A Human-centered Approach to make Networked Entertainment Green: A Case Study of CDN

Abstract

In our day to day lives, we are becoming more re-liant on clouds, which results in an excessive energy consumption by cloud infrastructures. In 2016, the US Department of Energy found that datacenters in the US made up roughly 2 % of all electricity consumption nationwide. Designing a cloud computing system that runs on sustainable energy and simultaneously satisfies increasing user demands with high availability is still an unsolved challenge. We focus on networked entertainment systems in this short paper. The first goal of this paper is to provide a broad research agenda on using a human-in-the-loop design to tackle this challenge, discussing more problems than solutions. Second, we present our design of an interactive Content Delivery Network (CDN) that is the backbone for much networked entertainment, e.g., Netflix and YouTube. Our CDN solution (i) considers user preferences and their tolerance (to unavailability) when tuning system configuration and design, and (ii) uses system metrics and availability predictions to encourage users to adopt more energy- efficient behavior. Finally, we advocate designs that integrate expertise from multiple disciplines, including Human-Computer Interaction, Machine Learning, and Cloud/Pervasive computing, and argue why cloud platforms can benefit from adopting such a holistic approach.