第1页 / 共7页
第2页 / 共7页
第3页 / 共7页
第4页 / 共7页
第5页 / 共7页
第6页 / 共7页
第7页 / 共7页
云计算中虚拟机迁移技术的全面综述-研究论文.pdf
A Comprehensive Review of Virtual Machine Migration Techniques in
Cloud Computing
Suruchi Talwani, Jimmy Singla
School of Computer Sci. & Engg., Lovely Professional University, Phagwara, 144411, India
Email: suruchitalwani14@gamil.com
Abstract As cloud users grow exponentially, virtualization of the cloud
data center (CDC) is accomplished by running simultaneous virtual
machines (VM) on one cloud server in parallel processing. The growing
demand for resources such as communication, storage, and computing
can be successfully met by using dynamic virtual machine migration.
For single or multiple VM migrations, most resource management
techniques are used to maximize the performance and efficiency of the
cloud data center. The Qos (service quality) declines proportionally with
fast increase in cloud users and associated workload. The powerful live
VM migration strategy can only be used to deliver better computing
facilities, manage large numbers of cloud users and reduce time and
energy on cloud data center. Modern cloud computing system depends
primarily on live VM migration. This paper analyzes with their
advantages and disadvantages all the
live VM migration
techniques. Their future range research and implementation has been
briefly mentioned. Comparisons have been made between pre-copy and
post-copy VM Migration techniques through simulation corresponding
to
as
parameters.This paper reviews various virtual machine migration schem
es while, through a comprehensive analysis of existing schemes, examinin
g the critical aspects of virtual machine migration schemes.Existing
systems are explored with key goals. In the VM migration domain, some
open research issues are also mentioned.
memory
network
latest
CPU
uses,
and
Keywords—cloud computing, cloud data center, virtualization, virtual
machine migration
1.
INTRODUCTION
leads
Virtual machine migration (VMs) is the key virtualization feature where
virtual machines migrate between hosts. To accomplish the mission effectively
and successfully, challenges like load balancing, reactive or proactive fault
tolerance, and upgrading or system maintenance are addressed. Migration is
accomplished through two patterns between servers that are alive or non-live.
In case of live migration technique, states of memory, CPU statements, and
VM storage data are moved from one host to another. Optimized live
migration takes place, such as total migration time is calculated, service
downtime is noticed, and data transfer (Google, "Google App Engine", 2012)
Migration also
to resource management goals such as server
consolidation, network load balancing, performance improvement, energy
efficiency making the CDC more efficient and efficient to meet user needs.
Migration and consolidation strategies for virtual machines (VM) also aid in
green computing and help in maintaining the balance of workload to avoid
excessive use of servers. In a non-live scenario, the application services get
stopped, therefore prediction of migration and downtime gets easy. In a non-
live migration strategy-pause, copy, and restart, there are three major steps.
The high demand for internet usage and worldwide digitalization has
strengthened the idea of large-scale data centers being introduced. Cloud data
center's main objective is to provide the best Internet access . According to the
survey, the energy consumed by GOOGLE datacenter in 2013 was close to
260 million megawatts and 0.01 percent of global energy. Such an immense
amount of energy is more than enough to supply 200,000 homes. Cloud data
centers now have sufficient processing power due to VM migration, and are
also able to reduce workload. By using live VM migration tool, the cloud
computing service provider enables Platform as a service (PaaS), Software as a
service (SaaS), and Infrastructure as a service (IaaS) (Amazon, "Amazon
Elastic Compute Cloud (Amazon EC2)",2012). Live VM migration is therefore
the main focus for both researcher and expert in industry. Our aim is to define
and identify available methods of live VM migration and to examine the pros
and cons of existing research on it.
Data centers are providers of internet services. We provide all the
of
the
to
the
barriers
research
modern computing services on the internet and charge the user according to
their use. To order to get the facilities, the user just requires an internet
connection and a pay scheme. In addition, a cloud data center provides
services through a web link. But due to internal fragmentation, cost
increases and performance reduces to provide the service. The remainder of
the paper was split into chapters. Various sections focus on the live
migration strategies of the literature. Section 2 presents various techniques
to be studied by different researchers. Section 3 discusses some drawbacks
and
literature.
Migration problems at the application level can be avoided through the migr
ation of a virtual machine. Residual dependencies are removed by VMM. V
irtual Machine Migration enables power savings, balancing of loads, efficie
nt use of resources. There are two types of virtual machine migration
methods: warm (live) migration and cold (non-live) migration. The VM
loses status and client may note the cold migration service interruption.
When transitioning, virtual machine continues to run and does not lose its
status. Through warm (live) migration, the client does not notice any
interruption through service. A virtual machine’s status for migration is
seen
migration.
