Energy-based anomaly detection in quality of experience

Subjective performance of smartphone-based high bandwidth- and energy-demanding applications and services such as video streaming, are highly influenced by the temporal impairments perceived by the user at the user interface; and the application´s energy consumption patterns. Therefore, we study the influence of the anomalies detected by objective measurements from the user interface and the network-level, on the power consumption during video streaming on the smartphone. In this paper, we study the inter-frame time (D_f) metric, i.e., the time gap between two consecutive displayed pictures at the user interface; the inter-packet time (D_p) and the initial signaling duration (D_s) at the network-level; and the instantaneous power consumption (P_n) at the power supply of the smartphone. We conduct experiments on the VLC media player, while streaming video via local-storage; via 3rd Generation (3G) using Real Time Streaming Protocol (RTSP) protocol; and via 3G using Hypertext Transfer Protocol (HTTP) protocol. We show that the anomalies detected at the instantaneous power consumption reveals the anomalies at the user interface and the network-level.