Research On Task Allocation Strategy In Mobile Cloud Computing

Mobile cloud computing is a new application service mode of integration of cloud computing and mobile Internet, and it provides strong support for the mobile Internet application services to meet the higher standards of user service requirements by using strong mature technology of cloud computing. With the rapid development of mobile Internet services, performance improvement of the mobile terminal device and upgrade of the wireless network, it is a new direction for future development of mobile cloud computing that the mobile device as a resource provider involved in cloud computing task processing. Although the existing research results prove the feasibility of this direction, it still faces many problems. On the one hand, the inherent attributes of mobile devices, such as mobility, low connectivity and limited power, makes it difficult to exploit the full potential of mobile devices. On the other hand, the lack of reasonable task allocation strategy makes the task and mobile computing capacity could not reasonably matched, resulting in the low efficiency and waste of resources.The application mode which taking mobile devices as resource providers is researched in this paper. The mobile MapReduce framework that the mobile devices involved in cloud computing task processing is designed and the ability of mobile devices as the task execution node of mobile cloud computing is measured. In addition, considering the characteristics of mobile devices, such as limited computation ability, network bandwidth and limited power, the task allocation algorithm is proposed to minimize the task completion time and devices energy consumption, improving the system resource utilization and meeting user’s differentiated service requirements.The main work of this paper includes:(1) The mobile MapReduce framework is designed that the mobile devices involved in cloud computing task processing. The MapReduce framework is improved to use mobile devices for MapReduce tasks task processing. The proxy server is introduced to manage and maintain the whole mobile cloud computing system. The JobTracker is implemented by a reliable proxy server to improve the reliability and stability of the mobile MapReduce system, and the TaskTracker is realized by the mobile devices to achieve the distributed computation.(2) The mobile device performance is measured in the mobile MapRedcue, including the availability and mobility of mobile devices. The mobility measurement of mobile devices is the focus of attention. A mobility measure algorithm is proposed that the entropy is used to calculate the trajectory of mobile device for mobility measurement. And based on the entropy measure method, a new mobility measure algorithm that increased inter symbol relations is proposed finally, ordering the mobility of mobile devices.(3) The task allocation algorithm is proposed for mobile cloud computing environment. Considering the characteristics of mobile devices, such as limited computation ability, network bandwidth and limited power, three task allocation algorithms based on genetic algorithm are designed to minimize the task completion time and devices energy consumption. The users can choose appropriate task allocation algorithm according to their own requirements.The experimental results show that the proposed mobility measurement algorithm can not only improve the measurement accuracy, but also can effectively reduce trajectory length. And the proposed task allocation algorithm can reasonably assign tasks for mobile devices, and effectively reduce task completion time and total energy consumption of devices, improving the system resource utilization.The main contribution and innovation of this paper are as follows:(1) The mobility measure algorithm is proposed for measuring mobility of large-scale mobile devices. By processing the trajectory of the mobile device, the mobility of mobile devices is sorted. And Combined with the availability of mobile devices, the high performance mobile devices were selected for task execution.(2) The task allocation algorithm is proposed for mobile cloud environment. It can reasonably assign tasks for mobile devices according to performance differences among computation capability, network bandwidth and device power. In addition to proposing task allocation algorithm with the shortest task completion time, the task allocation algorithm that takes into account the energy consumption of device to perform the task is also proposed, meeting user’s differentiated optimization requirements.