Design & Implementation of API Based Software

Defined Networking Architecture

Abstract—Software Defined Networking known as SDN ap-

peared to provide automation across any network system. The
SDN architecture is designed to build a flexible and manageable
network comparing to traditional networking system. Besides,
the SDN plays a vital role in network virtualization, cloud
computing, mobility and programmability to make an efficient
network system whereas the traditional networking system has
some limitations. In this paper, an SDN architecture has been
proposed with API integration using python library on Linux
based OpenDaylight framework. The comparative results show
that the performance of the proposed SDN architecture is
much better than traditional network architecture considering
data flow, automated IP address generation, packet transmission
consistency.
Index Terms—IoT, Software Defined Networking, IP address,
API, Python, Linux, OpenDaylight, Data flow.
Fig. 1. Traditional Networking Architecture.
I. I NTRODUCTION
Before this era, basic packet-switching terminology was the
most modern networking system. It was based on data which
are called packets and are routed so that there is destination for
each packet through the network. On this system, data packets
are included on device where switches are also included on
device i.e. the networking system is non–virtual. To control the
routing, user has to depend on switches manually (included
with hardware) which is not so smooth. This traditional net-
work is used all over the world because of its popularity among
Fig. 2. The pace difference between Traditional Networking and SDN
other networking system. There are three phases of traditional
networking system: Hub Based LANs, Switch Bases LANs
and routed networks where LANs are flat, shared-media which process in more likely most organizations. The IT team should
is quite simple in architectural design. Fig. 1 is showing an be notified of the problem first and then they contact the
architecture of traditional networking. network team and eventually the network team fix the problem.
Now SDN is not exactly a product that people need to The mentioned steps take times and are evidently hectic in a
start using by purchasing any package of it. SDN is rather sense.
an architecture or a model that needs implementation. The Automated APIs are faster from that perspective of opera-
central idea of Software Defined Networking architecture is tion. SDN architecture has addressed the need for centralized
to “decouple the hardware from the software”. When it has control or management and monitoring of network devices on
been dealt with the decoupling of hardware from the software, both the physical and virtual levels. It also roots for intro-
there are two basic concepts that raise concern. The control ducing the ability to override the certain existing forwarding
plane which determines where to send the traffic or packet and algorithms of the network devices to fit the need of the indi-
then there is the data plane that implements the decision and vidual networks. The new architecture also allows the external
performs the forwarding of the traffic. application-based network operation influencing abilities along
The SDN architecture is also faster than the traditional net- with rapidity and scalability in deployment of network for its
working architecture. The traditional existing technology for life cycle managing purposes [1]. Fig. 2 is showing the pace
networking is limited to certain network–oriented knowledge- difference between the traditional Networking and the SDN.
based personnel which made the process quite slow, time
consuming and rather dependent on humans. When some II. BACKGROUND S TUDY
problem in the system is intercepted, the need would arise This section discusses the basic descriptions and concepts of
to reach the network team which may be through a certain SDN as well as the basic knowledge of Python, OpenDaylight,
Openflow, Mininet. 4) Openness of the Architecture
As mentioned earlier, the whole point of the SDN archi-
A. Software Defined Networking
tecture is to keep on researching and manipulating the
There are four fundamental characteristics for SDN archi- control plane according to experimentation. This gives
tecture [2] which are mentioned below: the architecture openness for various researches and
1) Separation of the Planes development. Developers work on carving the network
The most basic characteristics of the SDN is the sep- to their own needs and deploy it in accordance to their
aration of the two planes (controller & data packets). own fitting deems. The previously retaining bindings and
The tables and logics for determining the flow of the limitations of vendor specific products and the need for
incoming packets based on IP, MAC, VLAN ID etc., having to be dependent on vendors for knowledge and
that is, the forwarding functionality has been entirely training has been eliminated. This adds a new horizon
kept with the data plane. It may forward, replicate, of potentials for development and new promises for the
drop or consume the inbound packets as it seems fit future of networking.
and the way the packets dispense. As for the control Factors those are concerned with implementation of the
plane; the protocols, the logic and algorithms of the proposed system of SDN are API & Python programming
forwarding motion and program has been separated into language with necessary libraries.
an abstraction in order to centralize control.