The state is made up of its contents of memory and local file system.It is no
t necessary to transfer the local file system (Microsoft,
"Windows
Azure.",2012). Next, VM is suspended, thenstate is moved, and eve
of
ntually, resuming
done
at the host destination. Online maintenance
along with
load balancing and energy management:
VM
facilitated
process
process
live
the
in
of
is
is
layer
Sapuntzakis and so on. Al. (P. S. Constantine, C. Ramesh, P. Ben, C. Jim, S.
L. Monica, and R. Mendel ,2009), shows the process of the state of the
operating machine that can be quickly moved across the network around,
including the disk state, registers, memory operations and devices for input
and output devices. Capsule ,a hardware layer(a very efficient and
operational
that contains the operating system as a whole)
that contains the operating system as a whole as well as running programs a
nd processes.Live migration algorithm's basic idea, first suggested by Clark
et. Allaah (C. Christopher, F. Keir, H. Steven, H. Jacob Gorm, J. Eric, L.
Christian,P. Ian, and W. Andrew. 2005).At some point, the VM is suspended
at the source to stop further writing of memory and transfer the remaining
pages. The VM resumes at destination Nelson et. al.( N. Michael, L. Beng-
Hong, and H. Greg,2005) after transferring all the memory content. defines
and ensure fast and transparent migration of software, and does not involve
modification of applications or operating systems. For hundred virtual
machines, performance
that application
downtime was less than a second for a variety of workloads.
is measured. Results were
Kuno, et. Al. (Y. Kuno, K. Nii, and S. Yamaguchi,2011) shows performance
assessment of live as well as non-live migration methods and shown a
severe decline in process performance on a virtual migratory machine. A
host OS interaction and memory writing are the main reasons for the
decline. We also look at the reasons for the decline in I / O results. These
results show that one of the main reasons for the slowdown in performance
is migration transmission.
Feng et. Al. (X. Feng, J. Tang, X. Luo, and Y. Jin,2011) contrasts VMotion
and XenMotion output. When migrating VM example, VMotion performs
better than XenMotion when generating full live migration information.
VMotion and XenMotion's efficiency in the network is deteriorating with
delay and packet loss. With loss of packet and mild delay, VMotion proved
worse than XenMotion in the network. Live migration technolo in life. The
following metrics measure the live migration’s performance: Preparation
Time, Downtime, Resume time, Pages Transferred, Total Migration Time,
Application Degradation. The current research study
is aimed at
undertaking a comprehensive survey of live VM migration literature and
1
Electronic copy available at: https://ssrn.com/abstract=3564971
evaluating the various mechanisms proposed to achieve different resource
management goals. Specific shortcomings are identified and the study
issues are highlighted in order to improve the reliability of live VM
migration. A comprehensive study on predictive load balancing approaches
will be carried out in the future and a new strategy will be implemented for
the same purpose. A host OS interaction and memory writing are the main
reasons for the decline. We also look at the reasons for the decline in I / O
results. These results show that one of the main reasons for the slowdown in
performance is migration transmission.
2. CLOUD LIVE VM MIGRATION TECHNIQUES
An extremely strong and powerful cluster tool is live migration and machine-
wide instances can be rearranged to ease the overloaded hosts. For performing
migration, its runtime state is moved to the destination from the source with
VM in running mode.
The main two primary approaches: transfer of memory after copy and pre-
copy[22]. First, suspension of migrating VM by the post copy takes place at
the source, then it copies the minimum processor status to the destination
node, then starts the virtual machine, and starts to collect memory blocks or
pages from the source over the network.
Pre-copy method has two phases: the stage of warm-up and the phase of stop-
and-copy. If there is a change in some memory pages during the process of
memory copying, they are re-copied till the rate is more than the rate of page
dirtying. The VM will be stopped at origin in the Stop and Copy process and
the dirty left over pages will be saved at the destination and then VM gets
resumed.
2.1 Post Copy Approaches:
Hines et. Al. (R. H. Michael, D. Umesh, and G. Kartik, 2009). describes the
design part, implementation part and methodology of a technique called post-
copy for the live virtual machine migration. Postcopy consists of four
components that are major: demand paging, aggressive pressing, planning, and
self-ballooning dynamics. They introduced and tested post-copy on a platform
based on Xen platform and Linux platform. It is shown by the results that there
is a reduction in the total time of migration and the transfer rate of number of
pages in comparison to pre-copy significantly. The pre-preparation bubbling
algorithm will substantially reduce the number of network errors.
Michael and so on. Al. (R. H. Michael and G. Kartik,2009) Contrast the post-
copy solution of Xen Hypervisor with the pre-copy methodology. It shows
changes in several migration metrics using a variety of VM workloads,
including migrated pages, overall migration time and overhead network. For
eliminating duplicate page transmissions, adaptive pre-paging is used with
post-copy. We remove the transition of free memory pages by a dynamic
algorithm of self-ballooning (DSB) method in both migration schemes. Free
pages are released time to time to the hypervisor in a guest machine and
dramatically accelerates migration with negligible degradation of efficiency.
