This study focused on comparing performance with Apache running on Ubuntu compared with a bare-metal server based on the theoretical framework queuing theory model. The theoretical framework, used in this study imitate server resources such as CPU utilization, memory utilization, and response time, as they predict the estimates of utilization of servers. Open source tools and the Linux operating system along with Windows 10 were the platforms for the test. The problem with the poor performance of a virtualized server compares quantitatively with a bare-metal server. The central processing unit (CPU) utilization, memory utilization, and average response time values were measured and compared between a virtual and bare-metal server. This researcher implemented the methodology by using JMeter to run a load against each environment and quantitatively compared. The Mann-Whitney U test indicated that CPU utilization was greater for the virtualized server (Mdn = 14.71) than for the bare-metal server (Mdn = 6.750), U = .000, z = -7.897, p < .000, r = -1.2186. This finding means that the bare-metal server's CPU used fewer resources and worked less to do the same amount calculations. The Mann-Whitney U test indicated that memory utilization was greater for the bare-metal server (Mdn = 33.292) than for the virtual server (Mdn = 32.403), U = .006, z = 2.729, p < .000, r = .421. The nonparametric test for memory utilization showed that the smaller utilization demonstrated that there was more free memory remaining during the test runs for the virtual server that showed that the bare-metal server performed worse than the virtual server. Concerning average response time, the Mann-Whitney U test indicated that there was no significant difference for virtualization (Mdn =. 583) compared with the bare-metal server (Mdn = .584), U = 1.252, z = 1,014, p = .211, r = .193. The overall conclusion indicated that bare-metal servers performed better (CPU utilization) than virtualized servers but worse for memory utilization when using Apache hosted on Ubuntu on Linux but no difference concerning average response time. A recommendation for future research may involve using another Linux distribution, i.e., Fedora, Debian, and openSUSE. |