Computing motif signatures for QoE disruptions
In one embodiment, a device obtains a set of probing motifs. Each probing motif groups similar patterns of path probing results for one or more path metrics in a network. The device generates signatures for the set of probing motifs. Each signature relates a probing strategy with a measure of performance of a classifier to detect that motif were path probing to be conducted in accordance with that probing strategy. The device selects, based on the signatures for the set of probing motifs, a particular probing strategy to use in the network. The device causes one or more probing agents in the network to conduct path probing in accordance with the particular probing strategy.
1 . A method comprising:
obtaining, by a device, a set of probing motifs, wherein each probing motif groups similar patterns of path probing results for one or more path metrics in a network;
generating, by the device, signatures for the set of probing motifs by training and evaluating, for each probing motif and each candidate probing frequency, a classifier to discriminate first samples of the probing motif from second samples of non-impacting motifs, wherein each signature of the signatures relates a given probing strategy with a measure of performance of the classifier to detect each probing motif included in the set of probing motifs if path probing were to be conducted in accordance with the given probing strategy;
selecting, by the device and based on the signatures for the set of probing motifs, a particular probing strategy to use in the network; and
causing, by the device, one or more probing agents in the network to conduct path probing in accordance with the particular probing strategy.
2 . The method as in claim 1 , wherein the one or more path metrics comprise at least one of: delay, loss, or jitter.
3 . The method as in claim 1 , wherein the measure of performance comprises a recall or precision metric.
4 . The method as in claim 1 , wherein the particular probing strategy comprises sending path probes at multiple probing frequencies.
5 . The method as in claim 1 , wherein the device selects the particular probing strategy based on a defined policy that specifies a threshold for the measure of performance and a threshold probing frequency.
6 . The method as in claim 5 , further comprising:
receiving, at the device, the defined policy from a user interface.
7 . The method as in claim 1 , wherein each of the set of probing motifs is associated with a corresponding impact to a quality of experience metric for a particular online application.
8 . The method as in claim 7 , further comprising:
detecting, by the device, degradation to the quality of experience metric for the particular online application by using the classifier to detect a set of probing motifs from probing results captured by the one or more probing agents.
9 . The method as in claim 1 , wherein causing the one or more probing agents in the network to conduct path probing in accordance with the particular probing strategy comprises:
sending a finite state machine that encodes the particular probing strategy to the one or more probing agents.
10 . The method as in claim 1 , wherein the one or more probing agents are executed by one or more edge routers in the network.
11 . 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:
obtain a set of probing motifs, wherein each probing motif groups similar patterns of path probing results for one or more path metrics in a network;
generate signatures for the set of probing motifs by training and evaluating, for each probing motif and each candidate probing frequency, a classifier to discriminate first samples of the probing motif from second samples of non-impacting motifs, wherein each signature of the signatures relates a probing strategy with a measure of performance of the classifier to detect a given probing motif included in the set of probing motifs if path probing were to be conducted in accordance with the probing strategy;
select, based on the signatures for the set of probing motifs, a particular probing strategy to use in the network; and
cause one or more probing agents in the network to conduct path probing in accordance with the particular probing strategy.
12 . The apparatus as in claim 11 , wherein the one or more path metrics comprise at least one of: delay, loss, or jitter.
13 . The apparatus as in claim 11 , wherein the measure of performance comprises a recall or precision metric.
14 . The apparatus as in claim 11 , wherein the particular probing strategy comprises sending path probes at multiple probing frequencies.
15 . The apparatus as in claim 11 , wherein the apparatus selects the particular probing strategy based on a defined policy that specifies a threshold for the measure of performance and a threshold probing frequency.
16 . The apparatus as in claim 15 , wherein the process when executed is further configured to:
receive the defined policy from a user interface.
17 . The apparatus as in claim 11 , wherein each of the set of probing motifs is associated with a corresponding impact to a quality of experience metric for a particular online application.
18 . The apparatus as in claim 17 , wherein the process when executed is further configured to:
detect degradation to the quality of experience metric for the particular online application by using the classifier to detect a set of probing motifs from probing results captured by the one or more probing agents.
19 . The apparatus as in claim 11 , wherein the apparatus causes the one or more probing agents in the network to conduct path probing in accordance with the particular probing strategy by:
sending a finite state machine that encodes the particular probing strategy to the one or more probing agents.
20 . A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising:
obtaining, by the device, a set of probing motifs, wherein each probing motif groups similar patterns of path probing results for one or more path metrics in a network;
generating, by the device, signatures for the set of probing motifs by training and evaluating, for each probing motif and each candidate probing frequency, a classifier to discriminate first samples of the probing motif from second samples of non-impacting motifs, wherein each signature of the signatures relates a given probing strategy with a measure of performance of the classifier to detect each probing motif included in the set of probing motifs if path probing were to be conducted in accordance with the given probing strategy;
selecting, by the device and based on the signatures for the set of probing motifs, a particular probing strategy to use in the network; and
causing, by the device, one or more probing agents in the network to conduct path probing in accordance with the particular probing strategy.