International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)

Volume 3 Issue 8, August 2014

Performance Monitoring for Better Management of

Cloud Server using Virtualization in Cloud
Computing
V. Pruthvi Chaithanya Varshu Dr. K. Santhi Sree Ph.D T. Mahesh Kumar
(M.Tech), CSE Department, Professor, CSE Department, (M.Tech), CSE Department,
Jawaharlal Nehru Technological Jawaharlal Nehru Technological Student Member of IEEE,
University, Kukatpally, Hyderabad, University, Jawaharlal Nehru Technological
[email protected] Kukatpally, Hyderabad, University, Kukatpally, Hyderabad,
[email protected] [email protected]

Abstract: Any server performance can be improved when The interface is almost same as the windows OS. We have a
we understand the current usage of the machines working desktop and task manager to see performance and control
under it. The performance of the server according to each
panel for the assistance. Windows Server 2012 has robust
category can viewed by installing several tools in the server.
we always have to analyze performance of server side recovery options to run the important workloads [2].
machines because in cloud computing environment the client We can build, deploy, operate, and monitor applications
always scale resource need according his requirements. This on premises and in the cloud. The server provide any
paper proposes an analysis of how cloud servers are applications for the infrastructure they want.
accessing/managing the performance of virtual machines
running under them because the better we understand the
current situation-the better we improve performance.

Index Terms---Cloud server, Resource Monitoring,

Windows Server, Virtualization, Hypervisor.

I. INTRODUCTION
A. Cloud Server
Cloud servers are the virtual servers which run on cloud
computing environment. In cloud hosting, a client takes rent
the virtual space instead of buying the physical servers.
They will pay to the server for which the amount of time
they use the virtual server. By using cloud servers the
clients can scaled up or scaled down. We access what we
need, when we need it, and then only pay for what we’ve
used [1].
Fig 1: Windows Server 2012 Start Screen
Cloud servers can be run on a hypervisor.  Hypervisor
will control the capacity of operating systems so it is
We mainly have “server manager” and “windows power
allocated where needed. With cloud computing multiple
shell” for managing the windows server. The windows
servers will be available to each particular client. If one
server can be installed either way with server core or with a
server fails, others will take its place. The security is
GUI. Server manager is used to manage the multiple
provided to each content in the cloud and the services
servers. The task manager in the windows server lists all
offered on demand. The client has to pay as much as he
details as in windows 8 system like application names and
used rather than paying for total equipment which can’t
status, as well as CPU, memory, hard disk and
beused always.
network utilization. The Performance tab shows "CPU",
B. Windows Server "Memory", "Disk", "Wi-Fi" and "Ethernet" graphs. The
Windows Server is a combination of operating systems. windows server has only 64 bit version. Windows Server
It can be used as a cloud at server side. Windows Server is 2012 has an IP address management role for discovering,
upgraded with the windows operating system. The cloud monitoring, auditing, and managing the IP address space
environment also do virtualization. There are many servers used on a corporate network.
like xen, but windows server has built in user interface.

ISSN: 2278 – 1323 All Rights Reserved © 2014 IJARCET 1

 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 3 Issue 8, August 2014

C. Citrix Xen Server M

Xen is an open source, Type-I hypervisor. When Xen Hot- Add:vCPU none No hot-
boots, it creates an initial virtual machine called Domain-0. add/remove , VMem add CPU,
It is a control domain through which other VMs can be virtual Add:Virtua memory
created and managed.Xen follows the approach para- resources l disk
virtualization where the guest OS is modified to take Power DPM: No cluster No
advantage of the knowledge that it is running on top of a Efficiency Cluster-leverl level cluster
VMM to optimize certain operations. Xen server deliver power management level
high performance on usually expensive operations such as management manageme
network I/O. The hardware support from both Intel and nt
AMD allows Xen to support fully-virtualized VMs that can Storage Thin Thin disk, Only by
host unmodified guest OS images [3]. Usage provisioning but no way of
Efficiency monitoring monitoring third party
II. PROPOSED WORK tools tools storage
vendor
By analyzing the performance of the server side
components we can take decisions about adding or Network Distribute none none
Management switch
removing the more virtual machines. We can also
dynamically allocate the memory or CPU cycles to the Efficiency Third-party
virtual switch
virtual machines. But to know which machine has to be
modified according to the client requirements is known by Automate Transparen Host Host
the performance monitoring. This paper proposes an d patching t host patch patching patching,
analysis of how cloud servers are accessing/managing the efficiency Auto guest causes VM but no auto
performance of virtual machines running under them Patching downtime guest
because the better we understand the current situation-the patching
better we improve performance.
By seeing the above table we can understand about their
III. EXPERIMENTAL RESULTS performance in several views. Three types of servers
compared here VMWare, Microsoft and Citrix. The table
A. Comparison of Servers shows that in all angles the VMWare shows good
There are many cloud service providers are available, performance capabilities. The servers are compared
everyone provides some resources and some feature according to their efficiency over a year period of time. We
oriented applications. They have their own way of using also compared xen and windows server in simple terms that
their resources, here we see some famous cloud server are easily understandable. The details are described in a
providers and compare their configurations. table below.

