IJRECE VOL.

6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)

A COMPARSION OF LOAD BALANCING STRATEGY IN

SOFTWARE DEFINED NETWORKING
Prabhjot Kaur Dr. Abhinav Bhandari
M.Tech (Scholar) Assistant Professor
Department of Computer Science & Engineering, Department of Computer Science & Engineering
Punjabi University, Patiala Punjabi University, Patiala
[email protected] [email protected]

Abstract— SDN is the present big thing in network In SDN (Software Defined Network), controller plays
industry. This report includes the SDN architecture, benefits important role to improve the control plane and monitor the
and simulation best practice using Mininet and POX behaviour of network as supervisor provides software
controller. It also explains and helps in understanding how interfaces of system and maintenance of global viewpoint of
the controller can be implemented on AWS EC2 Linux network. In the manner new functions and management job
machine and creating load balancer using POX load can be done by the several applications.
balancer algorithm. Cloud based SDN load balancer helps in
balancing the load on machines running in different or same
data centers. Software defined network load-balancesa are
programmable and allow it to implement and develop your
own load balancing stages. Other beenfits of software
defined network load balancer is it don’t require dedicated
software. In this paper, we implement and comparing the
various method like round robin and scheduling algorithm
with already designed random methods using NS-2
Simulator.

Keywords— SDN (Software Defined Networking), EC2

(Elastic Compute Cloud), VLAN (Virtual Local Area
network) and AWS (Amazon Web Services).
Fig.1. Programmatic Abstracted Interaction with Network
I. INTRODUCTION
Data plane forwards conventional flows allowing to the
Today internet applications require fast networking which protocols resulting from the supervisor and restrictions are
caused large amount of traffic and deploys number of put by the controller on all flow through consecutively of
distinct, dynamic applications and services. This user-defined control requests. Benefits of software defined
management makes network highly complex and is more network are lost due to the failure of controller in the
difficult for the network administrator to handle concepts network [3].
like network virtualization also interchangeable data is
A. Characteristics of SDN
added [1]. The configuration and installation of the network
demands skilled personnel. The network nodes such as The features of software defined network are Separating
switches and routers has complicated interactions. System switch plane after data-plane, network view, central
based approaches requires simulation of elements but due to supervisor, open boundaries between the equipment and
programming interfaces it is difficult to access [2]. network program ability [2]. System devices include control
plane and data plane that decide how network traffic is
The term software defined networking (SDN) was coined
handled, and traffic forwarding according to the control
few years back but its idea is changed since 1996 as driven
plane decision respectively [1]. Communication network is
by user control management need to forward in the network
made up of end users, hosts connected by the infrastructure
nodes. There are many projects such as Ipsilon, The Storm
of network and hosts share the infrastructure and use
and Internet Engineering Task Force (IETF) Forwarding and
routers, communication links, and switches to carry data [4].
Control Element Separation, 2000. Latest, Ethane (2007)
and Open Flow (2008) required the usage of the software The research work is partitioned into different section. The
defined network. Open Flow facilitates the flow table by first section is about the basic information about SDN. The
enabling entries which are defined by external server. section II involves the architecture of SDN. Third section

