Path selection for data traffic within a software-defined wide area network using traffic metrics

1. A method to select a route for data traffic within a software-defined wide area network (SD-WAN) topology, the method comprising:

receiving, at a branch node, Service Level Agreement (SLA) metrics of links between sites in the SD-WAN topology from at least a portion of the peers of the branch node, the SLA metrics received out-of-band in protocol extension packets through secure tunnels of the SD-WAN topology;

receiving a packet at the branch node in an established session, the packet having a source node and a destination node;

calculating performance of a plurality of possible paths from the source node to the destination node by combining SLA metrics for each link in the respective path, the source node and the destination node being within the SD-WAN topology; and

selecting a path at the branch node for the received packet from the source node to the destination node based on the calculated performance.

2. The method of claim 1 , wherein the received traffic metrics comprise existing site-to-site monitoring collected by point-to-point peers independent of receiving a poll.

3. The method of claim 2 , wherein the site-to-site monitoring comprises Service Level Agreement (SLA) monitoring.

4. The method of claim 1 , wherein calculating performance comprises performing path calculations by applying the received traffic metrics to thresholds.

5. The method of claim 4 , wherein the thresholds provide a value difference within which two paths are considered to be equal.

6. The method of claim 4 , wherein performing path calculations comprises using a Dijkstra Algorithm of thresholds with weight ranges.

7. The method of claim 6 , wherein using the Dijkstra algorithm comprises performing the Dijkstra Algorithm by punching in configured weight ranges in Shortest Path First (SPF) runs.

8. The method of claim 1 , wherein selecting the path comprises comparing traffic metrics for links of a candidate path to a cost range for other candidate paths.

9. The method of claim 1 , wherein the traffic metrics represent at least one of path latency and packet loss.

10. The method of claim 9 , wherein selecting the path comprises selecting the path having a lowest path latency.

11. The method of claim 1 , wherein selecting the path comprises applying path routing constraints.

12. The method of claim 1 , further comprising receiving an indication of a link failure in the selected path, the method further comprising selecting a second path from the source node to the destination node in response to the received link failure indication.

13. The method of claim 1 , wherein selecting the path comprises selecting the path at a control plane, the method further comprising sending the selected path to a data plane of the SD-WAN topology.

14. The method of claim 13 , further comprising the data plane receiving a list of Spoke-Hub-Hub-Spoke (SHHS) paths from a plurality of spokes and the data plane selecting the path using the list of SHHS paths.

15. A non-transitory computer-readable storage medium containing program instructions, wherein execution of the program instructions by the computer causes the computer to perform operations comprising:

receiving, at a branch node, Service Level Agreement (SLA) metrics of links between sites in the SD-WAN topology from at least a portion of the peers of the branch node, the SLA metrics received out-of-band in protocol extension packets through secure tunnels of the SD-WAN topology;

receiving a packet at the branch node in an established session, the packet having a source node and a destination node;

calculating performance of a plurality of possible paths from the source node to the destination node by combining SLA metrics for each link in the respective path, the source node and the destination node being within the SD-WAN topology; and

selecting a path at the branch node for the received packet from the source node to the destination node based on the calculated performance.

16. The medium of claim 15 , wherein calculating performance comprises performing path calculations by applying the received traffic metrics to thresholds, wherein the thresholds provide a value difference within which two paths are considered to be equal.

17. The medium of claim 15 , wherein the traffic metrics represent at least one of path latency and packet loss and wherein selecting the path comprises selecting the path having a lowest path latency.

18. A branch node comprising:

a communications interface configured to receive a packet at the branch node in an established session, the packet having a source node and a destination node;

a processor coupled to the communications interface and to a memory, and configured to implement:

a path metrics monitor configured to receive Service Level Agreement (SLA) metrics of links between sites in the SD-WAN topology from at least a portion of the peers of the brand node, the SLA metrics received at the branch node out-of-band in protocol extension packets through secure tunnels of the SD-WAN topology; and

a topology and path selection engine configured to calculate performance of a plurality of possible paths from the source node to the destination node by combining SLA metrics for each link in the respective path, the source node and the destination node being within the SD-WAN topology, and to select a path at the branch node from the source node to the destination node based on the calculated performance.

19. The branch node of claim 18 , wherein the received traffic metrics comprise existing site-to-site monitoring collected by point-to-point peers independent of receiving a poll.

20. The branch node of claim 18 , wherein the topology and path selection engine is further configured to compare traffic metrics of links of a candidate path to a cost range for other candidate paths.