Load Balancing in Cloud Computing Environment: A Comparative Study of Service Models and Scheduling Algorithms

Int. J. Advanced Networking and Applications
Volume: 08 Issue: 06 Pages: 3246-3252 (2017) ISSN: 0975-0290
3246
Load Balancing in Cloud Computing
Environment: A Comparative Study ofService
Models and SchedulingAlgorithms
Navpreet Singh
M.tech Scholar, CSE & IT Deptt., BBSB Engineering College, Fatehgarh Sahib,
Punjab, India
(IKG – Punjab Technical University, Jalandhar) navpreetsaini26@gmail.com
Dr. Kanwalvir Singh Dhindsa
Professor, CSE & IT Deptt.,
BBSB Engineering College, Fatehgarh Sahib,
Punjab, India
(IKG – Punjab Technical University, Jalandhar) kanwalvir.singh@bbsbec.ac.in
-------------------------------------------------------------------ABSTRACT---------------------------------------------------------------
Load balancing is a computer networking method to distribute workload across multiple computers or a
computer cluster, network links, central processing units, disk drives, or other resources, to achieve optimal
resource utilization, maximize throughput, minimize response time, and avoid overload. Using multiple
components with load balancing, instead of a single component, may increase reliability through redundancy. The
load balancing service is usually provided by dedicated software or hardware, such as a multilayer switch or a
Domain Name System server. In this paper, the existing static algorithms used for simple cloud load balancing
have been identified and also a hybrid algorithm for developments in the future is suggested.
Keywords: Round-Robin scheduling, Data Center, Priority based scheduling, Cloud computing, Load balancing.
--------------------------------------------------------------------------------------------------------------------------------------------------
Date of Submission: May 02, 2017 Date of Acceptance: May 16, 2017
--------------------------------------------------------------------------------------------------------------------------------------------------
1. INTRODUCTION
Cloud computing is a technology that hosts computing
services in centralized datacenters and provides access to
them through the Internet. The cloud is a pool of
heterogeneous resources [1]. Cloud computing is very
much a utility, like electricity: sold on demand, instantly
scalable to any volume, and charged by use, with the
service provider managing every aspect of the service
except the device used to access it [2]. Cloud load
balancing refers to distributing client requests across
multiple application servers that are running in a cloud
environment [3]. Fig.1 represents a basic cloud balancing
scenario.
Fig 1: Cloud balancing scenario
There are different types of clouds that can be
subscribed to depending on one’s needs. As a home
user or small business owner, one will most likely use
public cloud services.
1. Public Cloud - A public cloud can be accessed by
any subscriber with an internet connection and access to
the cloud space. E.g.- Google’s Gmail,dropbox.
2. Private Cloud - A private cloud is established for a
specific group or organization and limits access to just
that group. E.g.- Many healthcare, financial, trade and
banking institutions utilize private cloud computing to
maintain cloud confidentiality in highly sensitive
electronic records.
3. Community Cloud - A community cloud is shared
among two or more organizations that have similar
cloud requirements. E.g.- several organizations may
require a specific application that resides on one set of
cloud servers. Instead of giving each organization their
own server in the cloud for this app, the hosting
company allows multiple customers connect into their
environment and logically segment their sessions.
4. Hybrid Cloud- A hybrid cloud is essentially a
combination of at least two clouds, where the clouds
included are mixture of public, private, or community.

