The side-effects of virtualization to benchmarks are long time known. However virtualization technologies are advancing, and their footprint is smaller and smaller. In addition, in cloud environment user can never be sure if you got a virtual machine or real hardware or container, or contaner in virtual machine.
In this thesis, student will investigate how virtualization, and especially nested virtualization, technologies are affecting performance and stability of various Java benchmark types – CPU, time, network, bytecode, IO.
Anohter impact may be from the benchmarked software itself. Some software, including java, can be virtualization-aware and adapt themselves. Thus student will use JDK with self-awareness off, partial self awareness, and full awareness. The thesis will focus on running benchmarks in several virtualization scenarios, and in-depth analysis of results.