2) Device Simplicity & Centralized Control B. Python
In traditional networking architecture, the network de- As the proposed SDN architecture deals with API
vice would remain complicated with thousands of lines and automation, Python programming is needed to im-
of complicated programming and the control plane and plement the system [4]. Many complex problem can be
data plane persisting in the device itself and having to solved with different python packages. In this research,
run autonomously. In addition to that, developing them ”miniedit.py” of Python3 version is used on Linux plat-
for individual purposes was more of an impractical job. form. To run MiniEdit, the mentioned Command is executed:
Plus, the control then required specific language skills sudo /mininet/examples/miniedit.py
and proper knowledge in the department. The separation
added to the device simplification. Now, the device was C. MiniNet
only a simple forwarding device and the whole control MiniNet is the most vital tool to implement SDN architec-
was shifted out. The machine was put off from the ture. It works like a simulator to control the automation and
excess pressure and the control was now centralized connection among the routers, switches, devices etc. MiniNet
and not device specific. The software based centralized is package or library of Python for which proper environment
control made sure that the devices could be accessed of Python programming is needed. It also helps to build up
from a central control system and added flexibility to the API related to the proposed system [5].
the system.
3) Virtualization of Network & Automation D. OpenFlow
The basic structure of SDN is based on three abstrac- Proper data flow and communication is the first concern
tions, the forwarding, distributed state and the configu- of SDN architecture. OpenFlow is such a standard of SDN
ration processes. The natural evolution of SDN in the architecture which provides proper networking protocols for
network control world is almost similar to the evolu- uninterrupted data flow in the proposed system [6]. Besides,
tion of language from machine language to the high the maintenance of communication among several planes of
level languages we deal with today [3]. The distributed SDN architecture deals with OpenFlow along with the API
state defines the global view for the programmer to system. It is also a library of Python programming language.
understand the system network whereas the forwarding
abstraction gives a proper understanding of the behavior E. OpenDaylight
of the network even without the knowledge of ven- Controller is the brain of SDN architecture. It controls the
dor specific hardware. At the end is the configuration total networking system. OpenDaylight is such a library or
abstraction allows to understand the primary goal of package of Python programming language which is developed
the system rather than being lost in the middle. This as the controller for SDN architecture. It is needed to imple-
configuration is what gives the most basic level of ment OpenDaylight to the proposed API system as controller
virtualization of the network. As it can be seen, the [7]. Therefore, OpenDaylight has been used as controller
software and the API bring about the automation process protocol to make the system smoother.
to the system along with the programmability. Thus, the
problems of network virtualization and automation for III. L ITERATURE R EVIEW
reconfiguration and other short comings like having to The scopes of IoT are increasing day by day significantly.
know the basics of vendor specific hardware systems So, it is very much important to cope up with the developments
have been eradicated. of IoT in the mean time [8]. Integration of SDN causes the
 Again the planes are separated in SDN which provides better
scalability. It needs only a physical controller and the API i.e.
the software does the rest [13] where traditional networking
system is dependent on hubs, switches, routers etc.
A survey has been presented to show the network traffic
controlling, management and security placement to determine
the efficiency of SDN architecture [14]. Due to the decoupled
control plane & data plane, network traffic is filtered in
each plane so that SDN architecture can provide better traffic
management and also the security [15]. Besides, the data
packets flow much accurately to provide better user experience
of networking system [16]. But, to ensure this kind of system
Fig. 3. The Layers of SDN Architecture Under Scrutinized View. with better security, an efficient SDN framework is needed
[17]. To make the SDN more energy efficient, a survey has
been done which deals with three factors named traffic aware
betterment in IoT as the networking system becomes virtual. model, end-host aware model & rule placement [18]. The
Besides, SDN architecture is simplified so that the networking factors provide objective functions, required parameters and
system can be controlled easily with less physical devices & detailed information for energy efficiency of SDN architecture.
technologies. The security system should also be promised to As a future technology, network forensic of SDN has been
keep the network sustainable where SDN plays better role than determined to measure its importance [19]. This forensic
traditional networking architecture. Overall, SDN provides provides an expansion of SDN architecture to improve the
innovation to fulfill the demands of IoT along with wireless networking system along with further scope of research. On
communication system. the latest survey, SDN architecture has been introduced to
SDN reduces the number of link failure in network. As the facilitate the with network virtualization for block-chain which
control plane and data plane are separated in software-defined is a significant part of IoT [20]. Besides, SDN architecture is
networking system, the transportation of links become more introduced at 5G mobile networking architecture [21]. Thus
efficient than traditional networking system. This thing is done the SDN architecture has been opened for future applications.
