Flux Research Group / School of Computing

Performance Analysis of Virtual Environments

Nikhil Mishrikoti

Flux Technical Note FTN–2013–03, University of Utah. 2013.



Virtual machines are becoming an integral part of both industrial data center facilities and large academic experimentation environments. It is important to measure the effects of virtualization and identify performance bottlenecks in the virtual setups. This project aspires to provide tools to perform fine-grained analysis of resource utilization, virtualization overheads, and performance bottlenecks, in a virtualized software stack. We implement our performance analysis framework for the Xen virtual machine monitor, which is a full feature, open source hypervisor, and a de-facto industry standard for industrial data- centers.