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1018 | P a g e
 IJRECE VOL. 6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)
composed the literature survey and the fourth section is In the figure 3 the idea of SDN is shown that is detachment
about the proposed work and methods. It involves load of the control layer and merging into a single point of
balancing and the available protocols. Further the result is network. It refers to the processing of every single device in
discussed and at last the conclusion and future scope is the network pass the packets of data from the particular
given. place to the next coming place under the choices of SDN
controller.
II. SDN ARCHITECTURE
Through the API particular controller controls the particular
The information technology have incapablility to acquire switch and the application layer SDN applications is
the bestest implementation via the system such as the the directed with the controller. Following terms are the layers.
traditional software devices. For instance the routers, 1) Data or Information Layer: The hardware exists
switches and the load balances. Subsequently, for enhancing that are physically connected in the layer of the
the security the firewalls are considered to get the things network and the programming is processed on the
properly managed. The requests of the system customers or system gadgets which gives the basic control plane
the clients are expands one by one. There is the necessity to interface. This is later associatred to the upper
manage the every single gadget completely. The number of dimensions of the southbound API.
gadgets are varied and come under the thousands and more. 2) Control Layer: The control layer is the most
The process to manage and to set the protocols and other essential layer in the network and it specifically
devices are boring as well as difficult job. For instance the saced the topology track which contains the
need of switches are because to include and to diminish the cotroller and later it used for the communication of
gadgets from the particular system, ACL (Access Control different devices in the particular framework.
List), other switches and the firewalls to enhance the When the data is shuffled from the state of the
security. The process is entirely unpredictable, long and network to the upper layers via using the
certain problems occurred [5]. Northbound API. The controls worked to generate
the sensible directions for the gadgets.
In the system gadgets, there are three planes as the data, 3) Application layer: It defines the features, services
control and the applications. The basic employement of IP and policies. Applications create new features with
(Information Plane) is transmitted the data. The basic big decisions according to the network changes.
employement of software defined network is controlled the Whenever the topology or the feature changes the
transmitted plane. It is finalized the application plane via applications get the control to change the behaviour
driving the stream rules under the transmitted plane. In these of the network which can done from a single point.
system gadgets, the infromation, control and application
planes are managed as shown in fig.2. The overall setyp is B. North-bound API
easy and simple due to the deattachment of the information
and the control plane [6]. The open flow conventions are The controller gives the Northbound API and to manage
used to symbolize SDN (Software efined Network). The communication by the applications they have to use the API
programming characterized organizing is the fundamental to get to the controller. The whole SDN network is
convention which trained for transmission plance and controlled by this method. It is applied by restful API as
control plane. For this purpose, the message trade is used through this it makes easier to control with the usage of
via Open Flow convention [7]. HTTP methods such as POST, PUT, GET and DELETE.

C. South-bound API

The South-bound API gives the correspondence interface

among the controller and the different network approaches.
This is the basic and the central for the software defined
network (SDN) controller while the administration of
switches is done through the API by the controller. The
controller of SDN has the responsibility of seeing the
network. The configurations are essential flows for every
switch in the network. The switches are talking about the
transmission of packets. In some situations the switches are
become incapable to take the final decision then the process
is set to the default action.

Fig.2. SDN Architecture D. SDN Controllers

A. Three layers in SDN Architecture

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1019 | P a g e
 IJRECE VOL. 6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)
The Software Defined Network controllers coined in the systems. the dependency is on the Beacon. The flood-
short term as SDN controller and are considered as the mind light is the apache authorized and relied on the Jave
of the system. The software defined network controller with the open flow controllers. The framework of
covered the essential control points in the network that flood-light with the API is utilized differently with the
worked with the stream control and the switches via the big switch systems and it is seen in BNC (Big Network
south-bound API. The business logic is completed by using Controller).
the north-bound API. In this way, the controllers are 5) Open Day-light: An Open Day-light venture is come
assumed as the smartest networks. The software defined under the open souce phases specifically for SDN
networks are relaint upon the servers mainly to guide the (Software Defined Networking) which used the open
changes in which the data packets are send and depends conventions to give access for automatic control
upon the open flow protocol. andchecking of system gadgets. Similarly, the other
terms of software defined networks, Open Day-light
1) NOX: NOX is both an exceptionally old form of the under the Open Flow, and it manages the
controller and it is relied on the modules design to arrangements as a feature of its stage.
create the fundamental applications of SDN. Further,
NOX gives the bestest modules correction openflow III. LITERATURE SURVEY
that are later increased. The centre of NOX gives
partner techniques and APIs to interfacing with Sukhveer Kaur et al., [8] introduced an SDN application
OpenFlow switches, including an association handler for executing server load balancing. The basic idea of SDN
and occasion motor. Extra parts that advantage that consists of logical centralized network control in SDN
Programming Interface is available, including host controller. To solve the issue of unnecessary latency they
following, steering, topologies along with the Python implemented direct routing based load balancing algorithm.
line executed as a covering for the module
Programming Interface. Mu Sub et al., [4] proposed a two LVS based technique to
2) POX: POX is the more up to date; the dependency is decrease the problem of over-head of the supervisors and
upon the python which is associated to NOX. Further, forward loop, and to reduce controller synchronization.
the imporvements are returned NOX for its basic C++ When load of a server exceeds an amount more than the
and to generate the well defined Python base such as threshold the LVS handles the synchronization state.
Python 2.7. Consequently, it composed of the abnormal
states with the capability of topology charts as well as Luo et al. [9] proposed PTAA that can erase the problem of
the virtualization. balancing in a way of effective parallel. Addressed load
a) It has a Pythonic OpenFlow interface. balancing of in band manage rush and evaluated the
b) It has also reusability which are the examples to performance confirmation of demonstrating communication
determine the topology and to discover. of PTAA with least eighty percent delay reduction through
c) POX capable to process everywhere and it is an approach which is fast and had low complexity. The
comprssed with the freee PyPy implementation location of the controller is additional used for the finding
specifically for the easiest utilization. the optimize network end to end delay.
d) It explicitly focused on the Linux operating system,
Max operating system and Windows. D. Perepelkin et al. [10] developed software application
e) It bolsters are the easiest and the simple Graphical user and visualized design environment in software defined
interferences and contains the perception apparatuses networking with quality of service. The paper has focused
like NOX. on set pavement of reserved routes that are set accordingly
f) It makes all around contrasted with NOX services to the indicator of QoS and paved route based on paired
written in Python. shifting with balance the load. Giving to the indicators of
3) Ryu: Ryu is part based and the open source introduced way difference the alignment and optimization of load is
by NTT labs. The framework implemented entirely in done.
the programming of Python. Ryu informed the
S. Attarha et al., [11] to avoid the over utilization of link in
administration supports parts created in different
the environment of software defined networking proposed
dialects. The open flow wired bolster composed the
an efficient algorithm. The manager is accountable for the
Nicira expansions, occasion the administrators,
monitoring the condition of the system and decide switches
informing, in memory state the board, application the
to be used occasionally. The new movement is transmitted
administrators, foundation administrations benefit. For
towards the possible direct path but in case when path is not
instance NETConf libraries and the netlow libararies.
blocked with traffic otherwise controller find a back-up
Ryu likewise bolsters has the REST interface to its
path. To resolve the issue of traffic overloading the
OpenFlow tasks.
algorithm selects the lowest possible number of flows and if
4) Flood-Light: this is the rapidly fascinating in the SDN
not they are directed to the backup paths.
controllers that initialized to impact from the beginging
of Big switch systems specificially for the open source

