Downtime Analysis of Virtual Machine Live Migration
Felix Salfner, Peter Tr, Andreas Polze
Abstract
Felix Salfner, Peter Tr, Andreas Polze
Abstract
Virtual machine environments like VMware, XEN, KVM, and Hyper-V support live migration of guest operating systems, which is used in data centers to provide uninterrupted service during maintenance or to move computation away from failure-prone hosts. The duration of migration, as well as the virtual machine downtime during this process are essential when assessing if service availability agreements might be violated. We present the result of an experimental study that analyzed virtual machine live migration downtime and duration. We show that total migration time as well as downtime are dominated by specific memory utilization patterns inside the virtualized guest. We experienced that downtime involved by live migration can vary by a factor of more than 23, which can have significant impact on service availability. Keywords-virtual machine, live migration, downtime I. INTRODUCTION
OpenAlex reports 51 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Virtual machine environments like VMware, XEN, KVM, and Hyper-V support live migration of guest operating systems, which is used in data centers to provide uninterrupted service during maintenance or to move computation away from failure-prone hosts. The duration of migration, as well as the virtual machine downtime during this process are essential when assessing if service availability agreements might be violated. We present the result of an experimental study that analyzed virtual machine live migration downtime and duration. We show that total migration time as well as downtime are dominated by specific memory utilization patterns inside the virtualized guest. We experienced that downtime involved by live migration can vary by a factor of more than 23, which can have significant impact on service availability. Keywords-virtual machine, live migration, downtime I. INTRODUCTION
Key concepts: Downtime, Live migration, Virtual machine, Computer science, Service (business), Operating system, Process (computing), Duration (music)