2.2 Pre Copy Approaches:
Precopy approach has many categories
of many
technologies, improvement in existing approaches of precopy , migration of
multipleVMs, and consideration is given to specific application loads. The foll
owing methods were explained:
like:- Combination
2.3 Combined migration Technologies:
L.
Weining
and
(Liu and Al.
F.
Tao ,2009)defines a strategy that is novel. We combine device recovery softwa
re using
trace
or
replay) alongwith CPU scheduling to provide fast and clear migration. Do
wntime is less for this technology and the overall time is fair.
check pointing or
recovery or
techniques
like
Liu et. Al.( L. Haikun, J. Hai, L. Xiaofei, H. Liting, and Y. Chen
,2009)describes CR / TR-Motion's novel approach, which adopts recovery
checks and trace or may be replay software to ensure quick VM migration that
will be transparent also. This scheme will significantly reduce downtime on
conversion. In multi-processor (or multi-core) system, because there is a need
to record and replay costly memory race between different VCPUs, this makes
migrating SMP guest OS an inherent challenge to this approach. Once CPU
and/or memory heavy VMs get migrated,migration downtime has been
increased which can cause interruption of service or failure occurs and total
time of migration has been prolonged which is detrimental to overall system
performance.
Kumar Bose and so on. Al. (S. Kumar Bose, S. Brock, R. Skeoch, N.
Shaikh, and S. Rao,2011) suggests combining replication of VM with
scheduling of VM to tackle the problem of migration latencies They
selectively duplicate a VM image through multiple cloud locations, make
a duplicate of any VM image as the primary copy. Then propagation of
changes that are incrementally done to the remaining VM replicas is done
at the primary copy. This architecture called CloudSpiders for automated
replication as well as scheduling will reduce migration latencies that deal
with live scenario of migration of VM images through WANs.
2.4 Improvised Pre-copy Approaches:
Jin et. Al. (J. Hai, D. Li, W. Song, S. Xuanhua, and P. Xiaodong ,2009)
present an architecture of a new VM migration method (MECOM) based on
memory-compression. Firstly, memory compression is used for providing
fast and reliable virtual machine migration, while services of virtual
machine gets little bit affected by the characteristics of the memory site. We
also developed an adaptive zero-aware compression algorithm to balance
virtual machine migration efficiency and cost. Pages are compressed
quickly in source batches and recovered precisely on target.
Fei Ma and so on. Al.( M. Fei, L. Feng, and L. Zhen,2010) improvedXen3.3.0
pre-copy approach by introducing a bitmap section that labels those sites
that are updated very regularly. In the process of iterations, these pages are
stored in the bitmap tab, and only in the last iteration cycle can those pages
be transmitted. It allows the transmission of frequently updated pages for
one time.
Svard et al. (S. Petter, H. Benoit, T. Johan, and E. Eri,2011) implemented the
algorithm named as the delta compression that acted as a kind of a
modification to the KVM hypervisor. The quality is measured by the
migration of running VMs with various types of workload, which shows a
noticeable reduction in downtime for migration. It is demonstrated that there
is a high risk of loss of service when VMs migrate to low-bandwidth
networks with heavy workloads.
Ibrahim et al. ( K. Z. Ibrahim, S. Hofmeyr, C. Iancu, and E. Roman ,2011)
present the latest KVM implementation for memory-intensive applications
,performance analysis and research the pattern of live iterative pre-copy
migration. The increase in memory level of the scientific software (VM
includes multiple cores) comes higher as compared with draining migration
rate. They present a new algorithm for getting low downtime as well as low
impact on overall performance. KVM implements this algorithm.
2.5 Multiple VMs migration
Al-Kiswany et. Al. (S. Al-Kiswany, D. Subhraveti, P. Sarkar, and M.
Ripeanu,2011) introduces VMFlockMS, a platform designed for transfer a
cross-datacenter and team deployment of images provided with a solution at
application-level
(web application with three levels). VMFlockMS
implements two techniques: 1) transfer of duplicate data within the
VMFlock and between the VMs in the VMFlock and information that is
present in the datacenter of destination. 2) Accelerated software instance at
the target datacenter after only moving a partial array of data blocks and
prioritizing the remaining information based on previously established
access patterns from running VMs. It is possible to achieve a scalable and
high-performance migration network.
Ye et. Al. (Y. Kejiang, J. Xiaohong, H. Dawei, C. Jianhai, and W.
Bei,2011) said that Live Migration Reserve System is made of Decision
Makers nodes, Migration Controllers, Resource Reserve Controllers and
Capacity Monitor. The allocated source machine resource includes Xen and
storage resource that is dynamically changing VM memory size while it
contains all resources in the target machine. Three metrics are downtime,
total time, overhead workload quality to measure the efficiency.
