Process and apparatus for estimating real-time quality of experience
Disclosed are a process and apparatus for classifying video streams of an online streaming media service in real-time. The process includes processing data packets representing one or more video streams between a service provider and a user access network, generating flow activity data from the packets representing quantitative metrics of network transport activity, and applying a trained classifier to the flow activity data to classify each of the video streams as either a live video stream or a video-on-demand (VoD) stream.
1. A computer-implemented process for classifying video streams of an online streaming media service in real-time, the process being for use by a network operator, and including:
processing packets of one or more network flows representing one or more video streams of the online service at a network location between a provider of the service and a user access network to generate flow activity data representing quantitative metrics of real-time network transport activity of each of the one or more network flows of the online service, the quantitative metrics including, for each said video stream, a corresponding time series of request packet counter values;
applying a trained classifier to each said time series of request packet counter values to determine whether the request packet counter values for each said video stream are indicative of live video streaming; and
in dependence upon the determination, to classify each of the one or more video streams as either a live video stream or as a video-on-demand stream.
2. The process of claim 1 , including applying one or more further trained classifiers to the flow activity data to generate, for each video stream, corresponding user experience data representing real-time quality of experience of the video stream.
3. The process of claim 2 , wherein, responsive to determining that the request packet counter values are indicative of a live video stream, the step of applying one or more further trained classifiers includes applying further classifiers to chunk features of the live video stream to generate corresponding user experience data representing real-time quality of experience of the live video stream.
4. The process of claim 2 , wherein the user experience data represents a corresponding quality of experience state selected from a plurality of quality of experience states.
5. The process of claim 4 , wherein the plurality of experience states include a maximum bitrate playback state, a varying bitrate playback state, a depleting buffer state, and a playback stall state.
6. The process of claim 1 , wherein the user experience data represents one or more quantitative metrics of quality of experience.
7. The process of claim 6 , wherein the online service is a streaming media service, and the one or more quantitative metrics of quality of experience include quantitative metrics of buffer fill time, bitrate and throughput.
8. The process of claim 6 , wherein the one or more quantitative metrics of quality of experience include quantitative metrics of resolution and buffer depletion for live video streaming.
9. The process of claim 1 , wherein the online service is a Twitch™, Facebook™ Live, or YouTube™ Live, live streaming service.
10. The process of claim 1 , including, in dependence on the user experience data, automatically reconfiguring a networking component to improve quality of experience of the online service by prioritising one or more network flows of the online service over other network flows.
11. The process of claim 1 , including training the classifier by processing packets of one or more training network flows of the online service to generate training flow activity data and chunk metadata (for videos) representing quantitative metrics of network transport activity of each of the one or more training network flows of the online service;
generating corresponding training user experience data representing corresponding temporal quality of user experience of the online service; and applying machine learning to the generated training flow activity data and the generated training user experience data to generate a corresponding model for the classifier based on correlations between the quantitative metrics of network transport activity and the temporal quality of user experience of the online service.
12. Apparatus for classifying, in real-time, video streams of an online streaming media service, the apparatus being for use by a network operator, and including:
a flow quantifier configured to process packets of one or more network flows representing one or more video streams of the online service at a network location between a provider of the service and a user access network to generate flow activity data representing quantitative metrics of real-time network transport activity of each of the one or more network flows of the online service, the quantitative metrics including, for each said video stream, a corresponding time series of request packet counter values for the online service; and
a trained classifier configured to process each time series of request packet counter values to determine whether the request packet counter values are indicative of live video streaming, and, in dependence upon the determination, to classify each of the one or more video streams as either a live video stream or as a video-on-demand stream.
13. The apparatus of claim 12 , including one or more further trained classifiers configured to process the flow activity data to generate, for each video stream, corresponding user experience data representing real-time quality of experience (QoE) of the video stream.
14. The apparatus of claim 13 , wherein the one or more further trained classifiers are configured to process, in response to determining that the request packet counter values are indicative of a live video stream, chunk features of the live video stream to generate corresponding user experience data representing real-time quality of experience of the live video stream.
15. The apparatus of claim 13 , wherein the user experience data represents a corresponding quality of experience state selected from a plurality of quality of experience states.
16. The apparatus of claim 15 , wherein the plurality of experience states include a maximum bitrate playback state, a varying bitrate playback state, a depleting buffer state, and a playback stall state.
17. The apparatus of claim 13 , wherein the user experience data represents one or more quantitative metrics of quality of experience.
18. The apparatus of claim 17 , wherein the online service is a streaming media service, and the one or more quantitative metrics of quality of experience include quantitative metrics of buffer fill time, bitrate and throughput.
19. The apparatus of claim 17 , wherein the online service provides live video streaming, and the one or more quantitative metrics of quality of experience include quantitative metrics of resolution and buffer depletion for live video streaming.
20. The apparatus of claim 12 , wherein the online service is a Twitch™, Facebook™ Live, or YouTube™ Live, live streaming service.
21. The apparatus of claim 12 , including a user experience controller configured to, in dependence on the user experience data, automatically reconfigure a networking component to improve quality of experience of the online service by prioritising one or more network flows of the online service over other network flows.
22. At least one computer-readable storage medium having stored thereon processor-executable instructions that, when executed by at least one processor, cause the at least one processor to execute the process of claim 1 .
23. Apparatus for classifying, in real-time, video streams of an online streaming media service, the apparatus being for use by a network operator, and including a memory and at least one processor configured to execute the process of claim 1 .