Test And Performance Optimization Of Experimental Teaching Management System Based On Desktop Cloud

In recent years, cloud computing technology is developing rapidly, especially the use of desktop cloud, greatly changed the application model based on local computing. As an innovative service model in the field of information technology, desktop cloud is a new computing model, which can be gotten at any time and any place. Desktop cloud can be customized according to user’s needs, and it enables users to enjoy the results of technological development at low-cost, high efficiency, more flexible way. Experimental teaching management system based on desktop cloud, which applies desktop cloud to the traditional experimental teaching management, is an innovative mode of laboratory management. Since the system put into use, teachers and students, especially the laboratory managers, are interested in the system performance. But in the system testing, we find some unsatisfactory places in the system, such as poor user experience, vCPU drift, high memory consumption, high I/O utilization, etc. In this paper, after the completion of the basic functional testing of the system, we give a solution of the problems found in the test to optimize the system performance.In order to solve the problems found during the test, this paper analyses the causes of these problems. From the four aspects of CPU, memory, disk I/O, and network I/O, we propose a solution. By binding CPU, we have optimized the performance of vCPU in virtual machine; through KSM technology and large pages(Huge Page) technology, we have solved the problem of high memory consumption; and in the optimization of disk I/O and I/O network, we chose the virtio framework.This article chooses the functional testing and performance testing as the test method of the system. Before the test of the system, we chose some relevant test indexes, and then we introduce the testing process. First, we introduced a variety of testing methods and the purpose of test method; second, we give the test cases in detail. From the four aspects of CPU, memory, disk, network, we tested the performance of the system before and after optimization. When we analyze the test results, we introduced the analysis method of each performance index. Finally, through the test compared with VMware desktop cloud, we summarize the shortage of the system and carry out a further optimization scheme.