Dynamic input granularity estimation for network path forecasting using timeseries features
In one embodiment, a device identifies peaks of a timeseries of a path metric used to predict performance of a path in a network. The device determines one or more characteristics of the peaks of the timeseries. The device computes, based on the one or more characteristics of the peaks, a measurement frequency for the path metric. The device causes the path metric to be measured in the network according to the measurement frequency.
1. A method comprising:
identifying, by a device, peaks of timeseries information of a path metric, wherein the path metric is used to predict performance of a path in a network and is indicative of at least one of: delay, loss, jitter, or throughput of the path;
determining, by the device, one or more characteristics of the peaks of the timeseries information;
computing, by the device and based on the one or more characteristics of the peaks, a measurement frequency for the path metric; and
causing, by the device, the path metric to be measured in the network according to the measurement frequency.
2. The method as in claim 1 , wherein the peaks of the timeseries information identified by the device satisfy an imposed minimum peak height or a maximum peak width.
3. The method as in claim 1 , wherein the one or more characteristics of the peaks of the timeseries information indicate whether the peaks of the timeseries information are periodic or aperiodic.
4. The method as in claim 1 , wherein identifying the peaks of the timeseries information comprises:
excluding a fluctuation in the timeseries information as a peak based on a required minimum amount of time between peaks.
5. The method as in claim 1 , wherein the one or more characteristics of the peaks of the timeseries information indicate whether the peaks of the timeseries information are preceded by patterns that signal that a peak is imminent.
6. The method as in claim 1 , wherein the path metric is used to predict performance of the path by a prediction model of a routing engine that reroutes traffic conveyed via the path onto another path in the network in advance of a predicted degradation of the path metric.
7. The method as in claim 6 , wherein the device computes the measurement frequency based further in part on an accuracy measurement for the prediction model.
8. The method as in claim 1 , further comprising:
computing, by the device and based on the one or more characteristics of the peaks, a length of history of the path metric to be retained.
9. The method as in claim 1 , further comprising:
identifying, by the device, a second path in the network as being similar to that of the path based on a comparison of timeseries information of the second path to the timeseries information relating to the path metric; and
computing, by the device and based on the measurement frequency, a second measurement frequency for the second path.
10. An apparatus, comprising:
one or more network interfaces;
a processor coupled to the one or more network interfaces and configured to execute one or more processes; and
a memory configured to store a process that is executable by the processor, the process when executed configured to:
identify peaks of timeseries information of a path metric, wherein the path metric is used to predict performance of a path in a network and is indicative of at least one of: delay, loss, jitter, or throughput of the path;
determine one or more characteristics of the peaks of the timeseries information;
compute, based on the one or more characteristics of the peaks, a measurement frequency for the path metric; and
cause the path metric to be measured in the network according to the measurement frequency.
11. The apparatus as in claim 10 , wherein the peaks of the timeseries information identified by the apparatus satisfy an imposed minimum peak height or a maximum peak width.
12. The apparatus as in claim 10 , wherein the one or more characteristics of the peaks of the timeseries information indicate whether the peaks of the timeseries information are periodic or aperiodic.
13. The apparatus as in claim 10 , wherein the apparatus identifies the peaks of the timeseries information by:
excluding a fluctuation in the timeseries information as a peak based on a required minimum amount of time between peaks.
14. The apparatus as in claim 10 , wherein the one or more characteristics of the peaks of the timeseries information indicate whether the peaks of the timeseries information are preceded by patterns that signal that a peak is imminent.
15. The apparatus as in claim 10 , wherein the path metric is used to predict performance of the path by a prediction model of a routing engine that reroutes traffic conveyed via the path onto another path in the network in advance of a predicted degradation of the path metric.
16. The apparatus as in claim 15 , wherein the apparatus computes the measurement frequency based further in part on an accuracy measurement for the prediction model.
17. The apparatus as in claim 10 , wherein the process when executed is further configured to:
compute, based on the one or more characteristics of the peaks, a length of history of the path metric to be retained.
18. The apparatus as in claim 10 , wherein the process when executed is further configured to:
identify a second path in the network as being similar to that of the path based on a comparison of timeseries information of the second path to the timeseries information relating to the path metric; and
compute, based on the measurement frequency, a second measurement frequency for the second path.
19. A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising:
identifying, by the device, peaks of timeseries information of a path metric, wherein the path metric is used to predict performance of a path in a network and is indicative of at least one of: delay, loss, jitter, or throughput of the path;
determining, by the device, one or more characteristics of the peaks of the timeseries information;
computing, by the device and based on the one or more characteristics of the peaks, a measurement frequency for the path metric; and
causing, by the device, the path metric to be measured in the network according to the measurement frequency.