Research On Resource Scheduling Mechanism Based On QoS Benefit In Private Cloud

Cloud computing is the development combined product of grid computing, distributed computing, parallel computing, virtualization and other traditional technology and network. In cloud computing, dynamic expansion and load balancing technologies can organize and allocate resources high efficiently. Present methods of resource scheduling mechanism and guaranteed QoS don’t take full account of the integrated QoS benefit of the system in cloud computing, and private cloud has its own characteristics and requirements in resource management and the resource scheduling mechanism should be more targeted.This thesis aims at the development of small and medium-size private cloud, through the analysis of QoS parameters and constructing QoS benefit assessment model, is mainly to solve how to construct resource scheduling model and choose resource scheduling algorithm, in order to guarantee the QoS benefit. Through researching these problems, this thesis realizes the resource scheduling mechanism based on QoS benefit in private cloud YNU Cloud of Yunnan University. In this thesis, through the analysis of these issues, the main research contents are as follows:(1) Assessing the integrated QoS benefit of system for resource scheduling problems, this thesis analyzes the factors which affect the QoS of system, quantifies the parameters which affect the QoS performance, and establishes the mapping rules between the QoS parameters of cloud users and cloud resources, uses a QoS evaluation model to assess the QoS benefit of system, guarantees the QoS.(2) Based on the previous QoS benefit evaluation model, this thesis constructs a resource scheduling model based on QoS benefit in private cloud. Then the thesis proposes a resource scheduling algorithm based on genetic and ant colony hybrid algorithms, it takes full advantage of fast feature and global convergence capabilities of genetic algorithm, gets the initialization pheromone distribution of ant colony optimization at last, and then utilizes the positive feedback of ant colony optimization to improve the efficiency of algorithm and resource utilization. Experimental results show that the model and algorithm can achieve better QoS benefit.(3) This thesis compares open-source cloud platforms OpenStack and OpenNebula from architecture, virtualization hypervisor, scale and activity level of community and cloud security and so on, and selects OpenNebula as the basic research and development platform of private cloud YNU Cloud according to the results of comparison. For the proposed model and algorithm, realizes the resource scheduling mechanism based on QoS benefit in YNU Cloud based on the OpenNebula.