Using Viewing Statistics to Re- duce Wasted Traffic and En- ergy Consumption in HTTP Adap- tive Streaming

Art der Arbeit:
Masterarbeit Informatik
Abgeschlossene Arbeit
Volker Gruhn


The energy consumption of mobile devices is a growing concern. Smartphones improve
their capabilities faster than their batteries. Resource hungry applications like streaming
services consume a lot of energy. The wireless interface, which uses a lot of energy, is
empirically configured to provide a good performance for most applications. Adaptive
video streaming, which is a popular streaming application, inefficiently utilises the wireless
interface and causes traffic overhead. We develop an adaptation strategy, which utilises
viewing statistics to reduce the energy consumption and traffic overhead. Additionally, the
adaptation strategy optimises the QoS, to improve the users video streaming experience.
We test and evaluate our strategy, using a simulation, which simulates users in a mobility
scenario. We simulate users, who are watching multiple videos and can skip and abandon
theplayback. Weshowthatenergyandtrafficcanbesaved, withmostlynegligibleinfluence
on the QoS. Energy efficiency and traffic saving mostly describe a tradeoff. Different
contexts offer different opportunities to save either more energy or more traffic. Further,
we show, that throughput fluctuations are an important influence factor for our adaptation