Catch-up TV forecasting: enabling next-generation over-the-top multimedia TV services

Due to recent developments in Over-The-Top (OTT) technologies, Pay-TV operators have begun a migration process of managed IP Television (IPTV) services to more appealing OTT approaches. In these scenarios, being able to predict when and what resources will be necessary at any given point is crucial to a high-quality, efficient, and cost-effective operation, especially when dealing with the dynamic and resource-intensive requirements of IPTV multimedia services. To evaluate the advantages of demand forecasting for efficient Catch-up TV delivery on OTT scenarios, this research work explores several classes of machine learning models regarding their accuracy, computational requirement trade-offs, and deployability. The training process relies on a dataset comprised of Catch-up TV usage logs acquired from an IPTV operator’s live production service containing over 1 million subscribers. A predictive and dynamic resource provisioning approach is proposed and evaluated in terms of bandwidth and storage savings. Results demonstrate that forecasting Catch-up TV demand is practical, suitable for integration in OTT solutions, and useful in improving efficiency, with benefits to operators and consumers. Significant savings in bandwidth and storage are shown to be achievable, enabling green and cost-effective resource usage.

[1]  Jean-Samuel Beuscart,et al.  Audience dynamics of online catch up TV , 2012, WWW.

[2]  Pedro Larrañaga,et al.  A review of feature selection techniques in bioinformatics , 2007, Bioinform..

[3]  Wolfgang Kellerer,et al.  Software Defined Optical Access Networks (SDOANs): A Comprehensive Survey , 2015, ArXiv.

[4]  Randy Sharpe,et al.  Forecasting of access network bandwidth demands for aggregated subscribers using Monte Carlo methods , 2015, IEEE Communications Magazine.

[5]  George Mastorakis,et al.  Efficient entertainment services provision over a novel network architecture , 2016, IEEE Wireless Communications.

[6]  Fotios Petropoulos,et al.  forecast: Forecasting functions for time series and linear models , 2018 .

[7]  Hassan Haghighi,et al.  An adaptive RL based approach for dynamic resource provisioning in Cloud virtualized data centers , 2015, Computing.

[8]  Jon Crowcroft,et al.  Understanding and decreasing the network footprint of catch-up tv , 2013, WWW.

[9]  Rüdiger Wirth,et al.  CRISP-DM: Towards a Standard Process Model for Data Mining , 2000 .

[10]  Richard A. Johnson,et al.  A new family of power transformations to improve normality or symmetry , 2000 .

[11]  Carlos Becker Westphall,et al.  Cloud resource management: A survey on forecasting and profiling models , 2015, J. Netw. Comput. Appl..

[12]  Filip De Turck,et al.  Towards a predictive cache replacement strategy for multimedia content , 2013, J. Netw. Comput. Appl..

[13]  Max Kuhn,et al.  Building Predictive Models in R Using the caret Package , 2008 .

[14]  Wolfgang Kellerer,et al.  Software Defined Optical Networks (SDONs): A Comprehensive Survey , 2015, IEEE Communications Surveys & Tutorials.

[15]  Susana Sargento,et al.  Catch-up TV analytics: statistical characterization and consumption patterns identification on a production service , 2016, Multimedia Systems.

[16]  Rich Caruana,et al.  Ensemble selection from libraries of models , 2004, ICML.

[17]  Henning Schulzrinne,et al.  Intelligent content delivery over wireless via SDN , 2015, 2015 IEEE Wireless Communications and Networking Conference (WCNC).

[18]  Rajiv Ranjan,et al.  Cloud Resource Orchestration Programming: Overview, Issues, and Directions , 2015, IEEE Internet Computing.

[19]  Chris Tofallis,et al.  A better measure of relative prediction accuracy for model selection and model estimation , 2014, J. Oper. Res. Soc..

[20]  S. Karsoliya,et al.  Approximating Number of Hidden layer neurons in Multiple Hidden Layer BPNN Architecture , 2012 .

[21]  Susana Sargento,et al.  Over-The-Top Catch-up TV content-aware caching , 2016, 2016 IEEE Symposium on Computers and Communication (ISCC).

[22]  Rob J. Hyndman,et al.  Another Look at Forecast Accuracy Metrics for Intermittent Demand , 2006 .

[23]  Jeffrey O. Kephart,et al.  The Vision of Autonomic Computing , 2003, Computer.

[24]  Rajkumar Buyya,et al.  A Taxonomy of CDNs , 2008 .

[25]  Annette M. Molinaro,et al.  Prediction error estimation: a comparison of resampling methods , 2005, Bioinform..

[26]  David Geerts,et al.  Broadcast, Video-on-Demand, and Other Ways to Watch Television Content: A Household Perspective , 2015, TVX.

[27]  Ron Kohavi,et al.  Irrelevant Features and the Subset Selection Problem , 1994, ICML.

[28]  Max Kuhn,et al.  Applied Predictive Modeling , 2013 .

[29]  Dave Winkler,et al.  Bayesian Regularization of Neural Networks , 2009, Artificial Neural Networks.

[30]  LarrañagaPedro,et al.  A review of feature selection techniques in bioinformatics , 2007 .