الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 109 |  |  |
| | | from | 109 | | |
Abstract
The past decade witnessed a tremendous increase on mobile devices capabilities and excessive bandwidth demands. Moreover, the advances in mobile technologies coupled with the overwhelming amount of applications have enabled users not only access but also generate data in many forms. However, the mobile resource limitations such as computation capability, storage capacity and battery lifetime, have drawn significant research attention in order to allow computation to be offloaded and executed on cloud computing infrastructure. This thesis presents a framework to minimize the computation time and consumed energy while processing applications with intensive computations in mobile devices. The framework is called The RESTful Cloud-Assisted Mobile Computing Framework (RCAMC). The RCAMC framework is designed to provide a lightweight and energy-efficient architecture for computational offloading. It also considers the existence of a third party data provider, which minimizes the amount of data that should be sent to the cloud side and therefore, offloading brings more benefits. The different framework components track changes in the mobile context including network environment, the available cloud services, and battery level. In this framework, the mobile has two modes of operation. First, it acts as a user device that runs a certain application. Second, it acts as a service provider that have enough resources to be shared. The evaluation of the proposed architecture and algorithms is conducted using bench marking experiments to determine the application execution time and energy consumption. Sixteen various workloads are used as benchmarks in order to collect and analyze performance metrics, namely response time and consumed energy. Results show that, if the offloaded task contains a small data transfer coupled with a heavy computation, RCAMC provides a speedup range from 5−10 times and almost 80−90 % energy saving. Moreover, if the offloaded function transfers a large amount of data, offloading outcomes are dependent on the mobile device context i.e. network connection type and bandwidth steadiness.