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1020 | P a g e
 IJRECE VOL. 6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)
U. Mahlab et al., [12] proposed a strategy called connection, it chooses the administration with least
discontinuity aware load adjusting for optical systems that dynamic associations.
upgrade the heap dissemination over the system. The b) Round Robin Method: This strategy constantly pivots a
localization minimizes cost for the new link development. rundown of administration that is appended to it. At the
To measure load imbalance, proposed entropy based metric. point when the virtual server gets a demand, it appoints
the association with the main administration in the
IV. PROPOSED WORK rundown, and after that moves that support to the base of
the list.
In this section, the load balancing and several kinds of c) Least Response Time Method:This technique chooses the
validation protocols are described. administration with the least dynamic associations and
the most minimal normal reaction time.
A. Load Balancing d) Custom Load Method: When utilizing this technique, the
load balancing machine picks an administration that isn’t
Load balancing is necessary constraint for initiative
dealing with any dynamic exchanges. On the off chance
networks so it can provide high availability and scalability.
that the majority of the administrations in the load
Load balancing works as an aware routing protocol software
balancing setup are taking care of dynamic exchanges,
defined network (SDN). The entity that assists the
the application chooses the administration with the least
availability and scalability that leads to obtain the minimal
load.
response time of application. To solve this problem load
balancing techniques method increases the network
A. Various Available Protocol Validation Tools
efficiency [13]. Present framework method of balance the
load isn’t useful aimed at the framework of Software
Defined Network (SDN), where the main station or manager There are many validation tools available in the
of network basics is completed by programs. To manage the market through which the process of validating is
load of server is easy which is connected to the controller of performed. Some of the tools are:
the framework [14]. It also helps the clients from connecting
directly to the backend servers which secure the structure of 1) AWS: Amazon Web Services (AWS) is an entire and
internal network and prevents from attack on the network exhaustive, changing distributed computing stage given
[15]. via Amazon. This will blend IaaS (Infrastructure As A
Service), PaaS (Platform As A Service) and SaaS
(Packaged Software As A Service) advertising. Amazon
Web Services boosted in 2006 via thye internal
organizations of Amazon. It worked for managing and
processing of online marketing activities. Amazon web
services (AWS) is the fundamental organization which
represented the compensation to distribute the
computational models which serves the clients with the
availability of stockpiling and quantity.
a) Computer: The Elastic Compute Cloud (EC2) gives
virtual systems called occasions to ascertain measures.
The elastic compute clouds (EC2) benefit gives access
to the heaps of the present kinds via changing the
measurements and dimensions, customized to the
particular remaining task at hand types and applications,
for example, memory-serious and upgraded figuring
employments. AWS additionally gives an Auto Scaling
device to vigorously scale capacity to keep up
Fig.3. Load Balancing Architecture occurrence wellbeing and execution.
b) Storage: S3 is given by Amazon which stands for the
1) Load Balancing Algorithm and Methods: Load
simple storage services. S3 gives the better stockpiling
balancing utilizes different calculations, called load
for the information, reinforcement and for the better
balancing techniques, to characterize the criteria that
segments. Information Technology (IT) proficient stores
the ADC application uses to choose the administration
information and documents like the simple storage
to which to divert every customer demand. Unlike load
services (S3). It is implemented on 5GB and pails to
balancing systems utilize distinctive criteria.
save the data. The businesses must be organized with
a) Least Connection Method: The default strategy, when a
extra cashes to expand the storage level and to utilize the
virtual server is arranged to utilize the minimum
amazon storage for long term.

