Research On QoS-aware Virtual Machine Placement And Network Bandwidth Configuration In SDN Cloud Data Center

As an important part of internet architecture,cloud data center has provided the underlying support for more and more applications.However,the problems associated with high energy consumption and low resource utilization are becoming increasingly prominent,resulting in serious environmental pollution and waste of resources.Efficiently allocating resources in the cloud data center to meet various needs is not only beneficial to the interests of cloud service providers and customers,but also conforms to the basic national policy of resource conservation and environmental protection.Based on the research of cloud data center architecture and application architecture,multi-QoS grouping model is proposed in this thesis,and the virtual machine and network bandwidth allocation algorithm is implemented based on this model to ensure the network performance of important applications.The following work has been done:1.A QoS-aware multi-QoS grouping model that divides applications into different QoS levels.The current research on the efficient resource allocation of cloud data centers is inadequate.The resource allocation of applications with high QoS requirements is often caught in the dilemma of losing resource flexibility or failing to meet user satisfaction.To this end,a multi-QoS grouping model is proposed in this thesis.The K-means algorithm is used to perform clustering analysis on google cluster trace data,and the request is divided into multiple QoS levels.In the clustering process,full consideration is given to the parameters reflecting QoS requirements,such as priority,delay sensitivity,duration etc.The multi-QoS grouping model accurately divides the application's QoS requirements and provides a basis for efficient resource allocation in the cloud data center.2.A virtual machine and network bandwidth allocation algorithm based on the multiQoS grouping model.The heterogeneity of cloud data centers and the connectivity between application virtual machines are often ignored in current resource allocation schemes,which result in higher energy consumption and QoS conflicts.In this thesis,we analyzed the QoS level of application requests submitted to the cloud data center based on the multi-QoS grouping model.The proposed QoS-aware virtual machine placement algorithm places virtual machines of high QoS level applications on closely connected hosts,which reduces network congestion caused by other applications.The proposed QoS-aware network bandwidth allocation algorithm ensures the network bandwidth required by applications with high QoS levels by configuring priority queues on switches managed by the SDN controller.3.Verify the QoS-aware virtual machine and network bandwidth allocation algorithm.In this thesis,we modified the source code to implement algorithm in the Cloud Sim-SDN cloud simulation platform,and selected response time,QoS violation rate,energy consumption to verify the performance of the algorithm.Experimental results show that the algorithm proposed in this thesis has better performance than resource allocation algorithms such as round robin algorithm and first fit decreasing algorithm,especially for applications with higher network requirements.