A lightweight framework for intensive mobile application processing in mobile cloud computing / Muhammad Shiraz

Muhammad , Shiraz (2013) A lightweight framework for intensive mobile application processing in mobile cloud computing / Muhammad Shiraz. PhD thesis, Universiti Malaya.

PDF (Thesis PhD) Download (2938Kb)

Abstract

Mobile Cloud Computing (MCC) enables computational intensive and ubiquitous mobile applications by leveraging the services of computational clouds. Human dependency on contemporary smartphones increases rapidly in various domains such as enterprise, e-learning and entertainment, gamming, management information systems, and healthcare. However, mobile applications on the latest generation of smartphones and tablets are still constrained by battery power, CPU potentials and memory capacity of the Smart Mobile Devices (SMDs). Therefore, MCC employs computational offloading as a significant software level solution for alleviating the resources limitations in SMDs. Recently, a number of computational offloading frameworks are proposed for the processing of computational intensive mobile applications in MCC. The traditional computational offloading frameworks implement intensive techniques for computational offloading in MCC which results in high energy consumption and longer turnaround time of the mobile applications. Therefore, lightweight techniques are imperative for the processing of computational intensive applications in MCC. Lightweight techniques enable computational intensive mobile application deployment and execution with minimal resources utilization on SMDs. As a result, mobile users can utilize distributed cloud services with lower computational load on mobile devices, shorter turnaround time of the application and longer lasting battery lifetime. This research investigates the resources intensive features of traditional computational offloading frameworks and proposes a lightweight framework for the processing intensive mobile applications in MCC. The additional cost of runtime computational offloading is investigated by implementing application offloading mechanism in the real mobile cloud computing environment. Distributed and Elastic Application Processing (DEAP) framework is proposed as a lightweight solution for the intensive application processing in MCC. DEAP framework reduces the cost of migration of application binary file and data file of the running instances of the mobile application. As a result, the size of data transmission over the wireless network medium, turnaround time of the intensive operations and energy consumption cost on mobile device is reduced considerably. DEAP framework is evaluated in the emulation environment on the Android virtual device instance. The performance of DEAP framework is validated by benchmarking prototype application in the real mobile cloud computing environment. Results of different experimental scenarios are compared to validate the lightweight nature of DEAP framework. It is found that by employing DEAP framework the cost of migration of application binary file and data file of the running instances of the application is reduced. As a result, the size of data transmission over the wireless network medium, turnaround time of the intensive operations and energy consumption cost on mobile device is reduced. DEAP framework reduces resources utilization and the cost of distributed processing of the prototype mobile application in MCC as follows: RAM allocation on mobile device by 71.5 percent, CPU utilization on mobile device by 55 percent, the size of data transmission over the wireless network medium by 84 percent, turnaround time of the application by 79.8 percent and energy consumption cost by 81 percent. Hence, DEAP framework provides a lightweight application layer solution for intensive mobile application processing in MCC.

Actions (For repository staff only : Login required)