Cutting the energy cost of TV content distribution by 5, by understanding the popularity of the top ten programs

In this talk, I will describe work we have been doing with a major TV company to reduce the energy costs of their content distribution, which currently uses 3 different types of network. In the near future, when the networks are unified, and content is delivered to tablets, set top boxes and digital receivers alike, there are opportunities for data mining users' interests to provide accurate predictors for which content should be delivered when and where using which mode of CDN. The results can be an impressive reduction in energy consumption compared with today's systems.