Title: Improving Power Efficiency forInternet Mobile Streaming Services
As mobile devices, e.g., smartphones andtablets, are gaining prevailing popularity in current era, Internet streamingservices are one of the most popular entertainment activities on this platform.However, the resources on mobile devices are constrained, particularly theinherently limited battery power supply. While multiple system components, includingthe CPU, display and the network interfaces, of a mobile device aresimultaneously involved in a typical streaming session, power saving techniquesare proposed and applied on them individually. As the CPU consumes less batterypower during a streaming session due to the widely supported hardware decoder,we mainly focus on the power saving techniques for the display and transmissionsubsystems. To effectively reduce the power consumption of thedisplay subsystem, we propose our Content Adaptive Display (CAD) power savingsystems for various streaming scenarios. In CAD, the GPU is widely adopted toaccomplish the computation intensive tasks. For the transmissionsubsystem, we design our adaptive-push and DASH2M based on the newlystandardized HTTP/2, whose server push mechanism is exploited to help wirelessnetwork interfaces enter the sleep mode more efficiently. Experiments show thatour schemes can maximize the user's quality of experience (QoE) while savingthe battery consumption.
Mengbai Xiao is a PhD candidate of theDepartment of Computer Science at George Mason University. His research areasinclude Internet streaming, mobile computing, and storage systems. His work hasbeen published on ACM Multimedia Conference, WWW, ACM NOSSDAV, etc.