OPTIMIZING CONGESTION CONTROL BY USING DEVICES AUTHENTICATION IN SOFTWARE-DEFINED NETWORKS

International Journal of Network Security & Its Applications (IJNSA) Vol.15, No.1, January 2023
DOI: 10.5121/ijnsa.2023.15103 39
OPTIMIZING CONGESTION CONTROL BY USING
DEVICES AUTHENTICATION IN
SOFTWARE-DEFINED NETWORKS
Tamer Barakat, Hanan Eljawhri, Mohamed Merzban and Mahmoud Elbayoumi
Department of Electrical Engineering, Fayoum University, Egypt
ABSTRACT
The Internet and local networks (LAN) are essential in all organizations and aspects of our lives. These
networks' performance should be at high speeds to perform efficiently. This thesis suggests several motions
to improve performance. The first is a software-defined network (SDN) distinguished from traditional
networks by its ability to program traffic, agility, and support for applications that need big data and
virtualization. The second is to control congestion by rerouting traffic to the shortest path. Finally, to modify
the above, device authentication reduces congestion and improves network performance.
KEYWORDS
Congestion Control, Devices Authentication, Rerouting, Shortest Path, Software-Defined Networks
1. INTRODUCTION
Every enterprise should consider the advantages and disadvantages of different types of networks
for creating its networks. Consumer requirements for performance and flexibility, the demand for
modern networks helped in SDN's ascent, and the increasing use of SDN and other virtualization
systems. Traditional networks can not keep up with the ever-growing needs of users in the modern
workplace. Users who want to scale their network infrastructure with as little downtime as possible
are turning to SDN. It consists of the control plane and the data plane. It makes networks
programmable and manageable. The benefits of the split are flexibility and cost savings in both
capital and operational expenses.
Because of its many advantages over traditional networks as improved automation-based
reliability, more effective network management, cost savings, and quicker scaling, it is software-
based; but traditional networking is primarily hardware-based. Whereas SDN is software-based, it
is more flexible and adaptable to users to manage resources virtually via the control plane [1].
The SDN has an application plane, a control plane, and a data plane. The SDN application layer is
a set of programs; responsible for applications or services on a network and describes the service
performance of network resources programmatically. These programs connect with the controller
(control plane) through the northbound interface (NBI), and the controller modifies the activities
of network resources; presented to it through the application layer programs for internal decision-
making [2].
In SDN, the control plane (controller) works as a network operating system or the brain of the
network because of its role in managing and configuring the devices by installing appropriate flow
tables that contain packet routing rules on the network devices (switches, routers) at the data layer