3247
2. REVIEW OF LITERATURE
In this section, main focus of the discussion is on the
research work related to the load balancing in cloud
computing. It helps to compare the load balancing
techniques available and conclude an optimized solution.
Thapar et al. [1] proposed the concept of proportion
weights in order to assign the workload to data centres.
On the other hand, other service broker policies based
broker decisions on the basis of location. The proposed
policy took into consideration the efficiency of underlying
hardware, which meant greater number of hardware
machines, means more virtual machines, hence large
number of cloudlets could be served.
Fan [8] proposed that the development of cloud
computing has received a considerable attention.
For cloud service providers, packing VMs onto a small
number of servers is an effective way to reduce energy
costs, so as to improve the efficiency of the data centres.
However allocating too many VMs on a physical machine
may cause some hot spots which violate the service level
agreement (SLA) of applications. Load balancing of the
entire system is hence needed to guarantee the SLA.
Tziritas [9] discussed the problem of virtual machine
(VM) placement onto physical servers to jointly optimize
two objective functions. The first objective is to minimize
the total energy spent within a cloud due to the servers that
are commissioned to satisfy the computational demands of
VMs. The second objective is to minimize the total
network overhead incurred due to:
(a) communicational dependencies between VMs,
and (b) the VM migrations performed for the transition
from an old assignment scheme to a new one.
Guo [10] suggested that the instances which are used to
process big data need higher CPUs and disks than other
kinds of instances. If the instances of disk resource
consuming are placed in the same physical node, clearly,
the disk I/O bandwidth would be used up quickly that
would affect the performance of the entire node seriously.
Guo hence proposed an instance placement algorithm
FFDL, which based on disk I/O for private cloud
environment dealt with big data that would adopt the disk
I/O load balancing strategy and reduce competition for the
disk I/O load between instances.
Garcia [11] discussed that load management in cloud data
centres must take into account: (a) hardware diversity of
hosts (b) heterogeneous user requirements (c) volatile
resource usage profiles of virtual machines (VMs) (d)
fluctuating load patterns, and (e) energy consumption.
Garcia hence proposed distributed problem solving
techniques for load management in data centres supported
by VM live migration. Collaborative agents were endowed
with a load balancing protocol and an energy- aware
consolidation protocol to balance and consolidate
heterogeneous loads in a distributed manner while
reducing energy consumption costs.
Hsieh [12] suggested that as a cloud data centre may be
located over many regions and the network environment
within a cloud data center may differ from traditional
ones, how Virtual Machines (VMs) are deployed will
influence service performance. Author, based on the
Eucalyptus cloud computing and Software-Defined
Networking platform, proposed a load balancing
scheduling mechanism that works on the current network
status between users and associated VMs to improve
the cloud services. Author also set up a node controller
on the same subnet and different subnet.
Dinita [13] described an optimized and novel approach
to an Autonomous Virtual Server Management System
in a `Cloud Computing' environment and it presented a
set of preliminary test results. One key advantage of
this system is its ability to improve hardware power
consumption through autonomously moving virtual
servers around a network to balance out hardware loads.
This has a potentially important impact on issues of
sustainability with respect to both energy efficiency and
economic viability.
3. ROLE OF SERVICE BROKER
IN CLOUD COMPUTING
A cloud broker is an intermediary between the
provider and the purchaser of the cloud computing
service. It is generally a third party individual or
sometimes a business entity.
Cloud provider1 Cloud provider2
Service broker
Cloud consumer
Fig.2: Role of service broker
Fig.2 above represents the role of a service broker which
facilitates the distribution of work between different cloud
service providers.
The entire process of serving a client is a part of any one
of the services defined in the service model. It begins with
a request for a particular resource or application, be it for
development, or just accessing the storage of the service
provider. The request is serviced by the cloud service
provider through a series of steps, the first one passing
through a cloud service broker, which acts as the
intermediary between a cloud consumer and the cloud
service providers. The service broker makes use of any
one of the available service broker policies in order to
send the request to the most appropriate data center. The
role of a service broker is shown in fig.2 above.

3248
After choosing the data center that is going to perform
computation, the load balancer at the data center comes
into action. It makes use of the implemented load
balancing algorithms to select the appropriate virtual
machine to which the request has to be sent for execution.
The innermost abstraction layer comprises virtual machine
management. The virtual machine manager is responsible
for the management and migration of virtual machines in
the cloud data centers. Out of the above tasks, the use of
an efficient service broker policy is quite necessary to
ensure that the later tasks are carried out with efficiency
and least response time.
4. CLOUD SERVICEMODELS
Cloud computing is a convenient on-demand network
access model enabling the access to a shared pool of
configurable computing resources. Cloud model is
composed of three service models.
Cloud computing is able to provide a variety of services at
the moment but main three services are Infrastructure As-
A-Service, Platform-As-A-Service and Software-As-A
Service, also called as service model of Cloud computing
described in fig.4.
a) Software-As-A-Service (SaaS)
It describes any cloud service where consumers are
able to access software applications over the internet. The
applications are hosted in “the cloud” and can be used for
a wide range of tasks for both individuals and
organizations. There are a number of reasons why SaaS is
beneficial to organizations and personal users alike. e.g.-
Google Apps, Salesforce, Twitter, Facebook, etc.
 No additional hardware costs: the processing
power is supplied by the cloud provider.
 No initial setup costs: applications are ready to
use once the user subscribes.
 Pay for what you use: if a piece of software is
only needed for a limited period then it is only
paid for over that period and subscriptions can
usually be halted at any time.
 Usage is scalable: if a user decides they need
more storage or additional services, it can be
subscribed to at any time.
 Updates are automated: update is available
online
to existing customers, often free of charge.
 Cross device compatibility: SaaS applications
can be accessed via any internet enabled device,
 Accessible from any location with an internet
enabled device.
 Applications can be customized: with some
software, customization is available.
Fig.3: Cloud Computing Service model applications
Fig.4: Basic cloud service models description
b) Platform-As-A-Service (PaaS)
It provides a development platform to its users so that
the user can develop and maintain respective
applications and cloud specific utilities. It is different
from SaaS because SaaS is a developed and deployed
application whereas PaaS provides a platform or ground
to develop those applications. PaaS provides
development environment and platform, so all
supporting material i.e. programming environment,
development tools and infrastructure etc. must be
provided by cloud provider. e.g.-Google App Engine,
WordPress, etc.
 The users don’t have to invest in physical
infrastructure. This leaves them free to focus on
the development of applications.
 Makes development possible for ‘non-
experts’:
With some PaaS offerings anyone can develop
an application through their web browser
utilizing one-click functionality.
 Flexibility: Customers can ‘pick and choose’
the features they feel are necessary.
 Adaptability: Features can be changed if
circumstances dictate that they should.
 Teams in various locations can work together:
As an internet connection and web browser are

