Adaptive location-based SD-WAN policies

1. A method comprising:

receiving, by a software-defined networking in a wide area network (SD-WAN) system, WAN link characterization data for a plurality of WAN links of the SD-WAN system over a time period; and

for each site of a plurality of sites of the SD-WAN system,

generating, by the SD-WAN system, a local policy for the site based on a global policy comprising one or more rules and applicable to all of the plurality of sites and a machine learning model trained with the WAN link characterization data for the plurality of WAN links, and

providing the local policy to an SD-WAN edge device of the site.

2. The method of claim 1 , wherein generating the local policy for the site comprises generating, based on the machine learning model, one or more customized versions of rules of the global policy for inclusion in the local policy.

3. The method of claim 1 , wherein generating the local policy includes generating the local policy based on a location corresponding to the SD-WAN edge device.

4. The method of claim 1 , wherein the machine learning model is trained with policy parameters being applied at a time the WAN link characterization data was collected.

5. The method of claim 1 , wherein the WAN link characterization data includes one or more of jitter, latency, time to first packet, packet loss, and maximum transmission unit (MTU) for each of the plurality of WAN links.

6. The method of claim 5 , further comprising determining, by the SD-WAN system, the packet loss based on determining a number of packets transmitted by the SD-WAN edge device that did not reach destination devices specified in the packets.

7. The method of claim 1 , further comprising training the machine learning model based on historical WAN link characterization data received over a time period.

8. The method of claim 1 , wherein generating the local policy comprises generating a service policy for each service of one or more services.

9. The method of claim 8 , wherein the service policy specifies selection criteria for at least one of a network path or a service level agreement parameter.

10. The method of claim 1 , wherein the local policy includes a rule to determine whether or not to reassign an application or service from a first WAN link associated with the SD-WAN edge device to a second WAN link associated with the SD-WAN edge device, wherein the rule includes a Quality of Experience (QoE) parameter or a cost factor associated with each of the first WAN link and the second WAN link.

11. The method of claim 10 , further comprising optimizing the QoE parameter and the cost factor for each of the first WAN link and the second WAN link.

12. The method of claim 1 , wherein the local policy includes a rule to select a network path based on a path signature associated with the network path corresponding to one or more service network characteristics.

13. An SD-WAN system comprising:

a network analysis system comprising processing circuitry configured to:

receive WAN link characterization data for a plurality of WAN links of the SD-WAN system over a time period, and

for each site of a plurality of sites of the SD-WAN system,

generate a local policy for the site based on a global policy comprising one or more rules and applicable to all of the plurality of sites and a machine learning model trained with the WAN link characterization data for the plurality of WAN links, and

provide the local policy to an SD-WAN edge device of the site; and

the SD-WAN edge device comprising processing circuitry configured to:

receive the local policy, and

assign, based on the local policy, a service or application to a WAN link.

14. The SD-WAN system of claim 13 , wherein the network analysis system is configured to generate, based on the machine learning model, one or more customized versions of rules of the global policy for inclusion in the local policy.

15. The SD-WAN system of claim 13 , wherein the network analysis system is configured to generate the local policy based on a location corresponding to the SD-WAN edge device.

16. The SD-WAN system of claim 13 , wherein the WAN link characterization data includes one or more of jitter, latency, time to first packet, packet loss, and maximum transmission unit (MTU) for each of the plurality of WAN links.

17. The SD-WAN system of claim 13 , wherein the machine learning model is trained based on historical WAN link characterization data received over a time period.

18. The SD-WAN system of claim 13 , wherein the local policy includes a rule to determine whether or not to reassign an application or service from a first WAN link associated with the SD-WAN edge device to a second WAN link associated with the SD-WAN edge device, wherein the rule includes a Quality of Experience (QoE) parameter or a cost factor associated with each of the first WAN link and the second WAN link.

19. An SD-WAN edge device comprising:

one or more processors; and

a memory storing instructions, that when executed, cause the one or more processors to:

receive, from a network analysis system, a machine learning model trained with WAN link characterization data for a plurality of WAN links of a plurality of sites,

generate a local policy for the SD-WAN edge device based on the machine learning model, and

assign, based on the local policy, a service or application of the SD-WAN edge device to a WAN link.

20. The SD-WAN edge device of claim 19 , wherein the local policy includes a rule to determine whether or not to reassign the service or application from a first WAN link associated with the SD-WAN edge device to a second WAN link associated with the SD-WAN edge device, wherein the rule includes a Quality of Experience (QoE) parameter or a cost factor associated with each of the first WAN link and the second WAN link.

21. A method comprising:

receiving, by one or more processors of an SD-WAN edge device from a network analysis system, a machine learning model trained with WAN link characterization data for a plurality of WAN links of a plurality of sites;

generating a local policy for the SD-WAN edge device based on the machine learning model; and

assigning, based on the local policy, a service or application of the SD-WAN edge device to a WAN link.

22. The method of claim 21 , wherein the local policy includes a rule to determine whether or not to reassign the service or application from a first WAN link associated with the SD-WAN edge device to a second WAN link associated with the SD-WAN edge device, wherein the rule includes a Quality of Experience (QoE) parameter or a cost factor associated with each of the first WAN link and the second WAN link.