40
out of the southbound interface (OpenFlow), which allows the SDN controller to add, modify, and
delete flows in the flow table of an OpenFlow switches (data plane) and connects the controller to
an OpenFlow switches to manage devices, receive packets from the OpenFlow switches, and send
packets via the controller [3].
SDN controllers have many types; NOX was the first Open Flow compatible SDN controller. It
uses C++. POX uses Python; it only supports OpenFlow version 1.0 and does not operate in a
distributed way. Ryu is another Python-based SDN controller. In contrast to POX, it supports the
OpenFlow protocol in multiple versions: 1.0, 1.2, 1.3, and 1.4. Additionally, it bolsters other
protocols like Netconf and OFconfig. It has the benefit of being able to operate in a dispersed way
[4].
The data plane consists of network hardware such as switches, routers, and packet routing tables.
The network device that accepts packets through its ports can carry out network operations since
packets can be forwarded or destroyed or have their header changed for a particular action after
being received on ports. According to the packets sent by the switch upon a switch request (table-
miss), the controller modifies the forwarding table for each switch request. It then delivers a
response via a packet-out or flowmod message and keeps track of all application requests [5].
Dynamic and unstable network traffic is present. Network performance will drastically decline if
a subnet has excessive data packets. The main reason for packet loss is network congestion, which
happens when the network load exceeds the network capacity. This is why there is an urgent need
for congestion control algorithms. Congestion control is the process of developing a network
control policy based on the status and characteristics of the network [6].
The algorithm of congestion control has three main categories: loss-based algorithms, are define
congestion when buffers are filling, then packets are dropped; delay-based algorithms, which rely
on RTT measurements and identify congestion by an increase in RTT, which indicates buffering;
and hybrid algorithms, which combine elements of the first two categories [7].
Loss-based algorithms were the initial congestion control algorithms, with TCP Reno being the
first to be widely adopted, then the updating of loss-based algorithms, such as NewReno. But,
Packet loss occurs when several flows with different RTT values compete on the same network;
the flows with lower RTT values consume more bandwidth than the others, and a user with longer
RTT flows will not be able to receive sufficient bandwidth. So these modifications could not
resolve the RTT-fairness problems [8], BIC (Binary Increase Congestion Control), and Hybla as
solutions to this problem. The Slow Start and Congestion Avoidance phases of the NewReno
update by Hybla and made somewhat independent of RTT. However, the attained RTT fairness
resulted in more extreme behavior from flows with higher RTTs. BIC's key objective was to
employ a binary search method to receive close to the ideal congestion window size. Later
analyses, meanwhile, revealed that BIC can still have RTT fairness worse than Reno. In contrast
to loss-based algorithms, delay-based algorithms are proactive [9].
When bottleneck queues happen, delay-based approaches reduce network congestion to provide
lowlatency data transfer and convergence to a fair share amongst flows. Many delay-based
algorithms provide high and steady throughput by minimizing variation in data sending [10].
To improve congestion control in SDN by excluding unauthorized devices from authentication due
to the possibility of attackers among them. Also, unauthorized access makes a hacker add
conflicting network regulations or change them. On the other hand, protocol errors, device
impersonation, and assaults involving the insertion of false packets all can cause changes in
network topology [11].

41
Flooding of the controller's switch table in SDN: The acceptance of OpenFlow control packets at
the controller leaves attacks possible because there is no way of authenticating and verifying the
identity of the sender of incoming packets. So, an attacker can save data about fictitious switches
in the switch table of the target controller. Due to persistent flooding of its related switch table,
continuous execution of this attack degrades controller performance [12]. The attacks and
abnormal behavior, including eavesdropping, packet injection, packet crafting, traffic interception,
network information poisoning, and identity theft, are made possible by plain text channels and
improper or weak authentication mechanisms.
To implement a trust mechanism to authenticate and authorize network privileges to prevent
anyone from connecting to the network. In addition to encryption techniques, only data protected
and illegal network access stopped. Thus, SDN security is strong by utilizing authentication and
trust methods to encryption techniques, especially at the control plane. By exchanging signed
certificates, symmetric and asymmetric keys, or message authentication codes, the controller must
be able to identify and then authorize trusted devices (other controllers, switches, and extra
appliances) and network applications, degrades controller performance due to persistent flooding
of its related switch table, continuous execution of this attack [13].
In this thesis, we first employed rerouting traffic to take the shortest route that has minimized
utilized bandwidth possible away from the congestion path that has maximum utilization
bandwidth to provide low latency and high throughput, second; we designed a device
authentication program between the authentication server and users to check if these devices are
trusted. The server has the file of the media access control address (MAC addresses) of authorized
devices; therefore, it confirms that they are in this file and then sends that information to the
controller to allow those devices to access the network or block and disconnect them in the case
of unauthorized devices. This program improves control of congestion and increases network
performance.
1.1. Background and Related Work
Various related tasks provide networks the ability to route traffic. In [14], a work by Sofia Naning
et al., using comprehensive controller information, implemented a congestion control technique.
Congestion control techniques; are implemented by integrating multipath routing techniques with
rate adaptability. The controller uses data from switches to choose the shortest path between the
source and destination and the correct flow rate. But they did not illustrate the topology used.
In [15] work by Renuga Kanagavelu et al. propose a local rerouting strategy in SDN-based data
center networks. Based on our flow classification scheme, the rerouting approach will redirect
flows (at the point of congestion or one hop before) to possible paths; they cared about the size of
the transmitted data as a classification, regardless of its importance and priority in transmission.
In [16], Masoumeh Gholami et al. control the SDN data center congestion using the OpenFlow
protocol. When a network link becomes congested, the OpenFlow-equipped switches' port
statistics; are centrally monitored, and part of the flows is diverse through paths with more
available resources in the link. However, the average delay was reduced only crucial slightly in
the last period of the test.
Umme Zakia et al. [17] used a dynamic load management strategy based on SDNs to maximize
link utilization in Data Center Networks (DCNs) for taking flow priority into account. The program
estimates the cost of each link and determines the shortest routes between each host and the other
hosts. When a path becomes congested, its optimal alternative with the lowest link cost and the
lowest traffic flow takes its place. Throughput, latency, and packet loss in a DCN have assessed