Table 1: Comparing the different cloud providers Table 2: Servers Comparison

according to Efficiency.
XEN Windows
MOST Server
VMWAR MICROS Cost Open 190 days trial
EFFICIEN CITRIX
E OFT period
T
Hardware 64 logical 24 32 bit Support & Extra Included in
scale out cores 512 GB Logical logical Maintenance Purchase
ram cores cores128G Monitoring Tools Third Party Built-in
1TB B RAM User Interaction Moderate Friendly
TAM Updates and Extra Immediate
CPU CPU HW- Requires Requires Patches Purchase
Efficiency assist or BT HW-assist HW-assist Security Level Vulnerable Open Source
virtual for Win
specific The above table describes the various significant
scheduler differences between Xen server and windows server. Xen is
Memory Memory Memory Memory available in open source, while the windows server is
Efficiency HW assist HW-assist HW-assist commercial but it offers a trial version. Even though the
overcommit/s xen is open its usage and maintenance is hard. A novice
haring can’t use the xen server directly because it do not have a
Virtual 8 way 4 way on 8-way Graphical User Interaction. If any updates released in the
Hardware vCPU WIN08 vCPU xen server have to be known manually and added/ installed
Scale out 255GMvRA 64GBvRAM 32GBRAM manually. To maintain the server in full conditional we

ISSN: 2278 – 1323 All Rights Reserved © 2014 IJARCET 2

 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 3 Issue 8, August 2014

have to buy the support. This is a free to use but pay to get server or for a virtual machine. We can use the task
patched strategy. For managing the xen server we need to manager to see the performance graphs of all virtual
have an extra windows machine. For saving memory in the machines running under a server. These graphs shows same
server the xen developers given the dashboard i.e., xen information as the resource monitor.
center has to be installed on another windows machine and
it has to be connected to the server in LAN. Where as in
windows server it has a built in monitoring system to view
all its usage as the windows client had. The monitoring
tools are provided by xen server are basic tools and these
have to be installed on every virtual machine that we create.
And for better monitoring tools we should have to install
third party performance monitoring tools. Since xen server
is open source any one can use may be in the wrong way.
The windows server maintenance is fully protected and its
source code.
Now we see the sample performance results in Xen
server and Windows server.
B. Performance in Xen
The Xenserver handled by the xen center. In xen center
we install xen tools which are available in xen server to
monitor the server performance. Each virtual machine has
to be installed these tools. The all machines and CPU’s
performance can viewed in a graph format as shown below.
Fig 3: CPU Performance in Task Manager

We can see the performance of memory and CPU and the

network usage over a period of time. The above screen shot
represent the CPU performance of 16 virtual machines of a
server. Besides these details we can also see the percentage
of utilization and the speed and also the number of
processes, threads, and handles. It will also display all
details about L1, L2, caches. How much time the server has
been accessed can also be known here.
(ii) ServerHVHealth
We can see the performance of virtual machines by
“ServerHVHealth” which is inbuilt component of windows
server. Here the graph contain the colors to show the
different memory utilizations.

Fig 2: Performance in Xen server

The performance of each CPU and each NIC card can be

viewed here. The resource changes made on server are
viewed here. According to time all the fluctuations and
CPU, memory modulations can be observed using this
tools. We can select only specified CPUs and threads to
identify the idle processes. This kind of view helps in
achieving the Green Computing.
C. Performance in Windows Server
(i) Task Manager
Windows server has many new features regarding the
server management. Using Hyper-V we can create any
number of virtual machines in a server and able to connect
to other servers. We can also create virtual machines in
other servers. The performance can be monitored for a Fig 4: ServerHVHealth Performance Monitor

ISSN: 2278 – 1323 All Rights Reserved © 2014 IJARCET 3

 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 3 Issue 8, August 2014

Below the graph we have the options to select the

specified components. For e.g. every time we may not need
to compare all of the virtual machines usage, we can select
some of them and generate a separate graph for them for
clear understanding. And also we have the options for
selecting, clearing, viewing in different aspects of a
particular data. These kind of performance details useful for
achieving better performance and to achieve green
computing.

IV. CONCLUSION
The performance of every virtual machine is viewed in
different ways. We have compared the different types of
performance monitors in different servers. By the current
performance details we can take the decisions for better
performance.

V. REFERENCES
[1] Interoute - International Telecommunications Service
Provider http://www.interoute.com/what-are-cloud-
servers
[2] Wikipedia: The Free Encyclopedia
http://en.wikipedia.org/wiki/Windows_Server_2012
[3] Citrix Xen Server
http://www.citrix.com/products/xenserver/overview.ht
ml

V Pruthvi Chaithanya Varshu is a

student pursuing M.Tech in Information
Technology in Jawaharlal Nehru
Technological University Hyderabad. His
interested areas are Cloud Computing and
Networking and Big Data.. Email:
[email protected]

Dr. K. SanthiSree is a Professor, Dept

of CSE, School of Information
Technology, JNTUH Hyderabad. She has
14 years of teaching and Research
experience. Her interested areas are Data
mining, Cloud computing, Algorithms,
Information retrieval Systems. She is a
member of various professional bodies like ISTE, IETE,
IEI, and CSI. Email: [email protected]

T.Mahesh kumar is a student pursuing

M.Tech in Information Technology in
Jawaharlal Nehru Technological University
Hyderabad. He is a Student member of
IEEE. His interested areas are cloud
Computing, Multimedia and Big Data.
Email: [email protected]

ISSN: 2278 – 1323 All Rights Reserved © 2014 IJARCET 4