Reporting path measurements for application quality of experience prediction using an interest metric
In one embodiment, a device determines a first difference between current path measurements and prior path measurements. The device determines a second difference between current predictions and prior predictions made by a prediction model based on path measurements. The device computes, based on the first difference and the second difference, an interest metric for the current path measurements. The device sends at least a portion of the current path measurements for input to the prediction model, when the interest metric exceeds a predefined threshold.
1. A method, comprising:
determining, at a device, a first difference between current path measurements and prior path measurements;
determining, by the device, a second difference between current predictions and prior predictions made by a prediction model based on path measurements;
computing, by the device and based on the first difference and the second difference, an interest metric for the current path measurements; and
sending, by the device, at least a portion of the current path measurements for input to the prediction model, when the interest metric exceeds a predefined threshold.
2. The method as in claim 1 , wherein the prediction model is hosted in a cloud computing environment.
3. The method as in claim 1 , wherein the prediction model is updated based on the portion of the current path measurements.
4. The method as in claim 1 , wherein predictions made by the prediction model based on the path measurements comprise at least one of resolution, concealment, or smoothness.
5. The method as in claim 1 , wherein predictions made by the prediction model based on the path measurements comprise Quality of Experience predictions.
6. The method as in claim 1 , wherein computing, based on the first difference and the second difference, the interest metric for the current path measurements comprises:
applying, by the device, weightings to the first difference and the second difference.
7. The method as in claim 1 , wherein sending, by the device, at least the portion of the current path measurements for input to the prediction model, when the interest metric exceeds the predefined threshold comprises:
selecting, by the device, the current path measurements based on whether the current path measurements are of a type of metric.
8. The method as in claim 1 , wherein the current path measurements and the prior path measurements each comprise at least one of bit rate, forward error correction rate, packet loss, throughput, transmission delay, availability, or jitter.
9. The method as in claim 8 , wherein the current path measurements and the prior path measurements comprise statistical measurements.
10. The method as in claim 1 , wherein the current path measurements and the prior path measurements are made for one or more paths over which traffic for a real-time communication application executed by the device is sent.
11. An apparatus, comprising:
one or more network interfaces to communicate with one or more networks;
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:
determine a first difference between current path measurements and prior path measurements;
determine a second difference between current predictions and prior predictions made by a prediction model based on path measurements;
compute, based on the first difference and the second difference, an interest metric for the current path measurements; and
send at least a portion of the current path measurements for input to the prediction model, when the interest metric exceeds a predefined threshold.
12. The apparatus as in claim 11 , wherein the prediction model is hosted in a cloud computing environment.
13. The apparatus as in claim 11 , wherein the prediction model is updated based on the portion of the current path measurements.
14. The apparatus as in claim 11 , wherein predictions made by the prediction model based on the path measurements comprise at least one of resolution, concealment, or smoothness.
15. The apparatus as in claim 11 , wherein predictions made by the prediction model based on the path measurements comprise Quality of Experience predictions.
16. The apparatus as in claim 11 , wherein to compute, based on the first difference and the second difference, the interest metric for the current path measurements comprises:
applying weightings to the first difference and the second difference.
17. The apparatus as in claim 11 , wherein to send at least the portion of the current path measurements for input to the prediction model, when the interest metric exceeds the predefined threshold comprises:
selecting the current path measurements based on whether the current path measurements are of a type of metric.
18. The apparatus as in claim 11 , wherein the current path measurements and the prior path measurements each comprise at least one of bit rate, forward error correction rate, packet loss, throughput, transmission delay, availability, or jitter.
19. The apparatus as in claim 18 , wherein the current path measurements and the prior path measurements comprise statistical measurements.
20. A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising:
determining, at the device, a first difference between current path measurements and prior path measurements;
determining, by the device, a second difference between current predictions and prior predictions made by a prediction model based on path measurements;
computing, by the device and based on the first difference and the second difference, an interest metric for the current path measurements; and
sending, by the device, at least a portion of the current path measurements for input to the prediction model, when the interest metric exceeds a predefined threshold.