IRJET- Detection of Distributed Denial-of-Service (DDos) Attack on Software Defined Network (SDN)

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 12 | Dec 2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 918
Detection of Distributed Denial-of-Service (DDoS) Attack on Software
Defined Network (SDN)
Mr. Ajinkya Patil1, Mr. Pratik Jain2, Mr. Ravi Ram3, Mr. Venkatesh Vayachal4, Prof. S. P. Bendale5
1,2,3,4B. E. Student, Dept. of Computer Engineering, NBN Sinhgad School of Engineering, Ambegaon, Pune – 411041,
Maharashtra, India
5Professor, Dept. of Computer. Engineering, NBN Sinhgad School of Engineering, Ambegaon, Pune – 411041,
Maharashtra, India
----------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Software Defined Network (SDN for short)
enables better network flow, managing network traffic, and
optimizing the network to work better than traditional
network. Software-defined networking technology is a cloud
computing approach that facilitates network management
and enables efficient networkconfigurationprogrammatically
to improve the performance of the network and to facilitate
monitoring. SDN addresses the fact that the traditional
networks have a static architecturewhichis decentralized and
highly complex. The need of current networks is flexibility and
easy and efficient troubleshooting. SDN uses the concept of
centralization of network intelligence in a single main
network component. This is achieved by dissociation of the
forwarding process of network packets from the routing
process.
The rate of development of internet technology is
higher than ever. Due to this rapid development, the network
flow rates are now higher than ever. In addition, the
Distributed Denial-of-Service (DDoS) attacks which poses a
major threat to network security are now prevalent. In
computer networks, a Denial-of-Service (DoS) attack is a
cyber-attack where, the attacker or the mastermind's goal is
to make the network resources or a machine (such as Servers,
Network Controllers, Access Points, etc.) unable to process the
requests of the intended users. The attacker achieves this by
disrupting the services of a machine (host) connected to the
network. If any host in the network is unable to process or
function the requests from users, the network fails.
Using functionalities of Mininet such as OpenFlow
Switches, Ryu Controllers, Collection Modules and feature
extractions we are trying to simulate an SDN (Software
Defined Network). A DDoS attack on this network will be
simulated. We will try to detect this attack on the network
using detection methods based on data mining techniques.
Key Words: Software Defined Network (SDN), Denial-of-
Service (DoS), Distributed Denial-of-Service (DDoS).
1. INTRODUCTION
Software Defined Network (SDN in short), is an architecture
that is dynamic, it can adapt to different functionalities such
as high-bandwidth, profitable, and can be managed easily
compared to traditionalnetworkmodel.[1]SoftwareDefined
Networking provides number of benefits, centralized
network provisioning, better enterprise management,better
security, low operational costs, isolation and traffic control,
managing packet forwarding.TheSDNsuggestsaCentralized
Network by dividing the architecture into Network Control
Plane and Forwarding Plane. The network control plane is
directly programmable and consists of one or more
controllers which is also considered as Brain of SDN.
With the separation of Control Plane, the administrators
are able to dynamically adjust traffic flow in the whole
network, according to network needs.[2]Administratorscan
also configure and optimize the network security and secure
the network resources with the help of SDN programs.
The network implementation, configuration and
troubleshooting require high skilled network and system
engineers. The system managers can control different
components or “layers” (i.e., application, control and data
plane), they canallocateresources to network users through
application layer, manage the network entities through
control plane, and network devices on data plane.
The OpenFlow protocol was one of the important
elements forbuilding a SDN, itcanalsobecalledasOpenFlow
framework, first SDN standard. Most of the software defined
network have some version of SDN Controller, as well as
Southbound APIs andNorthbound APIs asshowninFigure1.
ThecontrollersandswitchesfollowOpenFlowstandards,and
OpenFlow runs between them acting as a communication
medium. There are different controller platforms which are
open source such as Beacon,OpenDayLight,Floodlight,Open
vSwitch.
Fig - 1: Software Defined Architecture