Deshpande and so on. Al.( S. Kikuchi and Y. Matsumoto,2011) (D. Umesh,
K. Unmesh, and G. Kartik,2011) outlines a de-duplication-based approach
for co-located VMs simultaneously. This approach transmits only once
identical memory content across VMs during migration to significantly
reduce both total migration time and network traffic. They used QEMU /
KVM's Linux system to migrate virtual machines to live gangs.
2.6 Application / Workload Specific Technologies:
When used on big computer systems such as SAP or ERP, conversion
technology is constrained. Such computers use a considerable units of
memory space. Hudzia made a system to support the open, live migration
2
Electronic copy available at: https://ssrn.com/abstract=3564971
of virtual machines that run large-scale workloads of business application,
deployment and evaluation. This minimizes service disruption due to the
use of VM delta compression algorithm for memory transfer, as well as
the implementation of an efficient warm-up step to reduce the stiffness of
massive VM migration.
For data intensive applications, Sato and so on. Al. proposed a relocation
algorithm in a geographically distributed environment on a virtual
machine. The proposed algorithm determines VM's optimal location for
accessing target file and minimizing the access time of expected file with
the help of DAG's shortest path search problems. Higher performance can
be achieved as compared to simple techniques.
Piao, and. Al.provides a networks
of
VM and migration strategy in cloud computing environments for data int
ensive applications. The suggested solution positions the VMs on physical
machines taking into account the constraints of the network.
conscious placement
AppAware comes under greedy heuristics algorithm for assigning one VM
to suitable physical machines, while at the same time trying to minimize
the
of
mapping.Using one of the approaches VMs are allocated to physical devic
in demonstrating
es. Simulations helped
the data about network
statistics that
network
traffic of
is reduced by 81 percent compared to a well
known alternative method of VM migration which is not application-
aware.
cost
2.7 Other Important Technologies:
Nocentino, etc. Al.introduces a dependency aware approach that helps in
reducing latency and overhead occurs during migration and results of
investigation were also shown. For detecting intrusion, some mechanism is
used here.. Dependence information is used during live migration to discern
processes.
Akoush and so on. Al. shows that link speed and how much a page is dirty
are the two factors that affect the migration process. A non linear effect is
produced by these factors on the performance of migration in large part to the
conditions of hard stoppage that force migration to reach the phase of final
stop-and-copy. No of times a migration will take place must be predicted
correctly so that VMs can be put more efficiently and smartly without
degrading efficiency.
Wood, and so on. CloudNet, made of cloud computing systems connected to
a VPN-based network infrastructure, offers smooth access between business
and cloud sites. CloudNet offers tailored support to live WAN migration,
which benefits from Internet connections, reducing the overall cost of
migrations.
Huang and so on. Al. present a live benchmark implementation, Virt-LM, to
compare performances in a data center scenario between different software
and hardware environments. The goals for designing Virt-LM are metrics
data, workload data, impartial scoring methodology, metrics for predicting
stability, compatibility issues, and usability factors. The availability of
resources will predict correct decisions about when should VM be migrated
and how the resources are provisioned between VMs.
In order to create a performance model, Wu and Zhao use statistical methods
like regression that is very useful in predicting migration duration and to take
resource provisioning and management decisions. Experiment was done by
migrating an intensive computational application running a xen-based VM
and allocating a variety of CPU shares. This indicates that the live migration
tools available have an effect on the migration period.
Jing and so forth. Al. proposes a migration model for optimization for
minimizing application downtime, that considers the memory transfer
statistical data of Xen VM. This architecture uses layered copy and memory
compression algorithms that optimize complexity in terms of time and space
and hence significantly reduces downtime for migration, that ultimately leads
to overall efficiency of migration.
Ashino, etc. Al. suggests VM migration for troubleshooting .It means that
while loading the device drivers or when we change configuration of any
device, Guest OS does not boot to destination. EDAMP's method of migration
is being proposed and is still being developed. There is just overwriting of the
files but device driver does not get deleted. EDAMP has a great usage in
many cloud services and can be combined into a single hypervisor.
3. RESEARCH CHALLENGES IN LIVE VM MIGRATION
3.1 Low Bandwidth in WAN environment
Depending on the storage cost of storage and demand, a virtual computer
can be designed to operate at geographically disparate locations. As VM
image has a large size, it is almost impossible in' reasonable' time to
translocate a live VM across high-latency low-bandwidth wide area
networks (WAN).
3.2 Virtual Machine having heterogeneous workload
Larger application systems like SAP or ERP face difficulty in the area of
live migration of VMs because these applications consume high amount of
resources like memory or other input output devices.
3.3 Link speed and page dirty rate
The main factors influencing the behavior of migration are connection
speed and site dirty frequency. Link efficiency has an inverse
proportionate relation with time taken sand downtime for migration. The
rate at which memory pages used by VM are changed directly affects the
number of pages in each pre-copy iteration that are transferred. This rate is
called page dirty rate.
