Examining Effects of Virtual Machine Settings on Voice Over Internet Protocol in a Private Cloud Environment

Abstract

The virtualization of computing resources, as represented by the sustained growth of cloud computing, continues to thrive. Information Technology departments are building their private clouds due to the perception of significant cost savings by managing all physical computing resources from a single point and assigning them to applications or services as needed while remaining in control of their systems and information. As part of this trend, real time communication applications including IP telephony can be integrated with other software applications into one platform and deployed in private clouds to reduce capital expenditure and lower overall costs of daily based maintenance and real estate required for computer hardware. As a critical component of private clouds, however, virtualization may adversely affect a real time communication application running in virtual machines as the layer of virtualization on the physical server adds system overhead and contributes to capacity lose. While real time communication services require a certain level of system performance and availability to address communication latency and overhead bottleneck, it is essential to investigate potential performance implications of private clouds on IP telephony applications. The purpose of this study was to investigate how to apply cost benefits of private clouds to Voice over Internet Protocol applications without compromise on communication performance. Through the experimental study, the statistical technique ANCOVA was used to examine the effects of virtual machine settings on voice quality when network condition remained the same. Linear Regression analysis was used to test whether the voice quality can be predicted from virtual computing resources and network bandwidth in private clouds. The results of this research provided a better understanding of the effects of virtual machine settings on voice quality of Voice over Internet Protocol applications in private clouds from the prospect of technology management.