42
the algorithm's performance. But did not been specified when congestion may occur on the link to
try to avoid or reduce it.
As for the authentication of SDN, there are related works like [18] by Diogo Menezes et al.,
Introducing AuthFlow, an access control and authentication system based on host credentials, the
framework for control applications that enables software-defined network controllers to specify
forwarding rules using the host identity as a new flow field. But they did not illustrate the topology
used.
In [19], Jing Liu et al. designed a secure and reliable access method using SDN. The procedure
involves a security access authentication mechanism and an access architecture design. The
particular organization and implementation of data exchange during the access procedure are laid
forth in the security access authentication protocol. The access device and the user's identification
evaluate for credibility using the architecture and protocol. But the hardware environment consists
of one access device, not more.
Our thesis first computes all the shortest paths between any two edges of an SDN network using
Ryu controller [20], OpenFlow 1.3 [21] for the Ryu controller and Mininet [22], and secondly
controls congestion by detecting the congested link with the lowest bandwidth, and then selecting
the shortest path with higher bandwidth than the congested one to alter it of the flows. The proposal
is to design an authentication protocol between client devices and the authentication server to
prevent access to unauthorized devices by communicating the protocol with the controller, which
increases network performance and reduces latency.
2. MOTIVATIONS AND CONTRIBUTIONS
2.1. Motivations
Security, access control, error minimization, connectivity, congestion control, traffic monitoring,
and other issues are all relevant to SDN.
In this research, we concentrate on two problems of Software Defined Networking (SDN):
congestion control and authentication. One of the main factors limiting SDN's performance is
secure authentication. To illustrate, a high percentage of SDN controllers now in use lack a user
access control function means that when the user and device access the network, there may be a
problem with identity authentication.
This problem can result in forgery attacks or unauthorized users and devices of identity to illegally
obtain information and cause congestion that lowers the quality of service and results in queuing
delays, packet losses, and the blocking of new connections.
2.2. Contributions
The objectives of this thesis are achieved by designing an efficient and secure authentication
protocol between access devices and an authentication server to achieve access authentication for
devices across the entire network to improve the network performance and reduce latency and by
designing an efficient congestion control model for the Ryu controller by centrally monitoring the
statistics from OpenFlow switches, then rerouting flows in congested links through the shortest
paths with free resources to improve the performance of SDN accordingly.

43
3. THE PROPOSED APPROACH
We use SDN, which has a structure illustrated in Figure 1. Ryu controller OpenFlow switches and
hosts (devices). We will apply the rerouting approach for congestion control and then improve this
with a secure protocol using the devices' authentication protocol.
Mininet generated the network topology diagram used in the proposal. The topology consists of a
Ryu controller, 13 virtual switches (s₁-s₁₃), and 16 hosts (h₁-h₁₆) in Figure 2.
Figure 1. Structure of SDN
Figure 2. The network topology
3.1. The Rerouting Approach
In this approach, the SDN controller chooses the shortest path for routing traffic. High bandwidth
utilization on links makes network congestion and lower network Quality of Service (QoS)
guarantees possible. It is possible to successfully prevent network congestion by monitoring the
underlying network data and controlling route forwarding to minimize forwarding delays.

