On distribution of user movie watching time in a large-scale video streaming system

Yishuai Chen, Yong Liu, Baoxian Zhang, Wei Zhu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Video watching time is a crucial measure for studying user watching behavior in online Internet video-on-demand (VoD) systems. It is important for system planning, user engagement study, and service quality evaluation. However, due to limited access to large-scale VoD systems, there is still a lack of accurate model for characterizing the distribution of user watching time on a per video basis. In this paper, we measure PPLive, one of the most popular commercial Internet VoD systems in China, over a three week period, and characterize user watching time distributions of 1,000 most popular movies. We find that a video's watching time can be modeled by a concatenation of exponential distribution (in the first several minutes of the video) and truncated power law distribution (in the remaining time of the video), when users watch the video without interruptions. For comparison, user watching time with user interactions such as seeking and/or pause operations does not follow such a distribution. We further reveal interesting characteristics regarding the relation between video's watching time distribution and various watching/video-related features (including time-of-day, user ratings, and movie genres). Our measurement and modeling results bring forth important insights for design, deployment, and evaluation of Internet VoD systems.

Original languageEnglish (US)
Title of host publication2014 IEEE International Conference on Communications, ICC 2014
PublisherIEEE Computer Society
Pages1825-1830
Number of pages6
ISBN (Print)9781479920037
DOIs
StatePublished - 2014
Event2014 1st IEEE International Conference on Communications, ICC 2014 - Sydney, NSW, Australia
Duration: Jun 10 2014Jun 14 2014

Other

Other2014 1st IEEE International Conference on Communications, ICC 2014
CountryAustralia
CitySydney, NSW
Period6/10/146/14/14

Fingerprint

Video on demand
Video streaming
Internet
Planning

Keywords

  • online video
  • user watching time
  • video-on-demand

ASJC Scopus subject areas

  • Computer Networks and Communications

Cite this

Chen, Y., Liu, Y., Zhang, B., & Zhu, W. (2014). On distribution of user movie watching time in a large-scale video streaming system. In 2014 IEEE International Conference on Communications, ICC 2014 (pp. 1825-1830). [6883588] IEEE Computer Society. https://doi.org/10.1109/ICC.2014.6883588

On distribution of user movie watching time in a large-scale video streaming system. / Chen, Yishuai; Liu, Yong; Zhang, Baoxian; Zhu, Wei.

2014 IEEE International Conference on Communications, ICC 2014. IEEE Computer Society, 2014. p. 1825-1830 6883588.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Chen, Y, Liu, Y, Zhang, B & Zhu, W 2014, On distribution of user movie watching time in a large-scale video streaming system. in 2014 IEEE International Conference on Communications, ICC 2014., 6883588, IEEE Computer Society, pp. 1825-1830, 2014 1st IEEE International Conference on Communications, ICC 2014, Sydney, NSW, Australia, 6/10/14. https://doi.org/10.1109/ICC.2014.6883588
Chen Y, Liu Y, Zhang B, Zhu W. On distribution of user movie watching time in a large-scale video streaming system. In 2014 IEEE International Conference on Communications, ICC 2014. IEEE Computer Society. 2014. p. 1825-1830. 6883588 https://doi.org/10.1109/ICC.2014.6883588
Chen, Yishuai ; Liu, Yong ; Zhang, Baoxian ; Zhu, Wei. / On distribution of user movie watching time in a large-scale video streaming system. 2014 IEEE International Conference on Communications, ICC 2014. IEEE Computer Society, 2014. pp. 1825-1830
@inproceedings{515b047458944e559ff3b22c1f5ee244,
title = "On distribution of user movie watching time in a large-scale video streaming system",
abstract = "Video watching time is a crucial measure for studying user watching behavior in online Internet video-on-demand (VoD) systems. It is important for system planning, user engagement study, and service quality evaluation. However, due to limited access to large-scale VoD systems, there is still a lack of accurate model for characterizing the distribution of user watching time on a per video basis. In this paper, we measure PPLive, one of the most popular commercial Internet VoD systems in China, over a three week period, and characterize user watching time distributions of 1,000 most popular movies. We find that a video's watching time can be modeled by a concatenation of exponential distribution (in the first several minutes of the video) and truncated power law distribution (in the remaining time of the video), when users watch the video without interruptions. For comparison, user watching time with user interactions such as seeking and/or pause operations does not follow such a distribution. We further reveal interesting characteristics regarding the relation between video's watching time distribution and various watching/video-related features (including time-of-day, user ratings, and movie genres). Our measurement and modeling results bring forth important insights for design, deployment, and evaluation of Internet VoD systems.",
keywords = "online video, user watching time, video-on-demand",
author = "Yishuai Chen and Yong Liu and Baoxian Zhang and Wei Zhu",
year = "2014",
doi = "10.1109/ICC.2014.6883588",
language = "English (US)",
isbn = "9781479920037",
pages = "1825--1830",
booktitle = "2014 IEEE International Conference on Communications, ICC 2014",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - On distribution of user movie watching time in a large-scale video streaming system

AU - Chen, Yishuai

AU - Liu, Yong

AU - Zhang, Baoxian

AU - Zhu, Wei

PY - 2014

Y1 - 2014

N2 - Video watching time is a crucial measure for studying user watching behavior in online Internet video-on-demand (VoD) systems. It is important for system planning, user engagement study, and service quality evaluation. However, due to limited access to large-scale VoD systems, there is still a lack of accurate model for characterizing the distribution of user watching time on a per video basis. In this paper, we measure PPLive, one of the most popular commercial Internet VoD systems in China, over a three week period, and characterize user watching time distributions of 1,000 most popular movies. We find that a video's watching time can be modeled by a concatenation of exponential distribution (in the first several minutes of the video) and truncated power law distribution (in the remaining time of the video), when users watch the video without interruptions. For comparison, user watching time with user interactions such as seeking and/or pause operations does not follow such a distribution. We further reveal interesting characteristics regarding the relation between video's watching time distribution and various watching/video-related features (including time-of-day, user ratings, and movie genres). Our measurement and modeling results bring forth important insights for design, deployment, and evaluation of Internet VoD systems.

AB - Video watching time is a crucial measure for studying user watching behavior in online Internet video-on-demand (VoD) systems. It is important for system planning, user engagement study, and service quality evaluation. However, due to limited access to large-scale VoD systems, there is still a lack of accurate model for characterizing the distribution of user watching time on a per video basis. In this paper, we measure PPLive, one of the most popular commercial Internet VoD systems in China, over a three week period, and characterize user watching time distributions of 1,000 most popular movies. We find that a video's watching time can be modeled by a concatenation of exponential distribution (in the first several minutes of the video) and truncated power law distribution (in the remaining time of the video), when users watch the video without interruptions. For comparison, user watching time with user interactions such as seeking and/or pause operations does not follow such a distribution. We further reveal interesting characteristics regarding the relation between video's watching time distribution and various watching/video-related features (including time-of-day, user ratings, and movie genres). Our measurement and modeling results bring forth important insights for design, deployment, and evaluation of Internet VoD systems.

KW - online video

KW - user watching time

KW - video-on-demand

UR - http://www.scopus.com/inward/record.url?scp=84906995491&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906995491&partnerID=8YFLogxK

U2 - 10.1109/ICC.2014.6883588

DO - 10.1109/ICC.2014.6883588

M3 - Conference contribution

SN - 9781479920037

SP - 1825

EP - 1830

BT - 2014 IEEE International Conference on Communications, ICC 2014

PB - IEEE Computer Society

ER -