Research And Implementation Of Virtualization Management Framework For High Availability

In recent years, more and more users consider constructing virtualization management center. There are three main reasons:lower hardware cost, more economical space, more environmentally friendly energy consumption. Such way not only can help users migrate applications from a physical server to a virtual machine (VM), but also provide optimization of energy consumption and load sharing mechanism based on the hardware level, so as to improve the efficiency of computing resources.The high availability application of virtualization can provide a higher platform, of which high availability server clustering is as to be a typical representative. The high availability server clustering can ensure that if a server fails, the virtual machines are automatically restarted on an alternate server without disrupting operations for users. Clusters consist of two or more servers (physical or virtual) running with the same configuration. Clustering software can keep the application data updated on both servers and restart the backup server in the event of a failure on the primary server.As the high standards for virtualization management framework, high availability has been the hot spot of researches. In this paper, we provide a virtualization management framework for high availability, based on NEU-Convirture (the secondary development software of Convirt) and load balancing function provided by the LVS Cluster Suite of RHCS. We also provide the comparison of several kinds of load balancing algorithm in LVS, so as to propose a new load balancing algorithm to improve the performance of the system, which can make a variety of resources reasonably distributed between multiple computers, avoiding some nodes being leisure, some heavy.Firstly, through the analysis of Convirt and the second development NEU-Convirture, we construct application platform of virtualization management framework for high availability to make sure the feasibility, reliability of virtual storage, virtual server, cluster and virtual machine operation to complete the basic functional requirements.Secondly, through the analysis of structure and main functions of RHCS with the most important scheduling algorithms, including Weighted Round-Robin Scheduling (WRRS), Least-Connection Scheduling (LCS) and Weighted-Least-Connection Scheduling (WLCS), we propose the improved load balancing scheduling algorithm to improve the performance of the system and make sure that the virtualization management framework for high availability is feasible, reliable and available.At the end of the paper, we complete the basic functions of operations, test and compare the performance of load balancing algorithms. Through testing the virtualization management framework, we can conclude that the algorithm provides the high availability and a stable operating environment for the system.