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1021 | P a g e
 IJRECE VOL. 6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)
c) Databases: AWS gives oversaw data-bases benefit done the host names via using Ethernet to establish with zero.
by ARDS (Amazon Relational Database Service), that The first interfaces are known as H1-eth0 and the third
incorporates alternatives (Oracle, SQL Server, MySQL, interfaces are known as H2- eth2. The first switch is
MariaDB) and a private superior data-base as namely as named as S1-eth1 and the numbering is start with 1 [16].
AA (Amazon Aurora). Amazon web services offered No
SQL data-bases via using Amazon Dynamo database. V. RESULT AND DISCUSSION
d) Alternarite Services: AWS has the objective for the
industry generation such as SaaS alternatives. It is Mininet is open source software and used to simulate a
beneficial for empowering the virtual video gatherings, SDN. As SDN separates the two planes from each other thus
requests and content established talks over the gadgets. making network devices fully programmable which means
2) Amazon Elastric Compute Cloud (EC2): Amazon EC2 the network will work upto the user need. Mininet provides
gives the essential registering capability mainly in AWS OVS controllers and switches. The primary feature that
cloud that are the amazon web services. The separates SDN devices from traditional network devices is
development of the elastric compute cloud described the the modification of protocols and functions is the points
different equipement for the better construction and to which differentiate SDN devices from traditional ones. It
solve the problems quickly. also supports the OpenFlow protocol that act as an interface
3) Mininet: The term mininet is the emulator which utilized between the planes. The OpenFlow controls the flow of
for assembling the extreme systems on the constrained packets on the instructions of API. Supporting various
terms in the particular PC (Personal Computer) and VM topologies and allow to customize the topologies. A little
(Virtual Machine). It gives the permissions for the data knowledge is required to work with the command line
packets to transmit successfully with the encryption of interface.
data. Additionally, it gives more availability and the
better sense without accessing more knowledge and
A. Features of Mininet
effort. It gives more access for the better utilization,
implementation and the versatility [16].
1) It is inexpensive network for developing
applications of OpenFlow.
2) Enables developers to make parallel work that can
independently done on the same topology.
3) Supports regression tests at the system level.
4) Enables complex topology testing.
5) Includes a CLI to debug or run the network-wide
assessments.
6) Supports random customized topologies, and
contains a simple set of parameterized topologies.
7) Can be used without programming.
8) Provides Python API to create network.

B. Test results: To well plan routing processing of the clients

to the particular servers mainly in the imporved approaches
and there are numerous load balancing.

Load Balancing Strategies:

Fig.4. Emulating Real Networks in Mininet [16]
1) Random: Traffic towards web servers go into random
The mininet is the simplest and most preferable open source
manner. You cannot control or predict about the load
programming which performed the survey on the open flow
balancing ratio. It is not recommendable for most of the
processes and to describe the software defined network
environment. SDN controller runs this random load
(SDN) controllers. It has the tendency to generate SDN
balancing application and diverts the traffic to web-servers
systems which executed the controllers for tests [17].
in random manner.
4) Mininet Topologies (MT): MT contains the large number 2) Weighted Round-Robin: In this type, we can add weight to
of the basic topologies for instance the minimal singles, the web servers. More the weight, better the chance that
reversed and the linear [18]. It is tolerated and proven traffic will be forwarded to that. This type of method is
approach for the terminals hosts and mainly for the used mainly when there are servers with different
several kinds of switches for the upcoming mininets. capacities in the network all we want is to transfer traffic
The switches are initialized from S1 to SN. The hosts are to servers with larger capacity. Every server gotten the
named as H1 and HN. The interfaces are combined with requests from the clients which are fixed.

