1 / 10

Developing a Predictive Model of Quality of Experience for Internet Video

Developing a Predictive Model of Quality of Experience for Internet Video. Athula Balachandran , Vyas Sekarz , Aditya Akellay , Srinivasan Seshan , Ion Stoica , and Hui Zhang SIGCOMM 2013. Goal and Challenges. To develop a predictive model of user QoE in viewing Internet video

bandele
Télécharger la présentation

Developing a Predictive Model of Quality of Experience for Internet Video

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Developing a Predictive Model ofQuality of Experience for Internet Video AthulaBalachandran, VyasSekarz, AdityaAkellay, SrinivasanSeshan, Ion Stoica, and Hui Zhang SIGCOMM 2013

  2. Goal and Challenges • To develop a predictive model of user QoE in viewing Internet video • Challenges: • Relationship between quality and engagement • Dependencies between quality metrics • Confounding factors

  3. Dataset • It was collected by conviva.com in real time using a client-side instrumentation library • 40 million video sessions collected over 3 months under two popular video content providers • One provider serves mostly VOD content, and another provider serves live broadcast for sports events.

  4. Relationship between quality and engagement • Engagement linearly decreases with increasing rate of buffering up to 0.3 buff events/min 0.3

  5. Confounding factors v.s. Engagement • Some factors may affect user viewing behavior and result in different observed engagements

  6. Confounding factors • Figure out effective cofounding factors by relative information gain

  7. Confounding factors-Device • For a VOD subset dataset, increased bitrate led to lower engagement in the case of TV

  8. Confounding factors-Device • It shows that mobile users are more tolerant toward low quality

  9. Quality model • It computes the mean performance(buffering ratio, rate of buffering and join time) for each combination of attributes (e.g., type of video, ISP, region, device) and control parameters (e.g., bitrate and CDN) using empirical estimation

  10. Predicted Average Engagement

More Related