You are on page 1of 3

Analysis and Implementation of Virtualization in Cloud Computing

Ahimsa Bhardwaj Preeti Sirohi


IMS Ghaziabad IMS Ghaziabad
Ahimsa.bharadwaj@imsgzb.com preeti.sirohi@imsgzb.com

Abstract— Cloud technology comprised of using a networked Terremark) In cloud computing the cloud provider offers
infrastructure, platform and software with the capacity for providing sharing of resources like infrastructure and software to its
an on-demand resources to the customer. Virtualization is an consumers. Resource allocation method is dynamic and
important technology for cloud environment as it provide users the flexible which fulfill the demands and release of services to
ability to allocate and de-allocate the resources, computation power,
storage space as per the user requirement. The cloud resources can be
clients at any point of time. [1].
taken by the user from any cloud service provider on pay as you go
cost model.

Ubiquitous network access: The service in cloud environment


Keywords— Virtualization; Cloud Computing is accessible through internet. The devices which have internet
connection can easily access the services of cloud. Also there
are number of data centers across the globe facilitating the
cloud service provider to take advantage by achieving
I. INTRODUCTION
maximum service utility.
“Cloud computing is the use of networked infrastructure
Service oriented: Cloud computing has service oriented
software and capacity to provide resources to users in an on-
architecture where Service Level Agreement plays an
demand environment. Sometimes known as utility computing,
important role. IaaS, PaaS and SaaS provide its services
clouds provide a set of typically virtualized computers which
according to SLA agreed between the provider and the
can provide users with the ability to start and stop servers or
consumers .
use compute cycles only when needed, often paying only upon
usage.”[1]
Scalability: Cloud consumer can store their data to cloud
without worrying about the space and backing up of data.
Cloud Computing Characteristics:
Resources can be allocated and de allocated according to
demand. The dynamic resource management features of cloud
Cloud computing provides various features which make it technology makes it unique from traditional computing
different from traditional service computing, the features are models. Users can request the services and resources
summarized below: whenever they are required .

On-demand self service. Cloud provides users services as per Utility-based pricing: Cloud computing is based on pay per-
the need of clients. The payment of the services is done by use pricing model. User only need to pay for the services
using “pay-and-go” method. Just like we pay our electric bills used .The pricing scheme of services depends upon the type of
according to usage here we need to pay only for the services services.Utility based pricing reduces the total service cost by
which we had used. charging customers on a per-use basis. Various cloud
providers reduce the unnecessary cost on the resource
Multi-tenancy: Cloud supports multi-tenancy approach i.e this consumption [1].
approach is helpful when multiple clients wanted to avail the
services that are co-located in the single data centre. Multi- II. RELATED LITERATURE
tenancy models should have policies for isolation, security and
separation.
In this era of competition we need customized cloud which
cater to individual needs Cloud can be of different types like
Shared resource pooling: The concept of cloud computing is
based upon sharing of resources .Cloud providers like internal and private, or a hybrid cloud that combines both.
Cisco, Citrix, Google, IBM (SoftLayer), Oracle, Microsoft Cloud is pool of resources which can be scaled up and down
(Azure), and SAP, Rackspace, and Verizon (which acquired according to the requirements of the clients.
Virtualization is the essential for cloud computing. VMware is  Second, you will solve the performance issue
leader in virtualization powers your cloud infrastructure mentioned before.
providing different platforms to build and run robust cloud  Third, you add a layer of mobility that allows you to
applications, and supplying end-users cloud-based services. take your virtual machines with you (virtual
[3] machines are just files).
 Fourth, you will reduce the fragmentation and wear
As x86 virtualization [3] has become established and cloud
computing emerges, a robust IT infrastructure actually on internal hard disks.
matters more than ever. Today many IT organizations trying
to improve themselves from consolidated simple server to External hard disks also solve any scalability issues you might
virtualized tier-1[4] applications. They were also looking have. Any time you need more space, you simply add another
forward to include cloud infrastructure designs in their IT external hard disk. Without meddling internal layout . They
strategies so that the need for complete, bullet-proof can be easily switched, upgraded, ported, moved about
availability increases for high-performance, high-density or without touching the productivity setup[7].
mission-critical services and applications.By the use of server
with fault tolerance characteristic improves the performance 3) Use remote connectivity to solve physical network
increases drastically and it becomes difficult to breach. layout complexity
Using fault-tolerant server hardware hardens against new
vulnerabilities and complements the resilience you can
Both Virtual Box and VMware support remote connections.
achieve with virtualization and cloud computing.
This means that you can run virtual machines located on a
separate machine without any physical access to that machine,
save a network connection [8].
III. BENEFITS OF VIRTUALIZATION
4) Split virtual machines into smaller files for easier
Virtualization products include VMware Player, VMware portability
Server, VirtualBox[5]. These products offer the desktop user a
simple and powerful way of encapsulating entire operating
systems as guest machines for the purpose of study, testing Virtual machines are files. We can split them into smaller
etc. The only downside of these products is that they are rather segments to make them more easily portable[9].
big, take lots of resources, mainly RAM and hard disk space.
Different ways to utilize precious and limited resources to 5) Pre-allocate disk space to reduce fragmentation, disk
the maximum, without making complex changes to your access times
environment. []
Virtual hard disk fragmentation restrict contiguous
1) Virtualization makes your machines faster by placing segmentation which reducing future fragmentation and thus,
them on non-OS hard disk disk access times. Thus pre allocation of disk space is not a
good idea if virtual machines are large. But the pre-allocation
benefits both the host and the virtual machines. They will not
If you only have a single hard disk, your virtual machines have to fight one another for space as virtual machines change
will run more slowly because the hard disk will have to dynamically with use over time, improving performance[10].
access both your real OS and the guest OS files
simultaneously. Placing virtual machines on the second /
third hard disk will significantly increase performance. [6] IV VIRTUALIZATION ISSUES

