Hypervisor and Virtual Machine Memory Optimization Analysis

Abstract

The goal of this thesis is to test memory optimization and reclamation tools in the most widely used hypervisors: VMware ESXi, Microsoft Hyper-V, KVM, and Xen. The aim is to measure how much memory could be reclaimed and optimized by different memory management algorithms across hypervisors mentioned above. Dedicated monitoring tools Zabbix and collectd are going to gather the data which will be analyzed. As a result, Hyper-V seems to be the most effective, with ESXi second and KVM falling somewhat behind in the third place. Xen failed to meet specifc criteria (automated memory optimization) which rendered it impractical to include in the testing process.