4. LITERATURE REVIEW
Different researchers conducted the study on the entire process of virtual
machine migration and proposed different schemes for automating live mi
gration. This section offers some of the proposed plays a brief review.
4.1 Live VMM techniques
Chang and Gu Al. developed a new resource-conscious VM choice
scheme to boost database usage in the migration process. The goal was to
reduce low-variable servers by calculating the cost of a square root
minimized metric MISR measured using two variable-standardized and
optimal resource usage. Both idle servers are designed to work on power-
saving mode. If no active virtual machines are minimized, cloud data
center energy consumption can be significantly reduced using live VM
migration technique. However, there are several breaches in accordance
with Service Level Agreements (SLA) and fines may also be levied on the
provider. For instance, researchers have proposed two step algorithms in
to preserve the trade between energy usage and total migrations. For this
reason, a novel host selection algorithm has been proposed, taking into
account the network's potential load. The proposed solutions are based on
a technique of exponential smoothing which takes into account weighted
moving averages. CloudSim was used for simulation and the results are
contrasted with existing techniques. Sharma and Chawla have suggested a
live migration algorithm in which the live migration precopy process was
divided into three sections. The number of pages transferred will be
reduced in the first phase. The second phase, often and often used,
categorizes the pages into two categories. Once pages are marked, the
movement of redundant pages is minimized. In, the authors focused on the
mechanism of multiple VM migration and suggested an enhanced strategy
for serial and mixed migration. The mixed approach is a combination of
competitive serial and parallel migration approaches. The researchers used
queuing theory to develop the scheme taking into account only the use of
CPU.XEN and KVM compared the performance of sequential, parallel
and mixed migration approaches, taking into account metrics such as
waiting time and queue size. It also reduced the cost of online VM
migration to offline switching. CloudSim studies have been conducted that
have reduced load and power consumption.
4.2 VMM using Prediction methods
Dambreville et.al. suggested a POD (Predict Optimize Dispatch)
algorithm to incorporate a predictive scheduling approach for reducing
energy consumption. Using this prediction model, this goal was achieved
by anticipating computation demands. The array of servers available will
be changed to fit the expected values and scheduling will finally take
place. In 2017, Nie.et.al. explained and implemented an automated live
migration algorithm called OPCA. In the memory chapter, Grey
Markov's method was used to figure out the possibility of change. The
work area was divided into two categories-hot work area and cold work
area based on this approximate potentials .High change probability pages
are kept warm and low change probability pages are kept in cold
workspace. The pages contained in hot space were copied in the final
round of migration to prevent multiple retransmissions. The experiments
show that OPCA reduces the number of iterations and downtime, raising
the level of use of the system and enhancing user experience. Researchers
3
Electronic copy available at: https://ssrn.com/abstract=3564971
proposed a strategy for detecting overloaded and underloaded hosts.
Gaussian processes are used by the technique to compare current
techniques and future work applications and their threshold calculated
values to long-term forecasting and probability model. Johnson also
calculated the likelihood of shifting memory pages using a time series
prediction model to pass high frequency pages to their final destination.
An approach to VM allocation that uses Best Fit Decreasing (BFD)
methodology and combo of an ensemble forecasting algorithm and
adaptive learning concepts has recently been proposed so that there is a
reduction in energy consumption of data centers of cloud . Automata can
also help in controlling the complex behavior or pattern of networking
nodes.
4.3
Bio Inspired VMM techniques
Tsygankov and Chen both investigated the issue of network aware VM load
balancing and proposed a theory about bandwidth solution and migration
time strategy so that migration time can be reduced and load gets properly
balanced and maintained. The goal of optimizing migration time and load
imbalance was achieved using Ant Colony algorithm that comes under the
category of bio-inspired optimization technique. Xu et.al. proposed an
Artificial Bee Colony (ABC) selection strategy in which binary search was
used with the basic algorithm with the aim of achieving balanced and optimal
homogeneous initialization, i.e. there should be minimum overhead. Bayes
and Boltzmann's principle was also incorporated in the PS-ABS approach
helped in reducing the VM failures that came during the migration process,
PS-ABC contributes in achieving lower power and energy consumption and
making the system resilient. Active hosts gets lessen in number and migration
number also lowered down. Euclidean distance was used to categorize the
different nodes in the category of under, partial or fully loaded nodes that
depend on how much disk or CPU is being used. VM reallocation process
also used the PSO method so as to ensure service quality by reducing energy
consumption. Effective use of resources was also accomplished by avoiding
the problem of local maxima. Proposed methodology produced optimized
results in the key areas of energy efficiency as compared with heuristic
techniques.
5.
