AbstractsPhysics

With the exponentially growing smartphone market, more and more people desire to have a multipurpose handheld device that not only supports voice communication and text messaging, but also provides video streaming, multimedia entertainment, etc. A crucial problem with a handheld device that enables video streaming is how to prolong the battery lifetime given the large amount of energy consumed by video transmission, decoding, and presentation. Thus, it is essential to have an in-depth understanding of power consumption required by video transmission, decoding, and presentation. The knowledge can be utilized to identify power-hungry components and to provide insight into how power consumption can be reduced for such components. Our experiments show that energy is mainly consumed by the wireless radio, the application processor, and the display system in a typical handheld device. More specifically, our research focuses on the streaming services where video is streamed over the 3G/4G network, decoded on ARM application processors and rendered on HD displays. To tackle the problem, we first propose power optimization algorithms for the 3G/4G radio, the ARM processor, and the display subsystem individually. By integrating all the algorithms, we build GreenVideo, a framework for energy-efficient video streaming to handheld devices. The system is validated with a large amount of real world videos from YouTube and the experimental results show that GreenVideo achieves significant power reductions.