2) Solve limited hard disk space; place virtual machines 1) Single Points of Failure
on external disks
If a number of virtual machines are migrated to a single
physical server like in cloud environment according to the
Virtual machines can easily take 4-6GB or more. Having tens client this can lead to single point of failure. Thus this
of virtual machines on your hard disk can eat valuable space flexibility can introduce single points of failure when too
that you might not be able to afford to spare for this task. many of the virtual machines that support a particular service
or application are concentrated on the same physical server.
Using external hard disks serves many purposes. Close attention must be paid to how many virtual machines
reside on a physical server, and which virtual machines are
 First, you will be able to free your internal hard disk placed on which physical server(s)[11].
space for other stuff.
2) I/O and Scalability Limitations
[3] Adams, Keith, and Ole Agesen. "A comparison of software
In virtualized and cloud environments virtual resources have and hardware techniques for x86 virtualization." ACM SIGOPS
to be mapped to physical resources, including the physical Operating Systems Review 40.5 (2006): 2-13.
server’s buses, I/O adapters and disks. I/O and scalability [4] Bi, Jing, et al. "Dynamic provisioning modeling for
concerns increase when a virtualization or cloud solution uses virtualized multi-tier applications in cloud data center." Cloud
software lock stepping to provide high availability or Computing (CLOUD), 2010 IEEE 3rd international conference
software-based fault tolerance. Which imposes additional on. IEEE, 2010.
overhead on the CPU, network[11].
[5] Li, Peng. "Selecting and using virtualization solutions: our
3) Failover and Fault Isolation experiences with VMware and VirtualBox." Journal of
Computing Sciences in Colleges 25.3 (2010): 11-17.
As mentioned earlier, virtualization software enables failover:
[6]Shields, Greg. The Shortcut Guide to Implementing
the ability to transfer processing from one virtual machine to
Virtualization in the Small Environment. Realtimepublishers.
another virtual machine on a different server in the event of a com, 2009.
hardware or application failure (and during scheduled
maintenance). Storage capacity may be virtualized as well. [7] Gupta, Karan, et al. "GPFS-SNC: An enterprise storage
Virtualization and cloud software solutions on the market framework for virtual-machine clouds." IBM Journal of
today cannot isolate and then determine the root cause of a Research and Development 55.6 (2011): 2-1.
hardware or software failure. Even though failover is
provided for, the cause of the failure may be propagated to [8]Zhang, Qi, Lu Cheng, and Raouf Boutaba. "Cloud
the secondary virtual machine. That allows the same failure computing: state-of-the-art and research challenges." Journal
to repeat along with the same threats of downtime and data of internet services and applications 1.1 (2010): 7-18.
corruption. In addition, finding and resolving the cause of a
[9] Bernstein, David, et al. "Blueprint for the intercloud-
failure within a virtualization software stack poses a
protocols and formats for cloud computing
challenge due to the increased software complexity and a lack interoperability." Internet and Web Applications and Services,
of self-diagnostics. 2009. ICIW'09. Fourth International Conference on. IEEE,
2009.
4) Data Integrity [10]Gokhale, Parag, and Michael F. Klose. "Systems and
methods of data storage management, such as pre-allocation
Virtualization or cloud software alone do not provide the of storage space." U.S. Patent No. 7,831,766. 9 Nov. 2010.
complete protection required. Avoiding a “split-brain” [11] Rimal, Bhaskar Prasad, Eunmi Choi, and Ian Lumb. "A
condition is crucial to preventing data corruption as well. Taxonomy and Survey of Cloud Computing Systems." NCM 9
Split brain occurs when a server or servers go down; (2009): 44-51.
[12] Chen, Deyan, and Hong Zhao. "Data security and privacy
they reboot or restart; and each virtual machine believes
protection issues in cloud computing." Computer Science and
itself to be the primary VM. Selecting a virtualization
Electronics Engineering (ICCSEE), 2012 International
solution that prevents split brain is good practice Conference on. Vol. 1. IEEE, 2012.
anytime you plan to provide redundancy by using
multiple physical servers and network paths in a
virtualized or cloud setup[12].

V CONCLUSION AND FUTURE SCOPE

The paper presents briefly the concepts of cloud computing


and the characteristics of the cloud environment.
Virtualization which is the main technology behind cloud
computing is discussed in the paper. The benefits of
virtualization is also discussed. The limitations is briefly
explained and will opens the scope of future research in the
area of virtualization in cloud computing

VI. REFERENCES
[1]Mell, Peter, and Tim Grance. "The NIST definition of
cloud computing." (2011).

[2] Xing, Yuping, and Yongzhao Zhan. "Virtualization and


cloud computing." Future Wireless Networks and Information
Systems (2012): 305-312.

You might also like