Research On Virtual Machine Placement By Glowworm Swarm Optimization In Cloud Computing Data Center

With the rapid development of information technology and internet technology, cloud computing is gradually developing as a new computing method and a new business model. The new generation data center is the integration product of cloud computing and data center. And it requires high resource utilization, environmental protection and high-quality service. Virtualization technology can abstract huge physical resources as virtual resources. This technology benefits the integration and unified management of resources, thus realizes the flexible scheduling and full use of resources. The construction of cloud computing data center need virtualization technology as the core technology, and the virtual machine placement is the key problem of the virtualization technology. Therefore the virtual machine placement research has important practical significance and practical value.The virtual machine placement problem is a combinational optimization problem, and swarm optimization algorithms are widely used in solving combinational optimization problems. Glowworm swarm optimization is one of the swarm optimization algorithms, and it has the characteristic of simple and easy-to-implement, strong robustness, etc. At present, this algorithm has been successfully applied in multi-peak problems and continuous problems, but has some limitations in solving discrete problems.In this dissertation, the relevant theories of cloud computing data center and virtualization technology were introduced. Then the virtual machine placement problem was analyzed and a multi-objective mathematical model was built with the comprehensive consideration of resource wastage, power consumption and load balance. Finally, the placement problem was abstracted as a bin packing problem, and it was solved by an improved adaptive discrete glowworm swarm optimization. By analyzing the simulation experiment results, it was proved that the proposed adaptive discrete glowworm swarm optimization has a great robustness and a fast convergence rate, and the proposed intelligent optimization strategy can solve the virtual machine placement problem preferably.