Abstract: Performance loss ratio of multiple virtual machines on a single physical server is the key challenge in virtual environment. In this research, it is concern about the eight specific conditions as the performances of virtual machine, through quantities analysis approach during period of the execution in the Dedicated Online Education System. The factors that impact on the performance measurement of virtual machine can be divided into five key parameters such as CPU Usage, Idle time, Memory Utilization, Network Latency and I/O wait time. The virtual machine performances while reducing the number of virtual machines and increasing user level, has been given a positive improvement for the first four parameters and slightly negative for the last parameter. The results shown that less number of virtual machine can be accessed for more clients based on above parameters with better performance. Keywords – Virtual Machine, Performance, Online Education System I. INTRODUCTION Web-based Education system is more effective than other OE systems, it still continue drawbacks such as Performance, Storage, implementation, maintenance, building space and security issues [1]. Many developers recommended virtualization based approaches as the best solution for above overhead. Virtualization technology provides benefits to such kind of OE system in performance isolation, online system maintenance, live migration, Customized OS and Virtual Machine (VM) which have emphasized the ease of system management and administration [2]. The performance overhead occurs commonly in the VM during the execution of application. Under this circumstance, developers consider vertical optimization or optimize the I/O or Hypervisor or Kernels are the solution to improve the above overhead. According to the previous knowledge summarized we decided that optimization of VM is a better solution for the performance overhead. There are many techniques available when creating the VM with the use of hypervisor or VMM, which can be implemented directly or separately on Physical Machine (PM) [2]. Hypervisor multiplexes across the VMs and ones situated between the PM and VMs [4]. Multiple VMs are also able to run on a unique PM, but the capacity of Manuscript received November 03, 2015; revised November 03, 2015. This work was supported in part by the Huazhong University of Science and Technology – Virtual Machine Performance Approaches in the Online Education System. Basheer Riskhan is with the Department of Computer Science, Huazhong University of Science and Technology, Wuhan, China. Phone: 0094714469676; Fax: 0094372224720; E-Mail: riskhan@ yahoo.com). Raza Muhammad is with the Department of Computer Science, Huazhong University of Science and Technology, Wuhan, China razacom_2000@yahoo.com). the physical resources should be enough to fulfill the need of all VMs working on it [5]. This study attempts to analyze the performance of VM using Dedicated OE System (DOES) model based on major eight conditions. CPU Usage, I/O wait time, idle time, Memory Utilization (MU) and Network Latency (NL) are the Variable measured during the execution of DOES model [4]. Our contribution to this paper; 1) Decrease the VMs or minimize the needed resource 2) Offer multi-users to access OE system in a single VM 3) Measure the matrix and report the overall detailed performance impact of each conditions. The VM optimization experiment result has indicated that our approach has improved performance efficiently and effectively. II. RELATED WORK This research is related to the VM Performance which was based on the VM and OE system. OE system has widely used in the form of client server technology and web base technology. But still continue some limitation due to processing memory, power constrains and different device platform. These overheads create issues to use the OE system properly and the performances of the devices not equally support to the OE system. To solve above mentioned overhead authors express, Cloud Computing is significant solution which provides the resources on demand [8]. Reference [7] addresses the application performance on VM and PM. Virtualization rescues many wastage cycles during the execution of applications. Apache JMeter was used as an experiment tool, experiment as followed by certain condition such as workload performance of PM which increases the CPU level in PM, workload performance of VM which increase the VCPU level in VM. Finally suggest that the benefits of virtualization application served by VM better than PM. III. PLATFORM ARCHITECTURE OVERVIEW OF DOES SYSTEM DOES model process among users, VM and PM. VM have given the permission to access the OE Application according to their resource limit. Beyond on VM limits, Hypervisor creates the VM and share the resources from PM. This experiment tests while users accessing the DOES model and measured the each variable matrix for the duration of 80 seconds. A. Introduction of Operation Principal of DOES Model Figure 1 describes the operation principle of DOES Model for multi user access in single VM. The Client request reaches to the VM after checking the personal identity. Then Virtual Machine Performance Approaches in the Online Education System Basheer Riskhan, Raza Muhammad Proceedings of the International MultiConference of Engineers and Computer Scientists 2016 Vol I, IMECS 2016, March 16 - 18, 2016, Hong Kong ISBN: 978-988-19253-8-1 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online) IMECS 2016