Heterogeneous Grid Computing for Energy Constrained Mobile Device

Min Li,Xiaobo Wu,Menglian Zhao,Hui Wang,Xiaolang Yan
DOI: https://doi.org/10.1007/978-3-540-30121-9_34
2004-01-01
Abstract:There is a lot of existing research focusing on low power mobile computing, but only a fraction of it takes systematic view toward the whole system. Our research is motivated by two important facts: the first is that processor, SDRAM and chipset consume more than 60% energy on mobile device, and the second is that mobile device usually moves inside a heterogeneous "computer community" that are interconnected by heterogeneous network. Targeting on low power mobile computing, we propose heterogeneous grid computing that enables mobile device to migrate work load to other computers, which have far more resources and are powered by AC. The off-loading can reduce majority of local computation related energy consumption. Meanwhile, since most mobile applications have asymmetric UI, the energy consumption on wireless interface will not be significantly increased. In order to support the proposed heterogeneous grid computing, we develop an UPnP based architecture that leverages a lot of industry standard. Three applications are developed on our tested: multimedia streaming, 3D game and Text to Speech. Experiments show that battery life time is radically prolonged.
What problem does this paper attempt to address?