3249
all that is required, developers spread across
several locations can work together on the same
application build.
 Security is provided, including data
security, backup and recovery.
c) Infrastructure-As-A-Service (IaaS)
As with all cloud computing services, it provides access to
computing resource in a virtualized environment “the
Cloud”, across a public connection, usually the internet. In
case of IaaS, the computing resource provided is
specifically that of virtualized hardware, in other words,
computing infrastructure. The definition includes such
offerings as virtual server space, network connections,
bandwidth, IP addresses and load balancers. Physically,
the pool of hardware resource is pulled from a multitude
of servers and networks usually distributed across
numerous data centers, all of which the cloud provider is
responsible for maintaining. The client, on the other hand,
is given access to the virtualized components in order to
build their own IT platforms.
e.g.- Salesforce.
 Scalability: Resource is available as and when
the client needs it and, therefore, there are no
delays in expanding capacity or the wastage of
unused capacity.
 No investment in hardware: The
underlying physical hardware is set up and
maintained by the cloud provider.
 Utility style costing: The client only pays for the
resource that they actually use.
 Location independence: The service can usually
be accessed from any location as long as there is
an internet connection.
 No single point of failure: If one server or
network switch, for example, were to fail, the
broader service would be unaffected due to the
remaining multitude of hardware resources and
redundancy configurations.
5. TYPES OF SCHEDULING
ALGORITHMS
The scheduling algorithms are aimed on improving the
performance and the quality of service by reducing the
execution time and costs. The various scheduling
algorithms are as follows:
5.1 Round-robin load balancing is one of the simplest
methods for distributing client requests across a group of
servers. Going down the list of servers in the group, the
round-robin load balancer forwards a client request to
each server in turn. When it reaches the end of the list,
the load balancer loops back and goes down the list again
(sends the next request to the first listed server, the one
after that to the second server, and so on).
Figure 5 below represents assigning of various jobs to
servers for their execution in Round Robin fashion.
Fig.5: Round-Robin balancing technique
It does not always result in the most accurate or efficient
distribution of traffic, because many round-robin load
balancers assume that all servers are the same: currently
up, currently handling the same load, and with the same
storage and computing capacity. The following variants to
the round-robin algorithm take additional factors into
account and can result in better load balancing.
5.2 Weighted round robin - A weight is assigned to each
server based on criteria chosen by the site administrator;
the most commonly used criterion is the server’s traffic-
handling capacity. The higher the weight, the larger the
proportion of client requests the server receives. If, for
example, server 1 is assigned a weight of 3 and server 2 a
weight of 1, the load balancer forwards 3 requests to
server 1 and for each 1 it sends to server 2.
As shown in figure 6 below, server 1 is assigned a weight
of 5 and server 2 is assigned a weight of 6. So a request 6
having weight 5 is assigned to server 2 while all others
are assigned to server 1.
Fig.6: Weighted Round-Robin load balancing
technique
5.3 Dynamic round robin - A weight is assigned to
each server dynamically, based on real-time data about
the server’s current load and idlecapacity.
5.4 Priority based scheduling – A priority is assigned
to each request and then the request is processed
depending on its priority. The requests of Equal priority
are scheduled in FCFS order [2]. Priority of a request
can be either defined externally or internally. Priorities
defined internally for a request are computed using
some measurable quantities or qualities. To each
admitted queue, a priority is assigned.