by applying appropriate machine learning (ML) algorithm
[9]. Therefore, the losses of link i.e. link failure problem in IV. P ROPOSED M ETHODOLOGY OF SDN A RCHITECTURE
network is reduced. It is one of the most significant purposes The idea of decoupling the two planes is that the data plane
to go with SDN architecture. or the execution plane will be left along with the hardware
Software-defined networking system ensures Quality of such as the network switch or router whereas the control plane
Services (QoS) while transporting data [10]. If the QoS of or determiner plane will be moved into a software layer. This
data flow becomes vulnerable, the whole networking system would enable the control of network by a centralized software
is hampered. So, the vulnerability of data flow is a significant application and the hardware would not have to be touched.
drawback of IoT. QoS guaranteed path selection algorithm has This facilitates the virtualization of network which was not
been proposed to provide better QoS in IoT while implement- possible previously using the traditional networking system
ing SDN. The algorithm calculates data flow and data loss architecture. Transformation from traditional networking ar-
so that the better QoS can be fulfilled where the accuracy of chitecture to software defined networking architecture has been
QoS is 99.73%. Again, the separation of control plane & data visualized on Fig. 4.
plane helps to provide uninterrupted network service mostly
[11]. Not only the control plane and data plane, but also the TABLE I
application layer i.e. management plane helps to provide better H ARDWARE , S OFTWARE & M AIN T ECHNOLOGIES FOR P ROPOSED SDN
A RCHITECTURE
QoS as the planes are separated. So, better QoS is ensured
in each plane for stable networking system. Therefore, SDN Components Requirements
architecture becomes popular in networking nowadays. Basic Hardware Laptop/Desktop, Router, Internet Connection
RAM 4 GB (min)
Cloud service deals with three type of sevices called Processor Intel Core i3, i5 or i7
Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS) & Memory 240/256/512 GB
Infrastructure-as-a-Service (IaaS). Networking system is the OS Ubuntu-16.04 LTS or above
main part of IaaS where SDN offers better structural architec- Python Version 2.7 or higher
Packages MiniNet [18], OpenFlow [19], OpenDaylight [20]
ture to fulfill the demand of IaaS [12]. SDN provides resource
allocation to service holders efficiently so that the IaaS doesn’t
interrupt quite often. To fulfill the proposed implementation process of SDN
SDN holds better system of networking in stead of tradi- architecture, it is needed to apply some techniques with
tional networking. As SDN is not device dependent system efficient hardware, software & proper technology. The crucial
it is quite comfortable to operate over traditional networking. hardware, software & important technology which are needed
 The generated controller APIs are very easy to control the
networking system. After running first API “my mn.py”, if
the terminal is pinged, it can be seen that the data flow from
host to host of a virtual network. Reversely, the data forward-
flow and the data backward-flow can be determined through
the API. Along with the data transmission rates can be shown
when running the API.
Data-flow or transmission control from one host or network
to another host or network in SDN is based on the following
formula [18].
Fig. 4. Transformation from Traditional Networking Architecture to j1 j1 3
F I(T ) : (SDN1 , C1 ) −−→ (SDN2 , C2 ) −−→ (SDN3 , C3 )
Software Defined Networking Architecture.
V. E XPERIMENTAL R ESULT AND A NALYSIS
to implement the proposed SDN architecture have been de- Two switches and two or more host has been designed for
scribed in Table I. the proposed system. After running the scripts “my mn.py”
The concern is to chalk out the performance of a SDN and pinging from Host1 to Host2 of same switch the results
controller. So, the basic thing is to meet the criteria of better and performance are shown as Table II.