44
The sdn_reroute application wrote in python language for the approach for the Ryu controller.
This application consists of four modules: the reroute module, monitor module, network
knowledge module, and network settings module, as illustrated in Figure 1.
In this application, the controller must obtain information about network resources or topology
discovery, such as the number of switch nodes, topology of the network, and link resource
information. Each OpenFlow switch must provide statistics to a controller, requested to OpenFlow
switches to obtain the per-table, per-flow, and per-port statistics.
With this information and statistics from the network knowledge module, we used the
library in this module as a result. Once the y library has been imported into a module, it
can be added using the function, which contains methods such as
for retrieving the list of switches and for retrieving the list of links [23].
Moreover, we used the library in the same module. Once imported, the library in the
module can add shortest-path algorithms that create a route between two nodes in a weighted graph
such that the path's edge sum is reduced to a minimum. The weight is fixed and set to 1 for each
link.
The procedure looks up the weights on the (outward) going (in weighted graphs) edges starting
from the source node to the destination node. It selects a path and the shortest path's total [24].
Any link has parameters that the approach can employ, such as the fact that the remaining
bandwidth of the link, , at the start is equal to the bandwidth of each link, w, and
then it decreases continually by subtracting the amount of bandwidth required for the data flow,
, from with each flow.
Additionally, the minimum bandwidth, , is initially equal to the maximum capacity of the
link, , and then it computes the minimum of the remaining bandwidths of the links,
. All of these parameters are predefined in the sittings module by their values, which indicate that
is set to , is set to , and and are both set to
.
The main module is the reroute module, which obtains all paths between any two nodes; these
paths are sorted from shortest to longest from the network knowledge module, the reroute module
also obtains the shortest path between any two nodes as well as host information, whether source
or destination, such as a MAC address and an Internet Protocol address (IP) from the network
knowledge module.
Moreover, this module detects the congested link that has the least as a result of
consuming bandwidth by more flows, i.e., the of the congested link is less than the
or the of a path is less than the , and then it reroutes to the shortest path that has
links that have a higher than the .
The monitor module allows the controller to assess the amount of bandwidth needed for the data
flow and track link bandwidth utilization, where the minimum bandwidth links, , and
bandwidth remaining at edges, , are computed during data flows for hosts.
To better illustrate the congestion control approach, the pseudocode is used in Algorithm 1, where
and are nodes, A is the source host, and B is the destination host. Moreover, two examples
in the Experimental Results and Analysis section will be mentioned later.

45
Algorithm 1 (reroute approach)
1- Send CO request to OD switches to obtain statistics for each port, flow and Table
2- The switches replay the requested statistics to Controller
3- receive all paths and select the shortest path between any two nodes
4- At the start let , and .
5- for each flow F has the shortest path from A to B do
6- compute for each edge (link) in the path
7- compute
8- if
9- print(“no congestion”)
10- end if
11- else if
12- print (“happen congestion”)
13- detect congestion link
14- select the shortest path that has of links from
15- end else if
16- end for
3.2. The Authentication Approach
The flexibility and programmability of networks afforded by SDN are increasingly needed as the
computing and communication industries evolve; however, security is one of the main factors
limiting SDN progress. The majority of SDN controllers now in use lack user access control
features. When a user or device accesses the network, authentication issues may arise.
This flaw can result in identity forgery, unauthorized access, and control of the entire network,
which could result in congestion or even paralysis. In this thesis, we proposed a trusted access
method for device authentication to address the issue of security checkups for access terminals.
We created and implemented a software-defined network trusted access protocol or technique that
might serve as the first network access defense to improve congestion control [25].
This technique is a device authentication protocol between clients and the authentication server in
python language. The clients’ devices to connect the network require a connection to the
authentication server to secure the network against unauthorized devices. The server asks for the
MAC of the devices from the clients to verify that they are authorized. If thus, the server sends a
confirmation to the client to allow it to connect to the SDN, or else sends the server to the controller
to drop the device from the network if the device is unauthorized.
When the client sends the data to the authentication server, it is encrypted by Elliptic Curve
Cryptography (ECC) to ensure the confidentiality of the data and not be sniping like a man-in-the-
middle attack.
Although the products and standards that employ RSA for encryption and authentication are
public-key, in recent years, the required key length for secure RSA use has increased.
ECC is more difficult to use than RSA. For high-level security functions such as encryption and
authentication, ECC employs arithmetic algorithms as the primary objective operations. Both
software and hardware can use ECC, which uses a standard to produce their public, and private

