Predicting the Performance of Virtual Machine Migration

TLDR

Live migration, the ability to move virtual machines between physical hosts, is a core feature of virtualisation. The study aims to predict migration times accurately by characterising parameters affecting live migration, especially on the Xen platform. The authors characterise migration‑affecting parameters, focusing on the Xen virtualization platform. The study shows that migration performance varies with workload and presents two simulation models that predict migration times within 90 % accuracy for synthetic and real‑world benchmarks.

Abstract

With the ability to move virtual machines between physical hosts, live migration is a core feature of virtualisation. However for migration to be useful, deployable feature on a large (datacentre) scale, we need to predict migration times with accuracy. In this paper, we characterise the parameters affecting live migration with particular emphasis on the Xen virtualisation platform. We discuss the relationships between the important parameters that affect migration and highlight how migration performance can vary considerably depending on the workload. We further provide 2 simulation models that are able to predict migration times to within 90% accuracy for both synthetic and real-world benchmarks.