Dynamic path switchover decision override based on flow characteristics

1. A method comprising:

receiving, at a device in a network, an alert regarding a performance change of a first path in the network;

receiving, at the device, a switchover policy for a particular type of traffic in the network;

in response to receiving the alert, determining, by the device, a predicted effect of directing a traffic flow of the particular type of traffic from the first path in the network to a second path in the network, the predicted effect corresponding to one or more predicted performance differences between the first path and the second path;

determining, by the device, whether the predicted effect of directing the traffic flow to the second path would violate the switchover policy;

causing, by the device, the traffic flow to be routed via the second path in the network, based on a determination that the predicted effect of directing the traffic flow to the second path would not violate the switchover policy for the particular type of traffic; and

preventing, by the device, switchover of the traffic flow to the second path in the network, based on a determination that the predicted effect of directing the traffic flow to the second path would violate the switchover policy for the particular type of traffic.

2. The method as in claim 1 , wherein the predicted effect comprises a predicted amount of packet reordering that would occur were the traffic flow directed to the second path, and wherein the switchover policy indicates a threshold amount of acceptable packet reordering for the particular type of traffic.

3. The method as in claim 1 , wherein the predicted effect comprises a predicted decrease in bandwidth for the traffic flow were the traffic flow routed via the second path, and wherein the switchover policy indicates a threshold amount of acceptable decrease in bandwidth for the particular type of traffic.

4. The method as in claim 1 , wherein the predicted effect comprises a predicted duration for the traffic flow, and wherein the switchover policy indicates an acceptable flow duration in order for the traffic flow to be directed to the second path.

5. The method as in claim 1 , further comprising:

generating, by the device, a machine learning model that models the particular type of traffic; and

using, by the device, the machine learning model to determine the predicted effect of directing the traffic flow to the second path.

6. The method as in claim 1 , further comprising:

determining, by the device, an estimated performance of the second path; and

using, by the device, the estimated performance of the second path to determine the predicted effect of directing the traffic flow to the second path.

7. The method as in claim 1 , further comprising:

receiving, at the device, feedback regarding one or more path switchovers; and

using, by the device, the received feedback to determine the predicted effect of directing the traffic flow to the second path.

8. An apparatus, comprising:

one or more network interfaces to communicate with a network;

a processor coupled to the network interfaces and configured to execute one or more processes; and

a memory configured to store a process executable by the processor, the process when executed operable to:

receive an alert regarding a performance change of a first path in the network;

receive a switchover policy for a particular type of traffic in the network;

in response to receiving the alert, determine a predicted effect of directing a traffic flow of the particular type of traffic from the first path in the network to a second path in the network, the predicted effect corresponding to one or more predicted performances differences between the first path and the second path;

determine whether the predicted effect of directing the traffic flow to the second path would violate the switchover policy;

cause the traffic flow to be routed via the second path in the network, based on a determination that the predicted effect of directing the traffic flow to the second path would not violate the switchover policy for the particular type of traffic; and

prevent switchover of the traffic flow to the second path in the network, based on a determination that the predicted effect of directing the traffic flow to the second path would violate the switchover policy for the particular type of traffic.

9. The apparatus as in claim 8 , wherein the predicted effect comprises a predicted amount of packet reordering that would occur were the traffic flow directed to the second path, and wherein the switchover policy indicates a threshold amount of acceptable packet reordering for the particular type of traffic.

10. The apparatus as in claim 8 , wherein the predicted effect comprises a predicted decrease in bandwidth for the traffic flow were the traffic flow routed via the second path, and wherein the switchover policy indicates a threshold amount of acceptable decrease in bandwidth for the particular type of traffic.

11. The apparatus as in claim 8 , wherein the predicted effect comprises a predicted duration for the traffic flow, and wherein the switchover policy indicates an acceptable flow duration in order for the traffic flow to be directed to the second path.

12. The apparatus as in claim 8 , wherein the process when executed is further operable to:

generate a machine learning model that models the particular type of traffic; and

use the machine learning model to determine the predicted effect of directing the traffic flow to the second path.

13. The apparatus as in claim 8 , wherein the process when executed is further operable to:

determine an estimated performance of the second path; and

use the estimated performance of the second path to determine the predicted effect of directing the traffic flow to the second path.

14. The apparatus as in claim 8 , wherein the process when executed is further operable to:

receive feedback regarding one or more path switchovers; and

use the received feedback to determine the predicted effect of directing the traffic flow to the second path.

15. A tangible, non-transitory, computer-readable media having software encoded thereon, the software when executed by a processor of a device operable to:

receive an alert regarding a performance change of a first path in the network;

receive a switchover policy for a particular type of traffic in the network;

in response to receiving the alert, determine a predicted effect of directing a traffic flow of the particular type of traffic from the first path in the network to a second path in the network, the predicted effect corresponding to one or more predicted performance differences between the first path and the second path;

determine whether the predicted effect of directing the traffic flow to the second path would violate the switchover policy;

cause the traffic flow to be routed via the second path in the network, based on a determination that the predicted effect of directing the traffic flow to the second path would not violate the switchover policy for the particular type of traffic; and

prevent switchover of the traffic flow to the second path in the network, based on a determination that the predicted effect of directing the traffic to the second path would violate the switchover policy for the particular type of traffic.

16. The computer-readable media as in claim 15 , wherein the predicted effect comprises a predicted amount of packet reordering that would occur were the traffic flow directing to the second path, and wherein the switchover policy indicates a threshold amount of acceptable packet reordering for the particular type of traffic.

17. The computer-readable media as in claim 15 , wherein the predicted effect comprises a predicted decrease in bandwidth for the traffic flow were the traffic flow routed via the second path, and wherein the switchover policy indicates a threshold amount of acceptable decrease in bandwidth for the particular type of traffic.

18. The computer-readable media as in claim 15 , wherein the predicted effect comprises a predicted duration for the traffic flow, and wherein the switchover policy indicates an acceptable flow duration in order for the traffic flow to be directed to the second path.

19. The computer-readable media as in claim 15 , wherein the software when executed is further operable to:

generate a machine learning model that models the particular type of traffic; and

use the machine learning model to determine the predicted effect of directing the traffic flow to the second path.

20. The computer-readable media as in claim 15 , wherein the software when executed is further operable to:

determine an estimated performance of the second path; and

use the estimated performance of the second path to determine the predicted effect of directing the traffic flow to the second path.

21. The computer-readable media as in claim 15 , wherein the software when executed is further operable to:

receive feedback regarding one or more path switchovers; and

use the received feedback to determine the predicted effect of directing the traffic flow to the second path.