Modeling channel popularity dynamics in a large IPTV system

  • Authors:

  • Tongqing Qiu;Zihui Ge;Seungjoon Lee;Jia Wang;Qi Zhao;Jun Xu

  • Affiliations:

  • Georgia Tech, Atlanta, GA, USA;AT&T Labs -- Research, Florham Park, NJ, USA;AT&T Labs - Research, Florham Park, NJ, USA;AT&T Labs - Research, Florham Park, NJ, USA;AT&T Labs - Research, Florham Park, NJ, USA;Georgia Tech, Atlanta, GA, USA

  • Venue:
  • Proceedings of the eleventh international joint conference on Measurement and modeling of computer systems
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Understanding the channel popularity or content popularity is an important step in the workload characterization for modern information distribution systems (e.g., World Wide Web, peer-to-peer file-sharing systems, video-on-demand systems). In this paper, we focus on analyzing the channel popularity in the context of Internet Protocol Television (IPTV). In particular, we aim at capturing two important aspects of channel popularity - the distribution and temporal dynamics of the channel popularity. We conduct in-depth analysis on channel popularity on a large collection of user channel access data from a nation-wide commercial IPTV network. Based on the findings in our analysis, we choose a stochastic model that finds good matches in all attributes of interest with respect to the channel popularity. Furthermore, we propose a method to identify subsets of user population with inherently different channel interest. By tracking the change of population mixtures among different user classes, we extend our model to a multi-class population model, which enables us to capture the moderate diurnal popularity patterns exhibited in some channels. We also validate our channel popularity model using real user channel access data from commercial IPTV network.