Advances, Systems and Applications
Table 3 Comparison of Live VM migration models
From: A critical survey of live virtual machine migration techniques
Approach | Objective | Technique | Performance metrics | Hypervisor /Simulator |
|---|---|---|---|---|
Huang et al. [71]-2011 | Select the software and hardware environments that gives the best live migration performance | Virt-LM, compare live migration performance on different software and hardware stacks | Compare the migration performance on different CDC environments | Xen 3.3 and KVM-84 |
Kikuchi et al. [51]-2011 | Construct a performance model of concurrent live migrations | The performance characteristics of live migration represent using PRISM model | Verify quantitative properties regarding live migration performance (Operation orchestration and Resource provisioning) | XenServer 5.6 |
Wu and Zhao [72]-2011 | Predict migration latency | Build the performance model using statistical methods such as regression | Availability of resources have an impact on migration latency | Xen 3.2.1 |
Liu et al. [69]-2011 | Quantitatively predict the energy cost and migration performance | Design a high-level linear model to estimate the migration energy | More than 90% prediction accuracy in measured cost, reduce the migration cost by more than 72.9% at an energy saving of 73.6% | Xen 3.4.1 |
Cerroni, and Callegati [74]-2014 | Proposed a model for cloud-based edge network | Considered sequential and parallel migration strategies scheduling alternatives | Optimize service downtime, total migration time and network bandwidth | Mathematical modeling |
Deshpande et al. [75]-2014 | Determines how fast the source server goes into offline mode | Find eviction time by Scatter-Gather method | Reduces the VM eviction time and maintain total migration time against pre-copy and post-copy | KVM/ QEMU 1.6.50 |
Xu et al. [52]-2014 | Focus on the cost and performance interference while handling VM migration | Design and implement of iAware, to avoid violations of performance SLAs | Load balancing and power saving can achieved without affecting application performance | Xen |
Zhang et al. [81]-2014 | Memory and storage migration over WAN | Propose a flexible and adaptive migration framework | Achieves better adaptiveness for various applications over a WAN | Xen 4.1.2 |
Deshpande and Keahey [82]-2017 | Reduction in network contention of migration traffic and VM traffic | Presented a traffic-sensitive approach for migration of co-located VM’s | Reduces application degradation and total migration time | KVM/ QEMU |