Fast convergence on link failure in multi-homed Ethernet virtual private networks
Techniques are described for providing fast convergence in the event of a link failure in an all-active multi-homed Ethernet virtual private network. A provide edge (PE) network device may pre-configure an interface next hop and secondary next hops. The secondary next hops may be logical links to other PE network devices in the same Ethernet segment. In the event of a link failure in the interface next hop between the PE network device and a customer edge (CE) network device, the PE network device may be configured to forward data traffic to the CE network device using the secondary next hops. In the event of a link failure between the PE network device and a core network, the PE network device may be configured to send an out-of-service message to the CE network device that instructs the CE network device to stop sending traffic to the PE network device.
1. A method comprising:
pre-configuring, at a first provider edge (PE) network device, a primary route to a customer edge (CE) network device and one or more secondary routes to the CE network device, the primary route using a first communication link between the first PE network device and the CE network device, and the one or more secondary routes using one or more other respective communication links between the first PE network device and one or more other PE network devices, wherein the first PE network device and the one or more other PE network devices are configured in an all-active mode as part of an Ethernet segment of a Layer 2 virtual bridge and the CE network device is multi-homed to the first PE network device and the one or more other PE network devices;
receiving, at the first PE network device, data packets addressed to the CE network device;
detecting, by the first PE network device, a link failure for the primary route; and
forwarding, by the first PE network device, the data packets addressed to the CE network device using the one or more secondary routes in response to detecting the link failure for the primary route.
2. The method of claim 1 , wherein pre-configuring the primary route and the one or more secondary routes comprises:
configuring a next hop list, the next hop list including an interface next hop for the primary route, and one or more secondary next hops for the one or more secondary routes.
3. The method of claim 2 , further comprising:
configuring the interface next hop with a low weight in the next hop list so that all network traffic addressed to the CE network device is routed via the first communication link when the first communication link is available; and
configuring the one or more secondary next hops with a high weight in the next hop list so that network traffic is load-balanced when the first communication link is not available.
4. The method of claim 2 , wherein forwarding, by the first PE network device, the data packets addressed to the CE network device to the one or more secondary routes in response to detecting the link failure in the primary route comprises:
configuring the next hop list to select the one or more secondary next hops; and
responsive to receiving the data packets, forwarding the data packets using the one or more secondary next hops.
5. The method of claim 4 , wherein the one or more secondary next hops are multiprotocol label switching (MPLS) next hops.
6. The method of claim 1 , further comprising:
withdrawing, by the first PE network device, from a routing table for the Ethernet segment in response to detecting the link failure.
7. The method of claim 1 , wherein the data packets are known unicast traffic.
8. The method of claim 1 , further comprising:
detecting, by the first PE network device, a link failure between the first PE network device and another PE network device providing access to a core network; and
sending, by the first PE network device in response to detecting the link failure between the first PE network device and the core network, a message to the CE network device that indicates the link failure and that indicates that the data packets should not be sent to the first PE network device.
9. The method of claim 8 , wherein the Ethernet segment is connected to the CE network device in an all-active multi-homed Ethernet virtual private network (EVPN) topology.
10. The method of claim 9 , wherein the first PE network device is part of a link aggregation group (LAG) with the one or more other PE network devices.
11. The method of claim 10 , wherein sending the message comprises sending a link aggregation control protocol (LACP) out-of-service (OOS) message.
12. A network device, wherein the network device is a first provider edge (PE) network device, comprising:
a forwarding engine configured to:
pre-configure a primary route to a customer edge (CE) network device and one or more secondary routes to the CE network device, the primary route using a first communication link between the first PE network device and the CE network device, and the one or more secondary routes using one or more other respective communication links between the first PE network device and one or more other PE network devices, wherein the first PE network device and the one or more other PE network devices are configured in an all-active mode as part of an Ethernet segment of a Layer 2 virtual bridge and the CE network device is multi-homed to the first PE network device and the one or more other PE network devices;
receive data packets addressed to the CE network device, and
forward the data packets addressed to the CE network device using the one or more secondary routes in response to detecting the link failure for the primary route; and
a routing engine configured to detect the link failure for the primary route.