46
keys, and parties agree on publicly available data items. ECC's primary advantage over RSA is
provided equivalent security for a much smaller key size, which lowers processing overhead. i.e.,
overall, ECC is more effective and secure than RSA [26].
To explain Elliptic Curve Cryptography (ECC) is based on the cubic equation:
Eq. (1)
Where and are real integers and and take values in the real numbers, defines an elliptic
curve over a finite field in two variables. The variables and coefficients for cryptography are
limited to elements in a finite field that are binary curves over GF(2m
) and prime curves over .
We utilize the cubic equation for a prime curve over [27].
Eq. (2)
Where all the variables and coefficients have values in the range of integers from 0 to p-1 and
where computations are done modulo p. Emphasizes that for software applications, prime curves
are ideal. Values =3, b= 1, and p = 3119 were offered from the curve that was used in Eq. (3).
Eq. (3)
We need to identify a "hard issue" that is the discrete logarithm or factoring the product of two
primes to create an elliptic curve-based encryption system. Consider the formula:
, where , , and are all variables.
If and are known, it is quite simple to compute , but, if and are known, it is relatively
challenging to find k which is known as the discrete logarithm problem for elliptic curves. The
difficulty of calculating k using KP and P determines ECC security. The encryption and
decryption system needs a point and an ellipse set as parameters. Is a base point on
the elliptic curve equation, generator point, denoted by the equation . To explain the
steps of encryption and decryption using ECC [28], it would be via our authentication approach
that includes this encryption and decryption. This approach consists of three Python modules: a
server module, a client module, and a coding module, which the pseudocode in Algorithm 2 and
Figure 3 illustrates.
Step 1: In the beginning, this protocol is, as we said earlier, between the authentication server and
clients to authenticate their devices, both the server and the client agree and both select a private
key where and for server and client, respectively. They compute the public key , which is
the product of and the private key, then they exchange public keys. The authentication server
requests device data - the MAC address- from any client who connects to the network.
Step 2: The client module selects a random number and encrypts his device's MAC address using
ECC before sending it to the server. The plaintext message (MAC address) must be encoded first
as a point Pm by calling the encode method in the coding module before being delivered.
This method requires that the ASCII values of the characters be converted to their equivalent
values. = 20 is an auxiliary base parameter; both the client and the server should agree upon
this. Receive all points on the elliptic curve that are used as points.
If is an ASCII value for each character, for each number take . By searching
about in P to find that checks the Eq. (3), try and then until one can

47
solve for y. In reality, one will encounter such a y before reaching . Then take
that point . As a result, the value of m is now represented as a point on the elliptic curve. In
this way, the entire argument is reduced to a series of points. Every pair will serve as a " " input
to the ECC mechanism.
Step 3: As for the encrypt step, in the client module, for each encoded text " " as follows, it
selects a random positive integer ( ), then it computes and , where is the public key of
the server using point multiplication, then calls the encode method in the coding module to receive
then computes + using point addition, and finally, it sends the ciphertext to the
server.
Eq. (4)
Step 4: In the server module, it decrypts each ciphertext as follows: it multiplies the first point in
the pair by the server’s private key nb and subtracts the result from the second point.
Eq. (5)
Then call the decodes method in the coding module to convert encoded Pm as point to m
(MAC address) to be the greatest integer less than [29]. The client has added to
the data to hide it. Although is a public key, only the client knows the value of , making it
impossible to remove the mask . The client does, however, moreover contain a "clue," which
is sufficient to remove the mask if one is aware of the server's private key . Attackers would need
to compute k by using and , which is thought to be difficult, to recover the message.
Step 5: The server module has a file that contains the MAC addresses of the devices authorized to
access the network (the auth file); it compares the client's MAC address with those addresses. If
it matches one of them, it confirms that it will allow the client device to access the network.
Otherwise, the server sends a message to the controller by calling the delete host method in the
reroute module, which removes the unauthorized device from the network.
Figure 3. Flowchart for authentication approach

48
4. EXPERIMENT TESTING AND ANALYSIS RESULTS
4.1. Experiment Setup
The test supports the suggested proposal. On the Ubuntu 18 system, the SDN simulation
environment using Ryu, Mininet, and the network topology is in Figure 2. The test on a computer
with an Intel(R) Core(TM) i7-4600CPU 2.90 GHz CPU and 8 GB of RAM.

