Evaluation of type-1 hypervisors on desktop-class virtualization hosts
Journal: IADIS INTERNATIONAL JOURNAL ON COMPUTER SCIENCE AND INFORMATION SYSTEMS (Vol.12, No. 2)Publication Date: 2017-12-14
Authors : Duarte Pousa; José Rufino;
Page : 86-101
Keywords : Bare Metal Virtualization; Synthetic Benchmarking; Performance Assessment; Commodity Hardware;
Abstract
System Virtualization has become a fundamental IT tool, whether it is type-2/hosted virtualization, mostly exploited by end-users in their personal computers, or type-1/bare metal, well established in IT departments and thoroughly used in modern datacenters as the very foundation of cloud computing. Though bare metal virtualization is meant to be deployed on server-grade hardware (for performance, stability and reliability reasons), properly configured desktop-class systems or workstations are often used as virtualization servers, due to their attractive performance/cost ratio. This paper presents the results of a study conducted on commodity virtualization servers, aiming to assess the performance of a representative set of the type-1 platforms mostly in use today: VMware ESXi, Citrix XenServer, Microsoft Hyper-V, oVirt and Proxmox. Hypervisor performance is indirectly measured through synthetic benchmarks performed on Windows 10 LTSB and Linux Ubuntu Server 16.04 guests: PassMark for Windows, UnixBench for Linux, and the cross-platform Flexible I/O Tester and iPerf3 benchmarks. The evaluation results may be used to guide the choice of the best type-1 platform (performance-wise), depending on the predominant guest OS, the performance patterns (CPUbound, IO-bound, or balanced) of that OS, its storage type (local/remote) and the required network-level performance.
Other Latest Articles
- Dynamic identification by enumeration for co-operative knowledge discovery
- Integration of context-information to support spatial decision systems
- Vhdl design for a speed limit sign segmentation algorithm
- Radial vs. rectangular: evaluating visualization layout impact on user task performance of hierarchical data
- Cliques detection vs maximum spanning tree for tweet contextualization
Last modified: 2019-12-13 21:04:28