Title :
A dynamic computation offloading framework for Android
Author :
Deqian Kong ; Tao Qi ; Tan Yang ; Yidong Cui
Author_Institution :
Sch. of Comput. Sci., Beijing Univ. of Posts & Telecommun., Beijing, China
Abstract :
The poor processing ability and battery life make Android smartphone difficult to execute some complex applications. Computation offloading intends to solve this problem by executing parts of the application on a remote server. This paper presents DPSF, a dynamic computation offloading framework, which automatically partition the Android application into two parts at the compile stage. Developers should import DPSF SDK library into their project. After compile there will generate an apk file and a common jar file, which will be executed on the local phone and remote server respectively. Benefit from the powerful server, applications can reduce the execution time a lot and then save power. The typical Five Chess and 3D Motorcycle Racing game have been used to evaluate this framework. Evaluation shows that applications can obviously reduce power consumption by 21%-87%.
Keywords :
Android (operating system); mobile computing; power aware computing; smart phones; 3D Motorcycle Racing game; Android application; Android smartphone; DPSF SDK library; complex applications; dynamic computation offloading framework; local phone; power consumption; remote server; Electronic publishing; Information services; Internet; Java; Power demand; Servers; Smart phones; Android; Computation offloading; Framework; Partition;
Conference_Titel :
Broadband Network & Multimedia Technology (IC-BNMT), 2013 5th IEEE International Conference on
Conference_Location :
Guilin
DOI :
10.1109/ICBNMT.2013.6823930