performance in this system. The proposed system can be
described as follows: TABLE II
P ING R ESULT & P ERFORMANCE FROM H OST 1 TO H OST 2
A. Generation of API
64 bytes from 10.0.0.2: icmp seq=1 ttl=64 time=66.0 ms
The most important task in the system is to generate a 64 bytes from 10.0.0.2: icmp seq=2 ttl=64 time=0.482 ms
64 bytes from 10.0.0.2: icmp seq=3 ttl=64 time=0.142 ms
controller API by which the control of networking system 64 bytes from 10.0.0.2: icmp seq=4 ttl=64 time=0.064 ms
with SDN can be checked. For this, the mininet technology 64 bytes from 10.0.0.2: icmp seq=5 ttl=64 time=0.131 ms
with openflow is needed. First of all, there needs to run the 64 bytes from 10.0.0.2: icmp seq=6 ttl=64 time=0.164 ms
64 bytes from 10.0.0.2: icmp seq=7 ttl=64 time=0.077 ms
mininet by the terminal of linux operating system i.e. ubuntu 64 bytes from 10.0.0.2: icmp seq=8 ttl=64 time=0.077 ms
16.04 LTS. After that, there needs to go on mininet examples 64 bytes from 10.0.0.2: icmp seq=9 ttl=64 time=0.075 ms
directory. There is a python scripts named “miniedit.py” into 64 bytes from 10.0.0.2: icmp seq=10 ttl=64 time=0.148 ms
the examples directory of mininet. That scripts is needed to 10 packets transmitted, 10 received, 0% packet loss, total time
9199ms Rate,
run for generating APIs. The APIs starts like a window. The Minimum = 0.064 ms
window looks like a cisco packet tracer. It is needed to design Average = 6.740
an architecture of SDN by running “miniedit.py” scripts and Maximum = 66.048
Mdev = 19.769 ms
opening a window. Several API of various design can be gen-
erated through the “miniedit.py” scripts. Two API controller
for this SDN system have been designed and generated. One The results and performance from Host2 to Host1 has been
is for two host-one switches and the other is for more than two shown on Table III.
host-more than two switches. Those two APIs on the mininet
examples directory with two names named “my mn.py” and TABLE III
P ING R ESULT FROM H OST 2 TO H OST 1
“my mn 2.py” respectively have been exported and stored.
In “my mn.py” API, there is a controller which controls a 64 bytes from 10.0.0.1: icmp seq=1 ttl=64 time=7.46 ms
switch. Again, the switch have two hosts. Thus, the network 64 bytes from 10.0.0.1: icmp seq=2 ttl=64 time=2.25 ms
64 bytes from 10.0.0.1: icmp seq=3 ttl=64 time=0.603 ms
flow in the hosts can easily be determined with the controller. 64 bytes from 10.0.0.1: icmp seq=4 ttl=64 time=0.075 ms
In “my mn 2.py” API, there is a controller which controls 64 bytes from 10.0.0.1: icmp seq=5 ttl=64 time=0.141 ms
several switches. The switches have included several host also. 64 bytes from 10.0.0.1: icmp seq=6 ttl=64 time=0.080 ms
64 bytes from 10.0.0.1: icmp seq=7 ttl=64 time=0.067 ms
There used seven hosts for this API controller. And also, it 64 bytes from 10.0.0.1: icmp seq=8 ttl=64 time=0.297 ms
can be determined that the network flow is controlled by the 64 bytes from 10.0.0.1: icmp seq=9 ttl=64 time=0.139 ms
controller API. 64 bytes from 10.0.0.1: icmp seq=10 ttl=64 time=0.132 ms
10 packets transmitted, 10 received, 0% packet loss, time
9149ms Rate,
B. Roles of API Minimum = 0.067 ms
In traditional networking system there is no controller to Average = 1.125 ms
Maximum = 7.467 ms
control the network virtually. So, in this criterion the controller Mdev = 2.206 ms
API plays a great role to control the networking system. Ear-
lier, there needs to be controlled the network system manually
including the hardware part. But, in SDN, the controller API From the net result, it has seen that Host1, Host2 and Host3
becomes separated from the hardware. are along with the Switch1, Host4 and Host5 are along with
 In the proposed SDN system, there is no need for con-
figuring the system manually. The difficulty in traditional
networking system is to provide the IP addresses to the hosts
and routing system manually with calculation. Let, there is
a two-host one-switch networking protocol in cisco packet
tracer. While providing IP addresses to the hosts, it must be
calculated manually. If anything is missed while calculating
the IP addresses the networking system would have become
null and void. In Fig. 5 it has been shown that there is a miss
calculation of IP addressing and that’s the reason why there
is not connection among the hosts.

VII. C ONCLUSION

As the software defined network i.e. SDN is the modern

networking system in it has been tried to implement for the
Fig. 5. Cisco Networking System (Wrong Calculation of IP Addresses and betterment of SDN. Actually, the concept of API is the start
Connection Loss). of finding the best controller for SDN. The API generation
has been tried to modify. Also, the test and comparison of
the result has been analyzed carefully for further development
the Switch2. Now, if it is tried to ping between the different
of SDN architecture. And the robustness of popularizing the
host of the switch, the “my mn 2.py” scripts is needed to run.