RESEARCH GAPS, FINDINGS AND CHALLENGES
After doing an exhaustive study of the techniques of VM migration, it is
concluded that there are lots of challenges. By addressing these hurdles, the
migration process can be streamlined, that will lead in development of more
effective cloud environment. Following are the listed gaps that are founded
during the research process-
Due to the increasing use of cloud services, data centers have gained a huge
importance. Because of the accumulated corporate competition to develop
cloud data centers, energy usage is also rised in these data centers. Reducing
this consumption of energy is one of the key challenge that needs to be
clarified.
Despite of the fact that many researchers have given graph solutions for
challenges like load monitoring and load balancing, but still there are two
factors that are pivotal in migration process like how to make green cloud and
how to make fault tolerant cloud.
Request for resources can take place simultaneously in a cloud environment
and further these requests find the suitable and available resources. Over the
past decade, there is one key challenge that have gained due interest is
resource utilization.
Different load prediction methods have been described for load balancing duri
ng VM migrations using the
or
machine learning concepts but their combination is still a major problem in
the area of cloud server consolidation.
of neural network or automata
concept
Combined approach made by heuristic as well as meta heuristic approach for
optimizing the performance of network is also a challenge.
from static
Current VM load balancing techniques
nature to dynamic nature because of
efficient and optimum
performance of dynamic techniques. So there is a need of investigating the
self adaptive dynamic VM balance algorithms for balancing load.
Correlation of various virtual machines while migrating especially in
wireless medium
like latency,
the
delay
are
the
shifting
for
avoiding
problems
and network degradation is also one of the challenge that needs to be taken
care of.
6. CONCLUSION
The study that is done in this paper aims at undertaking a comprehensive
survey of live VM migration literature and evaluating the different
mechanisms that have been for achieving goals related to resource
management. Many important shortcomings and key issues are mentioned
so that reliability can be enhanced. A comprehensive study on predictive
load balancing approaches will be done in the future and a novel and
efficient strategy will be implemented for achieving better migration.
REFERENCES
Google, "Google App Engine", (2012), [online]. Available:
cloud.google.com [Nov 1, 2012].
Amazon, "Amazon Elastic Compute Cloud (Amazon EC2)",
(2012), [online]. Available: aws.amazon.com/ec2/ [Nov 1,
2012].
Microsoft,
Available: windowsazure.com [Nov 1, 2012].
"Windows Azure.",
(2012),
[online].
"Virtual Machine."
(2012)[online].Available:http://searchservervirtualization.techtarget.com/
IBM, "SmartCloud." (2012), [online]. Available: ibm.com/cloud-
computing [Nov 1, 2012].
P. Mell and T. Grance, "The NIST definition of cloud computing (draft),"
NIST special publication, vol. 800, p. 145.
A. Desai,
definition/virtualmachine
VMWare, "vSphere ESX and ESXi Info Center.", (2012),[online].
Available: vmware.com/products/vsphere/esxi-and-esx [Nov 1, 2012].
Microsoft, "Windows Virtual PC.", (2012), [online]. Available:
http://www.microsoft.com/windows/virtual-pc/ [Nov 1, 2012].
Xen,"Xen Hypervisor.",
http://www.xen.org/products/xenhyp.html
Microsoft, "Hyper-V Server 2012.", (2012),
microsoft.com/server-cloud/hyper-v-server/ [Nov 1, 2012].
KVM, "Kernel-based Virtual Machine.", (2012), [online].Available: linux-
kvm.org [Nov 1, 2012].
[online].Available:
[Nov 1, 2012].
[online].Available:
(2012),
Oracle, "VirtualBox.", (2012), [online]. Available: virtualbox.org [Nov 1,
2012].
W. Voorsluys, J. Broberg, S. Venugopal, and R. Buyya, "Cost of Virtual
Machine Live Migration in Clouds: A Performance Evaluation," in 1st
International Conference on Cloud Computing, Berlin, Germany, 2009, pp.
254-65.
P. S. Constantine, C. Ramesh, P. Ben, C. Jim, S. L. Monica, and R. Mendel,
"Optimizing the migration of virtual computers," in 5th Symposium on
Operating Systems Design and Implementation, SIGOPS Oper. Syst. Rev.,
vol. 36, Issue SI, pp. 377-390, 2002.
C. Christopher, F. Keir, H. Steven, H. Jacob Gorm, J. Eric, L. Christian,P.
Ian, and W. Andrew, "Live migration of virtual machines," 2nd conference
on Symposium on Networked Systems Design & Implementation - Volume
2: USENIX Association, 2005.
N. Michael, L. Beng-Hong, and H. Greg, "Fast transparent migration for
virtual machines," Annual conference on USENIX Annual Technical
Conference Anaheim, CA: USENIX Association, 2005.
H. Wei, G. Qi, L. Jiuxing, and D. K. Panda, "High performance virtual
machine migration with RDMA over modern interconnects," in IEEE
International Conference on Cluster Computing, 2007, pp. 11-20.