3250
To increase the performance of the overall system, all the
resources shall be evenly distributed to satisfy the
customer requirement by distributing the load
dynamically among the nodes. Figure 7 below depicts a
basic scenario in which a job request is serviced by a
cloud platform to achieve the maximum performance.
Cloud Request Number of jobs
Define bandwidth
Scheduling Algorithm
Minimization of Make span
Computation Time, Load and respond time
Fig.7: A basic cloud load balancing scenario using any approach.
6. COMPARISON OF SERVICE MODELS AND SCHEDULING
ALGORITHMS
There are different types of service models in cloud
computing atmosphere as follows: Table 1: Service models in cloud computing
S. No. FEATURES SaaS PaaS IaaS
1 Feature Software delivered over web. Platform delivered over web,
for creation of software.
Infrastructure (software or hardware)
delivered on web as an on demand
service.
2 Offerings
User has nothing to worry about.
A pre configured package as per
requirement is given and payed
accordingly.
User gets what is demanded.
Hardware, Software, Web
environment, OS. Payment is
made accordingly and user gets
the platform to use.
User gets the infrastructure and
pays accordingly. Can install any
OS, composition or software.
3 Level
Complete pack of all
services.
Top of IaaS Basic layer of computing.
4 Feasibility
Used by a variety of users. Used
over web on various locations
(home, road, office).
All technical stack
requirements met by the
platform offerings.
For people or companies not
willing to invest too much on
hardware. For those trying to do
something temporarily.
5 Technical skill
requirement
No need of any technical
knowledge.
Knowledge of the subject is
required. Only the basic setup
is provided.
Technical knowledge is
required.
6 Deals with
Only applications (like
Gmail,Yahoo, etc ).
Social Networking sites (like
Facebook)
Runtimes, Database and
web servers.
Virtual machine storage, load
balancers, network, servers.
7
Consumption
graph
Most widely used among a
common man or companies
which that don’t have to worry
about technicalities.
Popular among developers as
they don’t need to worry about
traffic load or server
management.
High popularity among skilled
developers or researchers who
have need of custom
configuration.
8 Disadvantage
1. Security concern.
2. Certain organizations have
regulation related to where data
is stored.
1. Limitedflexibility.
2. Integration problem with the
in-house systems and the
application as it could trigger
an increase in complexity.
Dependence on a specific
provider. Also to mitigate any
security relates risk, it is important
to consider what data is to be sent
to the cloud.

3251
From the Table 1 above, it can be concluded that a cloud
can be scaled dynamically as per the needs of the users,
and also, there is no need for any company to deploy its IT
staff to manage this service, since the cloud service
provider is responsible for providing the software and
hardware necessary for the service.
The comparative study of the scheduling algorithms (as
shown in Table 2 below) has its advantages as it gives its
user a better picture of the appropriate class of scheduling
algorithms available for different types of required
services as per the requirements of consumers and service
providers.
Table 2: Comparison of various Scheduling algorithms
7. CONCLUSION
Scheduling is a major issue in the management of service
requests in cloud environment. Various phases have to be
used for the development of the load balancing system in
the cloud computing environment. These different phases
have to be implemented for the completion of the
proposed work. Load balancing has been done by dividing
different tasks into a number of jobs so that they can be
allocated to different resources for processing to complete
in less computation time. In cloud computing scenario,
number of tasks has to be assigned on various processes to
handle load on the cloud. These tasks have been divided
into sets and the dependency checking is done for
prevention of dead lock state or to prevent demand of
various extra resources for allocation. Hybrid algorithm is
better than others because unlike PB scheduling, it
automatically increases the priority for the old processes
having low initial priority, hence executing them
eventually.
REFERENCES
[1] K. Kishor, V. Thapar, “An efficient service broker
policy for Cloud computing environment”, International
Journal of Computer Science Trends and Technology
(IJCST), Vol. 2, Issue 4, July-Aug 2014.
[2] Pinal Salot, “A Survey of Various scheduling
algorithms in cloud computing environment”, ISSN: 2319
- 1163, Vol.2, Issue 2, pp. 131-135, June 2014.
[3] Z. Xiao, W. Song, and Q. Chen, “Dynamic resource
allocation using virtual machines for cloud computing
environment,” IEEE Transactions on Parallel and
Distributed Systems, Vol. 24, No. 6, pp. 1107–1117, 2013.
[4] L. D. Babu and P. V. Krishna, “Honey bee behavior
inspired load balancing of tasks in cloud computing
environments,” Applied Soft Computing Journal, Vol. 13,
No. 5, pp. 2292–2303, 2013.
[5]Y.Zhang, “Dynamic load-balanced multicast based on
the Eucalyptus open-source cloud-computing system”,
pp. 456 – 460, IEEE, 2011.
[6] R. Basker, V. R. Uthariaraj, and D. C. Devi, “An
enhanced scheduling in weighted round robin for the
cloud infrastructure services,” International Journal of
Recent Advance in Engineering & Technology, Vol. 2,
No. 3, pp. 81–86, 2014.
[7] Y. Wen, “Load balancing job assignment for cluster-
based cloud computing”, pp. 199 – 204, IEEE, 2014.
S. No. SCHEDULING
ALGO
Round-Robin Weighted Round Robin Priority based
1 Concept Designed specifically for
time sharing systems.
Designed to handle servers better
depending upon their processing
capabilities.
Designed to schedule the serving of
requests based on their priority.
2 Implementation
Similar to FCFS but each
request is served for a fixed
interval of time. All requests
are kept in a circular queue
known as ready queue.
Each server is assigned a weight,
integer value that describes the
processing capacity. Higher the
weight, higher the number of
connections received by the server.
It involves assignment of priority to
every request. Requests with high
priority are served first, while ones
with same priority are served in FCFS
order. In case of low priority job,
remedy to starvation is aging, in which
priority gradually increases for jobs
that are in queue for long period of
time.
Main advantage - it is Weights assigned to servers
3 Advantages Starvation free. create a longer time slice, hence
making it starvation free.
Important jobs are served first.

