Performance evaluation of virtual desktop operating systems in virtual desktop infrastructure

Desktop as a Service (DaaS) is a desktop virtualization service hosted in the public or private cloud. Virtual Desktop Infrastructure (VDI) is the key technology that enables virtual desktops. The virtual desktop that runs in a data center is delivered to the end user's device by using Remote Desktop Protocol (RDP). Currently, resource allocation in the VDI has become one of the key issues and big challenges for the organizations which want to migrate from traditional desktop to the virtual desktop. The main objective of resource allocation in the VDI is to provide resources to the virtual desktop without over-provisioning while reducing resource usage and response time. We believe resource usage and response time can be reduced without over-provisioning by selecting an Operating System (OS) in the VDI which has fewer RAM utilization, CPU response time and application response time before calculating virtual desktop density and deploying VDI in a large scale. The aim of this paper is to evaluate OS 1: Windows 8.1 Enterprise 32-bit and OS 2: Windows 10 Enterprise 32-bit as dedicated virtual desktops in the virtual desktop infrastructure by using different workloads to find which operating system has fewer resource usage and response time in terms of memory usage, CPU and application response time. We used experimental design and quantitative approach. The results indicate that OS 1 has the lowest memory usage and CPU response time than OS 2. In addition, results of application response time comparison between OS 1 and OS 2, show that OS 1 has the lower application response time than OS 2 as well.

[1]  Kamal Z. Zamli,et al.  Evaluation of Virtual Desktop Infrastructure host using User Workload Simulator , 2015, 2015 IEEE Conference on Open Systems (ICOS).

[2]  Li Chen,et al.  Research on virtual desktop management based on cloud computing , 2015, 2015 International Conference on Logistics, Informatics and Service Sciences (LISS).

[3]  Yongxin Zhu,et al.  Prototyping Efficient Desktop-as-a-Service for FPGA Based Cloud Computing Architecture , 2012, 2012 IEEE Fifth International Conference on Cloud Computing.

[4]  Filip De Turck,et al.  User subscription-based resource management for Desktop-as-a-Service platforms , 2014, The Journal of Supercomputing.

[5]  Makoto Kayashima,et al.  Desktop workload characteristics and their utility in optimizing virtual machine placement in cloud , 2012, 2012 IEEE 2nd International Conference on Cloud Computing and Intelligence Systems.

[6]  Regina Miseviciene,et al.  Educational Infrastructure Using Virtualization Technologies: Experience at Kaunas University of Technology , 2012, Informatics Educ..

[7]  Filip De Turck,et al.  Efficient resource management for virtual desktop cloud computing , 2012, The Journal of Supercomputing.

[8]  Lin Cai,et al.  CloudRank-V: A Desktop Cloud Benchmark with Complex Workloads , 2013, 2013 IEEE 10th International Conference on High Performance Computing and Communications & 2013 IEEE International Conference on Embedded and Ubiquitous Computing.

[9]  Khaled Salah,et al.  VDC-Analyst: Design and verification of virtual desktop cloud resource allocations , 2014, Comput. Networks.

[10]  Asoke Nath,et al.  Virtual Desktop Infrastructure in Higher Education Institution: Energy Efficiency as an Application of Green Computing , 2014, 2014 Fourth International Conference on Communication Systems and Network Technologies.

[11]  Ahmad Shawahna,et al.  Performance evaluation of VDI environment , 2016, 2016 Sixth International Conference on Innovative Computing Technology (INTECH).