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1022 | P a g e
 IJRECE VOL. 6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)
3) Round-Robin: This method makes one-by-one shift when made by the round robin using a circular queue. The request
traffic from clients go towards web servers. This method to the server with low load is based on the load based policy
can be used when all the servers running web application while load is the frequency of the request that is pending.
have same set of capacity [RAM, Processing Power, and
HDD]. Using Mininet , a custom topology is created. After that the
server is to set up. HTTP server is made with 80 port
C. VPS (Virtual Private Server) number of the node one and two. The servers should have
their individual IP addresses shown in the figure 13.
VPS is server that exists in on a server with respect to the
virtual machines. With a VPS, we have our own operating Now, traffic is sent to the server by curl command used by
system installation. the http clients shown in the figure 13. From this a webpage
is got from the server’s IP. Round robin algorithm helps the
client to get server in a circular manner. Figure 14 shows the
different nodes and the direction of the traffic to the servers.

To run ip_loadbalancer we need pass arguments../pox.py

misc.ip_loadbalancer –ip=<Service IP> –servers=<Server1
IP>,<Server2 IP>,… [–dpid=<dpid>]

Fig. 5. VPS IP address assigned to the instance by AWS

Now we will create topology with one controller, one switch

and 6 hosts in Fig.6.

Fig. 8. Directing Traffic

ip_load_balancer is a TCP load balancer which takes a

service ip and the list of internet protocol addresses of the
servers. Client requests are directed to various pre-defined
http servers while the round robin chooses the server.
Fig. 6. Mininet Topology Creation Traffic will be provided to that this type of method is used
namely when those are servers wise different capacities in
We have 2 two hosts as “h1”, “h2” that we run as server and
the network and all we want is to tend more traffic to server
from another “h3”, “h4”, “h5”, and “h6” we request HTTP
with larger capacity.
service by using curl command.
Round Robin method makes one-by-one shift when traffic
form clients go towards web-servers. These methods can be
used when the entire server running web-application same
have set of capacity like as a user, processing power and
HDD.