3252
[8] Z. Fan, “Simulated-Annealing Load Balancing for
Resource Allocation in Cloud Environments”, IEEE
International Conference on Parallel and Distributed
Computing, Applications and Technologies, pp. 1-6,
Taipei, 2013.
[9] N. Tziritas, “Application-Aware Workload
Consolidation to Minimize Both Energy Consumption
and Network Load in Cloud Environments”, IEEE
International Conference on Parallel Processing, pp.-449-
457, Washington D.C., USA, October 2013.
[10] J. Guo, “An instances placement algorithm based on
disk I/O load for big data in private cloud”, IEEE
International Conference on Wavelet Active Media
Technology and Information Processing, pp. 287-290,
2012.
[11] J. O. Garcia, “Collaborative Agents for Distributed
Load Management in Cloud Data Centres Using Live
Migration of Virtual Machines”, IEEE International
Conference on Services Computing, pp. 916-929, 2015.
[12] W. K. Hseih, “Load balancing virtual machines
deployment mechanism in SDN open cloud platform”,
IEEE International Conference on International
Conference on Advanced Communication Technology,
pp. 329-335,2015.
[13] R. I. Dinita, “Hardware loads and power
consumption in cloud computing environments”, IEEE
International Conference on International Conference on
industrial Technology, pp. 1291-1296, 2013.
[14] A. Goyal Bharti, “A Study of Load Balancing in
Cloud Computing using Soft Computing Techniques”,
International Journal of Computer Applications (0975 –
8887) Vol. 92, No.9, April 2014.
[15] N. Kaur, T.S. Aulakh, R.S. Cheema, “Comparison of
Workflow Scheduling Algorithms in Cloud Computing”,
International Journal of Advanced Computer Science and
Applications, Vol. 2, No. 10, 2011.
[16] M.S. Rana, S. Kumar, N. Jaisankar, “Comparison of
Probabilistic Optimization Algorithms for Resource
Scheduling in Cloud Computing Environment”
International Journal of Engineering and Technology, pp.
153-163, Vol. 3, No.6, July 2016.
[17] C. Kalpana, U. Karthick Kumar, R. Gogulan, “Max
- Min Particle Swarm Optimization Algorithm with Load
Balancing for Distributed Task Scheduling on the Grid
Environment”, IJCSI International Journal of Computer
Science Issues, Vol. 9, Issue 3, No. 1, May 2012.

Load Balancing in Cloud Computing Environment: A Comparative Study of Service Models and Scheduling Algorithms

Recommended

More Related Content

What's hot (20)

Similar to Load Balancing in Cloud Computing Environment: A Comparative Study of Service Models and Scheduling Algorithms (20)

More from Eswar Publications (20)

Recently uploaded (20)

Load Balancing in Cloud Computing Environment: A Comparative Study of Service Models and Scheduling Algorithms