4
Electronic copy available at: https://ssrn.com/abstract=3564971
L. Yingwei, Z. Binbin, W. Xiaolin, W. Zhenlin, S. Yifeng, and C. Haogang,
"Live and incremental whole-system migration of virtual machines using block-
bitmap," in IEEE International Conference on Cluster Computing, 2008, pp. 99-
106.
B. Robert, K. Evangelos, F. Anja, S. Harald, and berg, "Live wide-area
migration of virtual machines including local persistent state," 3rd International
Conference on Virtual execution environment, San Diego, California, USA:
ACM, 2007.
Y. Kuno, K. Nii, and S. Yamaguchi, "A study on performance of processes in
migrating virtual machines,", 10th International Symposium on Autonomous
Decentralized Systems, ISADS 2011, 2011, pp. 567-572.
X. Feng, J. Tang, X. Luo, and Y. Jin, "A performance study of live VM
migration technologies: VMotion vs XenMotion," The International Society for
Optical Engineering, 2011.
R. H. Michael, D. Umesh, and G. Kartik, "Post-copy live migration of virtual
machines," SIGOPS Oper. Syst. Rev., vol. 43, pp. 14-26, 2009.
R. H. Michael and G. Kartik, "Post-copy based live virtual machine migration
using adaptive pre-paging and dynamic self-ballooning,", ACM
SIGPLAN/SIGOPS international conference on Virtual execution environments,
Washington, DC, USA: ACM, 2009.
L. Weining and F. Tao, "Live migration of virtual machine based on recovering
system and CPU scheduling," in 6th IEEE joint International Information
Technology and Artificial Intelligence Conference, Piscataway, NJ, USA, May
2009, pp. 303-7.
L. Haikun, J. Hai, L. Xiaofei, H. Liting, and Y. Chen, "Live migration of virtual
machine based on full system trace and replay," 18th ACM International
Symposium on High performance distributed computing Garching, Germany:
ACM, 2009.
P. Svard, J. Tordsson, B. Hudzia, and E. Elmroth, "High performance live
migration through dynamic page transfer reordering and compression," 2011 3rd
IEEE International Conference on Cloud Computing Technology and Science,
CloudCom 2011, pp. 542-548.
S. K. Bose, S. Brock, R. Skeoch, and S. Rao, "CloudSpider: Combining
replication with scheduling for optimizing live migration of virtual machines
across wide area networks," 11th IEEE/ACM International Symposium on
Cluster, Cloud and Grid Computing, CCGrid 2011, May 2011, pp. 13-22.
S. Kumar Bose, S. Brock, R. Skeoch, N. Shaikh, and S. Rao, "Optimizing live
migration of virtual machines across wide area networks using integrated
replication and scheduling," in 2011 IEEE International Systems Conference,
SysCon 2011 - pp. 97-102.
J. Hai, D. Li, W. Song, S. Xuanhua, and P. Xiaodong, "Live virtual machine
in IEEE International
migration with adaptive, memory compression,"
Conference on Cluster Computing and Workshops, CLUSTER '09, pp. 1-10.
M. Fei, L. Feng, and L. Zhen, "Live virtual machine migration based on
improved pre-copy approach," in IEEE International Conference on Software
Engineering & Service Sciences ICSESS), 2010, pp. 230-233.
S. Petter, H. Benoit, T. Johan, and E. Erik, "Evaluation of delta compression
techniques for efficient live migration of large virtual machines," 7th ACM
SIGPLAN/SIGOPS
on Virtual Execution
Environments, California, USA: ACM, 2011.
K. Z. Ibrahim, S. Hofmeyr, C. Iancu, and E. Roman, "Optimized pre- copy live
migration for memory intensive applications," in International Conference for
High Performance Computing, Networking, Storage and Analysis (SC),2011,
pp. 1-11.
International Conference
S. Al-Kiswany, D. Subhraveti, P. Sarkar, and M. Ripeanu, "VMFlock: Virtual
machine co-migration for the cloud," IEEE International Sym. on High
Performance Distributed Computing, 2011, pp. 159-170.
Y. Kejiang, J. Xiaohong, H. Dawei, C. Jianhai, and W. Bei, "Live Migration
of Multiple Virtual Machines with Resource Reservation in Cloud Computing
Environments," California, USA, 2011, pp. 267-74.
S. Kikuchi and Y. Matsumoto, "Performance modeling of concurrent live
International
workshop
migration operations in cloud computing systems using prism probabilistic
model checker," 2011 IEEE 4th International Conference on Cloud
Computing, CLOUD 2011, July 2011, pp. 49-56.
D. Umesh, K. Unmesh, and G. Kartik, "Inter-rack live migration of multiple
virtual machines,", 6th
on Virtualization
Technologies in Dist. Computing, Delft, Netherlands.
E. Elmroth and L. Larsson, "Interfaces for placement, migration, and
monitoring of virtual machines in federated clouds," in 8th International Conf.
on Grid and Cooperative Computing, GCC 2009, pp. 253-260.
