Catch-up TV Big Data in MEO

Context

MEO’s Catch-up TV service (Gravações Automáticas) is a highly popular on-demand interactive TV service with over 22 million individual video streaming sessions each month, which imposes a severe strain on the content delivery network, particularly at prime time.
This project’s challenge is to devise a prediction model/system able to decide which programs should be prefetched (loaded in advance) to users’ STBs during the off-peak hours (late night) in order to reduce peak network bandwidth consumption.
The potential gains of such a system have already been identified in the literature, as illustrated by (http://www.inf.kcl.ac.uk/staff/nrs/pubs/www2013.pdf), which concluded that “even with a modest storage of 32GB, an oracle with complete knowledge of user consumption can save up to 74% of the energy, and 97% of the peak bandwidth compared to the current IP streaming-based architecture”.

Project objectives

Research and develop a forecasting model able to decide which programs to load in advance (prefetch) to MEO STB’s, using R tools and libraries as well as Catch-up TV consumption data to create and validate the predictive system.

Key skills required

Altice Labs product lines

TV, Content and Internet Platforms and Applications

Want to participate in this project?

If you think you have the appropriate competences to work with us at this project, please contact us to genius@inova-ria.pt.