Date of Award
Master of Science (MS)
The thesis implements an offloading framework for GoogleTM AndroidTM based on mobile devices. Today, the full potential for smartphones may be constrained by certain technical limits such as battery endurance and computational performance. Modern mobile applications own more powerful functions but need larger computation and faster frame rate, which consume more battery energy. Using the proposed offloading framework, mobile devices can offload computational intensive workload to servers to save battery energy consumption and reduce the execution time. The framework can also enable software developers to easily build and deploy services on the servers to support mobile devices to run computationally intensive jobs. Compared with other offloading schemes for android cell phones, the scheme enables developers to choose which parts of the codes are potentially offloading. As developers fully understand the data flow models of the apps, they are considered most capable of making offloading decisions. Developers can minimize communication overhead brought by offloading by carefully partitioning source code by data dependency. Experiment results and data showed that the proposed offloading scheme could significantly reduce computational time and battery energy consumption.