Research On Energy-Efficient Resource Allocation In Cloud Computing Environment

Recently, demand of large-scale computing is growing day by day, so as a better solution, cloud computing is becoming a well-known mode of calculation. Through aggregating and dynamic reusing resources, cloud computing achieves to optimize the service process and to improve the utilization of energy. With the booming developing of cloud computing, lots of research on its related technology has been done.. As the scale of server clusters of cloud computing is large, how to allocate the resource to improve the utilization efficiency of energy has become a research focus. How to implement a better energy-saving resource allocation algorithm is the key issue in this thesis.This thesis begins with a systematic introduction to the basic knowledge of cloud computing, including its origin, definition, classification and key technologies. On this basis, considering that the main application scenarios of cloud computing is providing services for numerous clients, the needs of clients of cloud computing are analyzed, and the resource allocation problem in cloud computing environment is decomposed and analyzed on each level combined with client’s needs, to clear the core mechanism and demand of resource allocation problem. Then, this thesis chooses to work on the angle of virtual machine initial allocation and migration to achieve a primary purpose of energy saving by the premise of not affecting user experience. With the goal, it gives a clear thinking of virtual machine initial allocation and migration combined with relevant theory and existing research results, In the end, the advantages and disadvantages of some existing resource allocation algorithms is studied, and a new virtual machine initial allocation algorithm, two-layer energy saving algorithm, is proposed according to the thinking by design changing and algorithm optimizing, combined the classic solutions and research results of problems such as bin packing problem, sever power consumption issues and plan selection problem. In this thesis, the main thought and procedure of the two-layer energy saving algorithm is introduced. A simulation platform is established by using java language, which imitates the structure of CloudSim, an open source cloud computing simulation platform. Through the simulation, the energy consumption data of our new algorithm and two existing algorithms is obtained and analyzed. By comparing the characteristics of energy consumption of the two-layer energy saving algorithm with the two existing algorithms, the effectiveness and superiority of two-layer energy saving algorithm in energy saving is proved.The two-layer energy saving algorithm proposed in this thesis provides a new solution for the resource allocation problem in cloud computing environment, in which the heterogeneous of the servers of cloud computing and the requirements of dynamic allocation of resources are fully considered. The new algorithm is close to the demand of actual application environment, and has an important significance in improving energy efficiency in cloud computing.