A. Celesti, F. Tusa, M. Villari, and A. Puliafito, "Improving virtual machine
migration in federated cloud environments," 2nd International Conference on
Evolving Internet, Internet 2010, pp. 61-67.
S. Chun-Hui, M. Kirchberg, and L. Bu Sung, "Efficient Migration of Virtual
Machines between Public and Private Cloud," in IEEE Third International
conference on Cloud Computing Technology and Science (CloudCom), Los
Alamitos, CA, USA, Nov 2011, pp. 549-53.
H. Stuart, Beno, and H. t, "Improving the live migration process of large
enterprise applications," 3rd International Workshop on Virtualization
Technologies in Distributed Computing, Barcelona, Spain: ACM, 2009.
K. Sato, H. Sato, and S. Matsuoka, "A model-based algorithm for optimizing
I/O intensive applications in clouds using vm-based migration," in 2009 9th
IEEE/ACM International Symposium on Cluster Computing and the Grid,
CCGRID 2009, pp. 466-471.
J. T. Piao and J. Yan, "A network-aware virtual machine placement and
migration approach in cloud computing," 9th International Conference on
Grid and Cloud Computing, GCC 2010, pp. 87-92.
V. Shrivastava, P. Zerfos, L. Kang-won, H. Jamjoom, L. Yew-Huey, and
S. Banerjee, "Application-aware virtual machine migration in data centers," in
IEEE INFOCOM, 2011, pp. 66-70.
L. Haikun, X. Cheng-Zhong, J. Hai, G. Jiayu, and L.Xiaofei, "Performance
and energy modeling for live migration of virtual machines," 20th
International Symposium on High Performance Distributed Computing, San
Jose, California, USA: ACM, 2011.
N. Anthony and M. R. Paul, "Toward dependency-aware live virtual machine
migration," 3rd International Workshop on Virtualization Technologies in
Distributed Computing, Barcelona, Spain: ACM, 2009.
A. Sherif, S. Ripduman, R. Andrew, W. M. Andrew, and H. Andy, "Predicting
the Performance of Virtual Machine Migration," in IEEE International
Symposium on Modeling, Analysis and Simulation of Computer and
Telecommunication Systems, 2010.
Y. Kanada and T. Tarui, "A "network-paging" based method for wide- area
live-migration of VMs," in International Conference on Information
Networking 2011, ICOIN 2011, Jan 2011, pp. 268-272.
T. Wood, P. Shenoy, K. K. Ramakrishnan, and J. Van Der Merwe, "CloudNet:
Dynamic pooling of cloud resources by live WAN migration of virtual
machines," 2011 ACM SIGPLAN/SIGOPS International Conference on
Virtual Execution Environments, VEE 2011, pp. 121-132.
H. Dawei, Y. Deshi, H. Qinming, C. Jianhai, and Y. Kejiang, "Virt-LM: a
benchmark for live migration of virtual machine," Second joint WOSP/SIPEW
International Conference on Performance engineering Karlsruhe, Germany:
ACM, 2011.
Y. Wu and M. Zhao, "Performance modeling of virtual machine live
migration,", IEEE 4th International Conference on Cloud Computing,
CLOUD 2011, pp. 492-499.
J. Yang, "Key technologies and optimization for dynamic migration of virtual
machines in cloud computing," Int. Conf. on Intelligent Systems Design and
Engineering Applications, ISDEA 2012, pp. 643-647.
Y. Ashino and M. Nakae, "Virtual machine migration method between
its evaluation," 26th IEEE
different hypervisor
International Conference on Advanced
Information Networking and
Applications Workshops, WAINA 2012, pp. 1089-1094.
implementations and
5
Electronic copy available at: https://ssrn.com/abstract=3564971
.
6
Electronic copy available at: https://ssrn.com/abstract=3564971
7
Electronic copy available at: https://ssrn.com/abstract=3564971
相关推荐
2023年江西萍乡中考道德与法治真题及答案.doc
2012年重庆南川中考生物真题及答案.doc
2013年江西师范大学地理学综合及文艺理论基础考研真题.doc
2020年四川甘孜小升初语文真题及答案I卷.doc
2020年注册岩土工程师专业基础考试真题及答案.doc
2023-2024学年福建省厦门市九年级上学期数学月考试题及答案.doc
2021-2022学年辽宁省沈阳市大东区九年级上学期语文期末试题及答案.doc
2022-2023学年北京东城区初三第一学期物理期末试卷及答案.doc
2018上半年江西教师资格初中地理学科知识与教学能力真题及答案.doc
2012年河北国家公务员申论考试真题及答案-省级.doc
2020-2021学年江苏省扬州市江都区邵樊片九年级上学期数学第一次质量检测试题及答案.doc
2022下半年黑龙江教师资格证中学综合素质真题及答案.doc
