to specify the type of Inter-Subnet Forwarding (ISF) and Subsequent Address Family Identifiers (SAFIs) you use to advertise IP prefix routes.[See uniform-propagation-mode.]