49
4.2. Experimental Results and Analysis
To receive results from the rerouting approach, use two terminals. One terminal to run the reroute
module, and another terminal to run Mininet to simulate topology (Fattree topology), then run the
command in the Mininet terminal to test connectivity between all hosts. For example,
Table 1, illustrated the paths from , the algorithm
selected the shortest path, Path 1 because it has fewer weights of edges than Path 2 (less in the
number of hops); the link bandwidth utilization for Path 1 becomes 94.
Table 1. paths from
Paths Edges (Switches) The link bandwidth utilization
Path1
Path2
The higher the flows, the greater the consumption of link bandwidth thus if the link bandwidth
consumption rate is less than the threshold of the link bandwidth utilization, then this link will be
the congested link, such as some of the paths from in Table 2.
Table 2. paths from to
Paths Edges (Switches) The link bandwidth utilization
Path1
Path2
Path3
Path4
The algorithm selected the shortest path, path1 because it has fewer weights of edges than other
paths, thus the link bandwidth utilization for Path1 is in Table 2. But, the algorithm detected
congestion in this path by computing rem_bw for a link in the path that is less than ,
thus it is a congested link.

50
The algorithm reroutes to another link that has the shortest path that is path 3 which has rem_bw
less than , unlike other paths in Table 2.
The network bandwidth measuring tool and traffic generator both use the tool [30]. We use
to measure throughput between in TCP flow by run” ” in Mininet
terminal, start the TCP server ( ) at with port . Moreover, monitor the results
every second ( ) to run in h1 terminal “ ”. Start the TCP client (–c) at
h10. Set the transmission duration (-t) to 100 seconds after -c, one need to specify the server IP
address 10.0.0.4 to run in h1 terminal “ –c 10.0.0.2 –p 5566 –t 100”.
By running the Mininet emulator in one terminal and the Ryu controller's reroute module in
another terminal, then typing " ping " in the Mininet terminal, the round trip time (RTT) of
the network can be measured since it has a significant impact on determining the speed and
dependability of the network connection. To receive results from the authentication approach, run
in Mininet and, using the Ryu controller's sdn_reroute application for all hosts, apply the
device authentication algorithm by running the server module in 1 and the client
module in another .
The server module confirms hosts or devices can access the network except for unauthorized
devices; refer to the auth file. In our algorithm, select h3, h9, h6, and h13 as unauthorized devices
that were dropped by the Ryu controller. We use to measure throughput between and
again after applying the authentication approach in the same steps in the rerouting approach.
In Figure 4 TCP Flow with Authentication, we use to measure throughput between and
by running " " in Mininet Terminal, following the same steps as in the first
stage of the congestion control mechanism. By comparing them, we find that the throughput has
improved when we apply the device authentication algorithm.
Figure 4. The throughputs between
The round-trip time (RTT) of the network can be measured again after the device authentication
approach, as in the same steps in the congestion control approach.
In Figure 5, the round trip time between and is illustrated in the congestion control
mechanism and after the device authentication algorithm. By comparing it, we notice that the RTT