SDN architecture can be more efficient as application in real
Therefore, The results and performance from Host1 to Host4
life, because 4IR is the key factor nowadays.
or vice-versa has been shown on Table IV.

TABLE IV R EFERENCES
P ING R ESULT FROM H OST 1 TO H OST 4 OR V ICE - VERSA
[1] Maged, A., 2015. Introduction to Software Defined Networking, Dubai:
From 10.0.0.1 icmp seq=1 Destination Host Unreachable Middle East Network Operational Group.
From 10.0.0.1 icmp seq=2 Destination Host Unreachable [2] Goransson, P. & Black, C., 2014. Chapter 4 – How SDN Works. In:
From 10.0.0.1 icmp seq=3 Destination Host Unreachable Software Defined Networks - A Comprehensive Approach. s.l.: Elsevier
From 10.0.0.1 icmp seq=4 Destination Host Unreachable Inc., pp. 59-69.
From 10.0.0.1 icmp seq=5 Destination Host Unreachable [3] Shenker, S., Casado, M., Koponen, T. and McKeown, N., 2011. The
From 10.0.0.1 icmp seq=6 Destination Host Unreachable future of networking, and the past of protocols. Open Networking
From 10.0.0.1 icmp seq=7 Destination Host Unreachable Summit, 20, pp.1-30.
[4] Noman, H.M. and Jasim, M.N., 2020, July. Pox controller and open flow
performance evaluation in software defined networks (sdn) using mininet
emulator. In IOP conference series: materials science and engineering
VI. C OMPARISON (Vol. 881, No. 1, p. 012102). IOP Publishing.
[5] M. Hasan, H. Dahshan, E. Abdelwanees and A. Elmoghazy, ”SDN
From the results of the proposed API based SDN, it is Mininet Emulator Benchmarking and Result Analysis,” 2020 2nd Novel
clearly seen that packet or data loss can be prevented which Intelligent and Leading Emerging Sciences Conference (NILES), Giza,
is not always possible in traditional networking system. The Egypt, 2020, pp. 355-360, doi: 10.1109/NILES50944.2020.9257913.
most important thing in networking is the flow of data packets. [6] Wazirali, R.; Ahmad, R.; Alhiyari, S. SDN-OpenFlow Topology Dis-
covery: An Overview of Performance Issues. Appl. Sci. 2021, 11, 6999.
Traditional cisco packet tracer sometimes creates loss in the https://doi.org/10.3390/app11156999
data packets. For this, there arises problem on networking [7] L. Mamushiane, A. Lysko and S. Dlamini, ”A comparative evalua-
system. The result of flowing data packets for cisco has shown tion of the performance of popular SDN controllers,” 2018 Wireless
Days (WD), Dubai, United Arab Emirates, 2018, pp. 54-59, doi:
in Table V. 10.1109/WD.2018.8361694.
[8] A. H. Mohammed, R. M. KHALEEFAH, M. k. Hussein and I. Amjad
TABLE V Abdulateef, ”A Review Software Defined Networking for Internet of
I NCONSISTENT DATA F LOW ON T RADITIONAL N ETWORKING S YSTEM Things,” 2020 International Congress on Human-Computer Interaction,
Optimization and Robotic Applications (HORA), Ankara, Turkey, 2020,
64 bytes from 10.0.0.2: icmp seq=1 ttl=36 time=78.0 ms pp. 1-8, doi: 10.1109/HORA49412.2020.9152862.
64 bytes from 10.0.0.2: icmp seq=2 ttl=40 time=0.512 ms [9] Ali, Jehad; Lee, Gyu-min; Roh, Byeong-hee; Ryu, Dong Kuk;
64 bytes from 10.0.0.2: icmp seq=3 ttl=64 time=0.693 ms Park, Gyudong (2020). Software-Defined Networking Approaches
64 bytes from 10.0.0.2: icmp seq=4 ttl=00 time=0.492 ms for Link Failure Recovery: A Survey. Sustainability, 12(10), 4255,
64 bytes from 10.0.0.2: icmp seq=5 ttl=54 time=0.089 ms doi:10.3390/su12104255.
64 bytes from 10.0.0.2: icmp seq=6 ttl=64 time=undefined [10] Sun, Weifeng; Wang, Zun; Zhang, Guanghao (2020). A QoS-guaranteed
64 bytes from 10.0.0.2: icmp seq=7 ttl=64 time=undefined intelligent routing mechanism in software-defined networks. Computer
64 bytes from 10.0.0.2: icmp seq=8 ttl=64 time=undefined Networks, 107709–. doi: 10.1016/j.comnet.2020.107709.
