DVFS And Thermal-aware Task Scheduling For Mobile Cloud Computing

With the rapid growth of mobile device hardware performance, mobile computing is more and more capable to meet people's computing needs. However, due to limitations of mobile device, the resulting computing experience is still not good enough in many cases. In order to expand the capabilities of mobile computing, mobile cloud computing is therefore proposed. As an extension of the cloud, mobile cloud computing faces many problems, such as the difference between the computing device, the communication complexity of the environment and energy consumption inefficiencies. So how to allocate computing resources more reasonably to make the overall cost reduction in this complex environment is a problem worth of studying.This paper studies those successfully applications which based on mobile cloud computing in various fields, and analyzes the advantages of mobile cloud computing. From these two aspects of communications and computing, the paper points out more specific problems of mobile cloud computing. With simple description of system process, the formal definition of the problem is given, then it is abstracted into a distribution problem. Taking into account the environment complexity and subsequent expansion of mobile cloud computing, a multiple factors representation model based on tensor is therefore be proposed. Next the three components of the energy consumption of mobile cloud computing- mobile devices, cloud and network energy consumption, is specifically analyzed. Through the detail modeling in each part of the energy, a global energy optimization model of mobile cloud computing is proposed, taking heterogeneity into account and increasing the versatility of the model.This allocation is actually an optimization problem, and in this paper, the optimal scheme to a given environment is solved based on simulated annealing algorithm. With the results, global energy consumption trends with deadlines and quality of experience is analyzed in different number of mobile devices, as well as a comparison of each part of the global energy. Also, a comparison of energy consumption with the existing model in the same situation is analyzed and it shows that the proposed model is better than the existing model.Studies have shown that DVFS technology will bring the problem of decreased reliability. In our study, DVFS is the major technology to reduce energy, so a detailed analysis of how DVFS affects the reliability of mobile cloud computing is followed. Next, a triple objective optimization problem of energy, reliability and Qo E is proposed, and a multi-objective simulated annealing algorithm combined with Pareto theory to solve this problem is given. The proposed model has guided significance for computing resource allocation in mobile cloud computing environments to save energy and to preserve the system reliability.