51
is lower after the device authentication algorithm is applied, as the minimum RTT decreases from
0.119 to 0.050 .
Figure 5. The round trip time between and
5. CONCLUSION
To improve network performance and ensure the quality of the SDN service, the first part of our
thesis computes all shortest paths between any two edges of an SDN network using a Ryu
controller and Mininet. The second part of our thesis controls congestion by identifying the
congested link with the lowest bandwidth and then choosing the shortest path with higher
bandwidth than the link's maximum capacity to change the flow path. A device authentication
mechanism is then added between client devices and the authentication server to prevent
unauthorized devices from accessing the network by the controller, increase throughput, and
decrease round-trip times.
REFERENCES
[1] M. R. Belgaum, S. Musa, M. M. Alam, and M. M. Su’ud, "A systematic review of load balancing
techniques in software-defined networking," IEEE Access, vol. 8, pp. 98612-98636, 2020.
[2] J. E. T. Regencia and W. E. S. Yu, "Latency and throughput advantage of leaf-enforced quality of
service in software-defined networking for large traffic flows," in Intelligent Computing: Springer,
2021, pp. 606623.
[3] A. Shirmarz and A. Ghaffari, "Performance issues and solutions in SDN-based data center: a survey,"
The Journal of Supercomputing, vol. 76, no. 10, pp. 7545-7593, 2020.
[4] A. L. Stancu, S. Halunga, A. Vulpe, G. Suciu, O. Fratu, and E. C. Popovici, "A comparison between
several Software Defined Networking controllers," in 2015 12th international conference on
telecommunication in modern satellite, cable and broadcasting services (TELSIKS), 2015: IEEE, pp.
223226.
[5] S. Rowshanrad, S. Namvarasl, V. Abdi, M. Hajizadeh, and M. Keshtgary, "A survey on SDN, the
future of networking," Journal of Advanced Computer Science & Technology, vol. 3, no. 2, pp. 232-
248, 2014.
[6] Y. Maleh, Y. Qasmaoui, K. El Gholami, Y. Sadiq, and S. Mounir, "A comprehensive survey on SDN
security: threats, mitigations, and future directions," Journal of Reliable Intelligent Environments,
pp. 1-39, 2022.