D. Result:

Fig. 7 Server Setup on h1, h2 and Traffic from Client Node h3, h4

The rewriting of the destination IP address is done by the

load balancer. The decision of where to send a request is

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1023 | P a g e
 IJRECE VOL. 6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)
we ready for SDN? Implementation challenges for
software-defined networks. IEEE Communications
Magazine, 51(7), 36-43.
[3] Xie, J., Guo, D., Hu, Z., Qu, T., & Lv, P. (2015).
Control plane of software defined networks: A
survey. Computer communications, 67, 1-10.J.
[4] Zehua, G., Mu, S., Yang, X., Zhemin, D., & Luo, W.
(2014). Improving the performance of load balancing
in software-defined networks through load variance-
based synchronization [J]. Computer Networks, 68,
95-109.
[5] Nunes, B. A. A., Mendonca, M., Nguyen, X. N.,
Obraczka, K., & Turletti, T. (2014). A survey of
Fig 9. Comparison software-defined networking: Past, present, and
future of programmable networks. IEEE
The above figure shown that the comparison between Communications Surveys & Tutorials, 16(3), 1617-
proposed and existing work. Random algorithm best 1634.
perofrmnace analysis based on one at a time transfer in [6] Xia, W., Wen, Y., Foh, C. H., Niyato, D., & Xie, H.
minimum time and heavy or weighted algorithm to (2015). A survey on software-defined
calculated the weight accordingly data transferring and networking. IEEE Communications Surveys &
Round Robin Algorithm defined used the same link Tutorials, 17(1), 27-51..
activated in equal memory used. In this comparison define [7] Ramos, F. M., Kreutz, D., & Verissimo, P. (2015).
the load balancing the customer requirement. Software-defined networks: On the road to the
softwarization of networking. Cutter IT journal.
VI. CONCLUSION AND FUTURE SCOPE [8] Kaur, S., & Singh, J. (2016). Implementation of
Server Load Balancing in Software Defined
SDN has changed the network industry and the plethora of
Networking. In Information Systems Design and
benefits that it provides makes it a super success in Data
Intelligent Applications (pp. 147-157). Springer,
Center and service provider industry. Load Balancing is a
New Delhi.
need in Data Centers where multiple servers are running
[9] Lin, S. C., Wang, P., & Luo, M. (2016). Control
same application for better redundancy and processing. Pox
traffic balancing in software defined
comes by default in mininet which can be used the
networks. Computer Networks, 106, 260-271.
controller and with mininet based switches (OVS). The
[10] Perepelkin, D., & Byshov, V. (2017, April). Visual
basic protocol is the openflow and utilized among controller
design environment of dynamic load balancing in
besides infrastructure layer. With load balancing, all the
software defined networks. In Radioelektronika
servers in the server zone works parallel and which
(RADIOELEKTRONIKA), 2017 27th International
automatically makes our data center network much more
Conference (pp. 1-4). IEEE.
efficient than before. AWS is amazon’s cloud which runs
[11] Attarha, S., Hosseiny, K. H., Mirjalily, G., &
application, host virtual machine and stores data. POX
Mizanian, K. (2017, May). A load balanced
controller is s implemented in the AWS EC2 cloud over a
congestion aware routing mechanism for Software
Linux machine and we made a cloud based controller which
Defined Networks. In Electrical Engineering (ICEE),
is acting as a load balancer for other servers. Lots of
2017 Iranian Conference on (pp. 2206-2210).
research work is still on in the SDN and as a result lots of
IEEE.S.
new applications and technologies are coming up as a part [12] Mahlab, U., Omiyi, P. E., Hundert, H., Wolbrum, Y.,
of research like 5G, SD-WAN etc. I want to take my
Elimelech, O., Aharon, I., & Zarakovsky, S. (2017,
research further in SDN based LBs clustered in the cloud,
July). Entropy-based load-balancing for software-
which can be used in application and networks
defined elastic optical networks. In Transparent
infrastructure forming the cloud and makes the cloud based
Optical Networks (ICTON), 2017 19th International
infrastructure much more efficient than using physical
Conference on (pp. 1-4). IEEE.U.
traditional network infrastructure. [13] Kumari, P., & Thakur, D. (2017). Load Balancing in
Software Defined Network.
REFERENCES
[14] Raghul, S., Subashri, T., & Vimal, K. R. (2017,
[1] Sood, M. (2015). SDN and Mininet: Some Basic March). Literature survey on traffic-based server
Concepts. International Journal of Advanced load balancing using SDN and open flow. In Signal
Networking and Applications, 7(2), 2690. Processing, Communication and Networking
[2] Sezer, S., Scott-Hayward, S., Chouhan, P. K., Fraser, (ICSCN), 2017 Fourth International Conference
B., Lake, D., Finnegan, J., ... & Rao, N. (2013). Are on(pp. 1-6). IEEE.

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1024 | P a g e
 IJRECE VOL. 6 ISSUE 4 ( OCTOBER- DECEMBER 2018) ISSN: 2393-9028 (PRINT) | ISSN: 2348-2281 (ONLINE)
[15] Pathak, P., & Mahajan, E. K. (2015). A Review on
Load Balancing In Cloud Computing. International
Journal Of Engineering And Computer
Science, 4(06).
[16] Kaur, K., Singh, J., & Ghumman, N. S. (2014,
February). Mininet as software defined networking
testing platform. In International Conference on
Communication, Computing & Systems
(ICCCS) (pp. 139-42).
[17] Lantz, B., Heller, B., & McKeown, N. (2010,
October). A network in a laptop: rapid prototyping
for software-defined networks. In Proceedings of the
9th ACM SIGCOMM Workshop on Hot Topics in
Networks (p. 19). ACM.
[18] Sharma, K. K., & Sood, M. (2014). Mininet as a
container based emulator for software defined
networks. International Journal of Advanced
Research in Computer Science and Software
Engineering, 4(12).

INTERNATIONAL JOURNAL OF RESEARCH IN ELECTRONICS AND COMPUTER ENGINEERING

A UNIT OF I2OR 1025 | P a g e