64 bytes from 10.0.0.2: icmp seq=9 ttl=64 time=undefined [11] Keshari, S.K., Kansal, V. & Kumar, S. A Systematic Review of Quality
64 bytes from 10.0.0.2: icmp seq=10 ttl=64 time=undefined of Services (QoS) in Software Defined Networking (SDN). Wireless
Pers Commun 116, 2593–2614 (2021). https://doi.org/10.1007/s11277-
020-07812-2.
[12] 06. Arwa Mohamed, Mosab Hamdan, Suleman Khan, Ahmed Abde- defined network for the Internet of Things. Trans Emerging Tel Tech.
laziz, Sharief F. Babiker, Muhammad Imran, M.N. Marsono, Software- 2022; 33( 1):e4391. doi:10.1002/ett.4391.
defined networks for resource allocation in cloud computing: A survey, [21] Abdul Ghaffar, A., Mahmoud, A., Sheltami, T. et al. A Survey on
Computer Networks, Volume 195, 2021, 108151, ISSN 1389-1286, Software-Defined Networking-Based 5G Mobile Core Architectures.
https://doi.org/10.1016/j.comnet.2021.108151. Arab J Sci Eng 48, 2313–2330 (2023). https://doi.org/10.1007/s13369-
[13] Haji, Saad H. and Zeebaree, Subhi R. M. and Saeed, Rezgar Hasan 022-07183-9.
and Ameen, Siddeeq Y. and Shukur, Hanan M. and Omar, Naaman [22] Olena, T., Saad, I. and Salim, M.J., 2015. Mathematical methods for
and Sadeeq, Mohammed A. M. and Ageed, Zainab Salih and Ibrahim, analysis of software-defined networks. Inform. Technol. Knowl., 9(2),
Ibrahim Mahmood and Yasin, Hajar Maseeh (2021) Comparison of Soft- pp.111-123.
ware Defined Networking with Traditional Networking. Asian Journal [23] ”Mininet”, http://mininet.org, last accessed on: August 25, 2023.
of Research in Computer Science, 9 (2). pp. 1-18. ISSN 2581-8260. [24] ”Openflow”, https://openflow.inc, last accessed on: August 25, 2023.
[14] Gupta, N.; Maashi, M.S.; Tanwar, S.; Badotra, S.; Aljebreen, [25] ”Opendaylight”, ”https://www.opendaylight.org, last accessed on: Au-
M.; Bharany, S. A Comparative Study of Software Defined Net- gust 25, 2023.
working Controllers Using Mininet. Electronics 2022, 11, 2715.
https://doi.org/10.3390/electronics11172715.
[15] Hussain, M.; Shah, N.; Amin, R.; Alshamrani, S.S.; Alotaibi, A.; Raza,
S.M. Software-Defined Networking: Categories, Analysis, and Future
Directions. Sensors 2022, 22, 5551. https://doi.org/10.3390/s22155551.
[16] Mattia Fogli, Carlo Giannelli, Cesare Stefanelli, Software-
Defined Networking in wireless ad hoc scenarios: Objectives
and control architectures, Journal of Network and Computer
Applications, Volume 203, 2022, 103387, ISSN 1084-8045,
https://doi.org/10.1016/j.jnca.2022.103387.
[17] Madhukrishna Priyadarsini, Padmalochan Bera, Software defined net-
working architecture, traffic management, security, and placement: A
survey, Computer Networks, Volume 192, 2021, 108047, ISSN 1389-
1286, https://doi.org/10.1016/j.comnet.2021.108047.
[18] Rout, S., Sahoo, K.S., Patra, S.S. et al. Energy Efficiency in Software
Defined Networking: A Survey. SN COMPUT. SCI. 2, 308 (2021).
https://doi.org/10.1007/s42979-021-00659-9.
[19] Waseem, Q.; Alshamrani, S.S.; Nisar, K.; Wan Din, W.I.S.; Alghamdi,
A.S. Future Technology: Software-Defined Network (SDN) Forensic.
Symmetry 2021, 13, 767. https://doi.org/10.3390/sym13050767.
[20] Babiker Mohamed, M, Matthew Alofe, O, Ajmal Azad, M, Singh Lallie,
H, Fatema, K, Sharif, T. A comprehensive survey on secure software-