52
[7] B. Turkovic, F. A. Kuipers, and S. Uhlig, "Interactions between congestion control algorithms," in
2019 Network Traffic Measurement and Analysis Conference (TMA), 2019: IEEE, pp. 161-168.
[8] J. Zhang, Z. Yao, Y. Tu, and Y. Chen, "A survey of TCP congestion control algorithm," in 2020
IEEE 5th International Conference on Signal and Image Processing (ICSIP), 2020: IEEE, pp. 828-
832.
[9] Z. Gál, G. Kocsis, T. Tajti, and R. Tornai, "Performance evaluation of massively parallel and high
speed connectionless vs. connection-oriented communication sessions," Advances in Engineering
Software, vol. 157, p. 103010, 2021.
[10] G. Kumar et al., "Swift: Delay is simple and effective for congestion control in the data center," in
Proceedings of the Annual Conference of the ACM Special Interest Group on Data Communication
on the applications, technologies, architectures, and protocols for computer communication, 2020,
pp. 514-528.
[11] L. P. Verma and M. Kumar, "An IoT-based congestion control algorithm," Internet of Things, vol. 9,
p. 100157, 2020.
[12] A. Shaghaghi, M. A. Kaafar, R. Buyya, and S. Jha, "Software-defined network (SDN) data plane
security: issues, solutions, and future directions," Handbook of Computer Networks and Cyber
Security, pp. 341387, 2020.
[13] J. C. C. Chica, J. C. Imbachi, and J. F. B. Vega, "Security in SDN: A comprehensive survey," Journal
of Network and Computer Applications, vol. 159, p. 102595, 2020.
[14] S. N. Hertiana and A. Kurniawan, "A joint approach to multipath routing and rate adaptation for
congestion control in OpenFlow Software Defined Network," in 2015 1st International Conference
on Wireless and Telematics (ICWT), 2015: IEEE, pp. 1-6.
[15] R. Kanagevlu and K. M. M. Aung, "SDN controlled local re-routing to reduce congestion in cloud
data center," in 2015 International Conference on Cloud Computing Research and Innovation
(ICCCRI), 2015: IEEE, pp. 80-88.
[16] M. Gholami and B. Akbari, "Congestion Control in Software Defined Data Center Networks Through
Flow Rerouting," in 2015 23rd Iranian conference on electrical engineering, 2015: IEEE, pp. 654-
657.
[17] U. Zakia and H. B. Yedder, "Dynamic load balancing in SDN-based data center networks," in 2017
8th IEEE Annual Information Technology, Electronics and Mobile Communication Conference
(IEMCON), 2017: IEEE, pp. 242-247.
[18] D. M. Ferrazani Mattos and O. C. M. B. Duarte, "AuthFlow: authentication and access control
mechanism for software defined networking," annals of telecommunications, vol. 71, no. 11, pp. 607-
615, 2016.
[19] J. Liu, Y. Lai, Z. Diao, and Y. Chen, "A trusted access method in software-defined network,"
Simulation Modelling Practice and Theory, vol. 74, pp. 28-45, 2017.
[20] S. Bhardwaj and S. N. Panda, "Performance Evaluation Using RYU SDN Controller in Software-
Defined Networking Environment," Wireless Personal Communications, vol. 122, no. 1, pp. 701-
723, 2022.
[21] B. Isyaku, M. S. Mohd Zahid, M. Bte Kamat, K. Abu Bakar, and F. A. Ghaleb, "Software Defined
Networking Flow Table Management of OpenFlow Switches Performance and Security Challenges:
A Survey," Future Internet, vol. 12, no. 9, p. 147, 2020.
[22] D. Dholakiya, T. Kshirsagar, and A. Nayak, "Survey of Challenges, Opportunities, and
Application in Software-Defined Network (SDN)," in International Conference on Information and
Communication Technology for Intelligent Systems, 2020: Springer, pp. 213-221.
[23] O. Oginni, P. Bull, and Y. Wang, "Constraint-aware software-defined network for routing real-time
multimedia," ACM SIGBED Review, vol. 15, no. 3, pp. 37-42, 2018.
[24] P. Wills and F. G. Meyer, "Metrics for graph comparison: a practitioner’s guide," Plos one, vol. 15,
no. 2, p. e0228728, 2020.
[25] K. Nisar et al., "A survey on the architecture, application, and security of software defined
networking: Challenges and open issues," Internet of Things, vol. 12, p. 100289, 2020.
[26] F. Mallouli, A. Hellal, N. S. Saeed, and F. A. Alzahrani, "A survey on cryptography: comparative
study between RSA vs ECC algorithms, and RSA vs El-Gamal algorithms," in 2019 6th IEEE
International Conference on Cyber Security and Cloud Computing (CSCloud)/2019 5th IEEE
International Conference on Edge Computing and Scalable Cloud (EdgeCom), 2019: IEEE, pp. 173-
176.

53
[27] W. Stallings, "Cryptography and Network Security Principles and Practice Seventh Edition Global
Edition British Library Cataloguing-in-Publication Data," 2017.
[28] A. Srivastava and A. Kumar, "A review on authentication protocol and ECC in IoT," in 2021
International Conference on Advance Computing and Innovative Technologies in Engineering
(ICACITE), 2021: IEEE, pp. 312-319.
[29] P. Bh, D. Chandravathi, and P. P. Roja, "Encoding and decoding of a message in the implementation
of Elliptic Curve cryptography using Koblitz’s method," International Journal on Computer Science
and Engineering, vol. 2, no. 5, pp. 1904-1907, 2010.
[30] J. Dugan, S. Elliott, B. A. Mah, J. Poskanzer, and K. Prabhu, "iPerf-The ultimate speed test tool for
TCP, UDP and SCTP," línea]. Available: https://iperf. fr.[Último acceso: 23 Mayo 2018], 2019.

OPTIMIZING CONGESTION CONTROL BY USING DEVICES AUTHENTICATION IN SOFTWARE-DEFINED NETWORKS

Recommended

More Related Content

Similar to OPTIMIZING CONGESTION CONTROL BY USING DEVICES AUTHENTICATION IN SOFTWARE-DEFINED NETWORKS (20)

Recently uploaded (20)

OPTIMIZING CONGESTION CONTROL BY USING DEVICES AUTHENTICATION IN SOFTWARE-DEFINED NETWORKS