13. The network device of claim 12 , wherein the forwarding engine is further configured to configure a next hop list, the next hop list including an interface next hop for the primary route, and one or more secondary next hops for the one or more secondary routes.
14. The network device of claim 13 , wherein the forwarding engine is further configured to:
configure the interface next hop with a low weight in the next hop list so that all network traffic addressed to the CE network device is routed via the first communication link when the first communication link is available; and
configure the one or more secondary next hops with a high weight in the next hop list so that network traffic is load-balanced when the first communication link is not available.
15. The network device of claim 13 , wherein the forwarding engine is further configured to:
configure the next hop list to select the one or more secondary next hops; and
responsive to receiving the data packets, forward the data packets using the one or more secondary next hops.
16. The network device of claim 15 , wherein the one or more secondary next hops are multiprotocol label switching (MPLS) next hops.
17. The network device of claim 12 , wherein the routing engine is further configured to withdraw from a routing table for the Ethernet segment in response to detecting the link failure.
18. The network device of claim 12 , wherein the data packets are known unicast traffic.
19. The network device of claim 12 , further comprising:
a routing engine configured to detect a link failure between the first PE network device and another PE network device providing access to a core network, and
wherein the forwarding engine is further configured to send, in response to the routing engine detecting the link failure between the first PE network device and the core network, a message to the CE network device that indicates the link failure and that indicates that the data packets should not be sent to the first PE network device.
20. The network device of claim 19 , wherein the Ethernet segment is connected to the CE network device in an all-active multi-homed Ethernet virtual private network (EVPN) topology.
21. The network device of claim 20 , wherein the first PE network device is part of a link aggregation group (LAG) with the one or more other PE network devices.
22. The network device of claim 21 , wherein sending the message comprises sending a link aggregation control protocol (LACP) out-of-service (OOS) message.
23. A network system comprising:
a plurality of provider edge (PE) network devices connected to a customer edge (CE) network device and configured in an Ethernet segment of a Layer 2 virtual bridge, wherein each of the plurality of PE network devices are configured to:
pre-configure a primary route to the CE network device and one or more secondary routes to the CE network device, the primary route being a first communication link between a respective one of the plurality of PE network devices and the CE network device, and the one or more secondary routes being one or more other respective communication links between the respective one of the plurality of PE network devices and others of the plurality of PE network devices, wherein the first PE network device and the one or more other PE network devices are configured in an all-active mode as part of the Ethernet segment of the Layer 2 virtual bridge and the CE network device is multi-homed to the first PE network device and the one or more other PE network devices,
detect a link failure for the primary route,
receive data packets addressed to the CE network device, and
forward the data packets addressed to the CE network device using the one or more secondary routes in response to a detection of a link failure for the primary route.
24. The network system of claim 23 , wherein the each of the plurality of PE network devices is further configured to withdraw from a routing table for the Ethernet segment in response to detecting the link failure.
25. The network system of claim 23 , wherein the data packets are known unicast traffic.
26. The network system of claim 23 , wherein
the plurality of PE network devices is connected to the CE network device and a core network, and wherein each of the plurality of PE network devices are configured to:
detect a link failure between a respective one of the plurality of PE network devices and another PE network device providing access to the core network; and
send, in response detecting the link failure, a message to the CE network device that indicates the link failure and that indicates that the data packets should not be sent to the respective one of the plurality of PE network devices.
27. The network system of claim 26 , wherein the Ethernet segment is connected to the CE network device in an all-active multi-homed Ethernet virtual private network (EVPN) topology.
28. The network system of claim 27 , wherein the plurality of PE network devices are part of a link aggregation group (LAG).
29. The network system of claim 28 , wherein sending the message comprises sending a link aggregation control protocol (LACP) out-of-service (OOS) message.