The Traditional network are static and can be
programmed at the time of installation, while the Software
Defined Network (SDN) are programmable at deployment
time as well at later stages. Traditional networks have
distributed control plane and SDN have centralized control
plane. The Traditional network are hardwareappliancesand
works using protocols. The SDN are configured using open
software and use APIs to configure the network as per the
needs.
As the Software Defined Network is based on centralized
control there is high-valuetargetexposedinthenetwork.The
attackers can have total control over the network by taking
control over it, and a single vulnerability could cause lot of
damage, so security must be primary concern when
deploying the SDN, it also increases the workload of the
administratorassecuritymustbedeployedmanually.Attacks
such as Network Manipulation, Traffic diversion, App
manipulation, API exploitation Traffic sniffing, Denial-of-
Service (DOS) and Brute force. We are focusing on
Distributed Denial-of-Service (DDoS) attack in particular.
The Distributed Denial-of-Service (DDoS) are most
threatening challenges to Internet Security nowadays. As
shown in Figure 2, [3] The DDoS attacker first takes control
over the network devices(computersorotherdevices)called
as hosts in the network to carry out the attacks, the goal of
the attacker is to overload the data packets being sent to
target (server’s) and making the legitimate packet flow from
accessingavailableservices.[4]Theattackerusesmanyhosts
they can send multiple new packets at the same time, the
incoming new packets will search for its destination
information which isstoredinswitchforwardingtable,asthe
legitimate packets cannot be differentiated by harmful
packets hence the attack becomes successful.
Fig -2: Distributed Denial-of-Service (DDoS) Attack
Detecting these DDoS attacks can be very hard done by
conventional detection methods. The attackers may be
distributed and located under different switches, the
detection process may not benefit by performing detection
process into the switches because switches may not detect
attacks completely.
2. OBJECTIVES
As the requirements are dynamic in the today’s world it
becomes a necessity forthe systemconfigurationtoevolveas
per the requirement. The networking systems have a drastic
role in the system communication. As the requirements
change it is need to configure a network which could be
modified with the help of the software remotely instead of
configuring hardware as and when required. The Software
Defined Network (SDN) is a way through which this could be
attained easily.UsingtheSoftwareDefinedNetwork.Withthe
help of SDN it is easy to configure the network as per the
traffic and as per theuserdemandsincrease,thetopologiesof
the network could be easily changed, the networktrafficflow
could be monitored and also it is possible to re-distributethe
traffic to avoid the bottlenecks.
The SDN allows us a vast fieldof network management to
ease the networking solutions. The SDN is vulnerable to
attacks from outside sources and the data needs to be
protected. The attacks fromparasitescouldcauseahugedata
loss leading to loss in data integrity as well as data
consistency. The familiar types of attacks in the SDN are the
Denial-of-Services attack and Distributed Denial-of-Service
attack.
DoS and DDoS attacks are the types of attack in which the
system is flooded with number of un-legit requests, thus
keeping the servers busy serving the fakes and denying
service to the legit requests. This causes system breakdown
and leads to mass communication failure in the cases of
heavy traffic servers of social media platforms.
Our study mainly focusesonthedetectionofDDoSattacks
in the software defined network, this would allow us to
protect the system from denying services to the users and
protecting the serviceintegrity.Oncetheattackonthesystem
is recognized it is easy to control the attack by disabling the
particular sets of internet protocol addresses and thus
securing the systems.
3. RELATED RESEARCH
Currently, there are various methods available to detect the
DDoS attack. Some of the methods are mentioned below.
In a research paper, D. Kotani [5] proposed packet-in
filtering mechanism which makes the SDN control panel
secure. The mechanism works by recording the values of
packet headers, before passing the packets then filtering the
packets which aren’t recorded. This method works until the
attacker doesn’t generate new flows.
S. Mousavi [6] introduced a detection method before a
DDoS attack is initiated on thenetwork.Itisbasedonentropy
variation of data flow on IP address. It assumes that the IP
addresses (destination) are evenly distributed. If any attack
flow is present then the IP addresses must be in small
quantity to generate low-traffic flow.
P. Dong [7] proposes detection method for DDoS attack
against controllers. The switches contain the information of

incoming flow and reports it to the controller. The attack
detection module runs on the northbound interface of the
controller distinguishing between flow statistics.
Yang Li [8] proposes featureselection and featureweight
mechanisms to reduce the computational cost and boost
performance. The method is based upon Transductive
Confidence Machines for K-Nearest Neighbors (TCM-KNN)
algorithm.
S. Jaiswal [9] uses K-Nearest Neighbor (KNN) classifier
and Ant Colony Optimization (ACO) techniques. It is applied
on multiple classifiers, which aims on misclassified features
taking more amount of timeto calculate and classify,theACO
optimizes the category until the features are accurately
classified. It employsID3(decisiontree)algorithmforfeature
reduction.
H. Peng [10] focuses on the anomalous flow, andpresents
anomaly flow detection method, it collects the information
(such as flow details, flow type, all features by which we can
distinct between multipleflows)fromswitchesorcontrollers
and then Detection mechanism is responsible for classifying
the features and network flow with Double P-value of
Transductive Confidence Machines for K-Nearest Neighbors
(DPTCM-KNN) algorithm.
4. RESULT ANALYSIS
Based on the experimental results performed on KDD’99
Dataset which is most used data set for anomaly detection
methods and research. [11] KDD99 is feature extracted
version of DARPA which is the base raw dataset. In the 1998
DARPA Intrusion Detection System Evaluation Program was
prepared, consistingofanattackscenariotoAir-Forcebase.It
consists of host and network dataset files; Host dataset file is
small dataset containing system calls. Network Dataset is
mostly used because it consists of seven weeks of network
traffic (TCP/IP dump).
Table -1: Conversion Matrix
Confusion Matrix is used to measure the performance of
the classification algorithm or classifier. We can derivemany
ratios from confusion matrix:
 True Positive (TP): These are cases in which we
predict ‘Yes’ and the actual result shows ‘Yes’.
 True Negatives (TN): These are cases in which we
predict ‘Yes’ but the actual result shows ‘No’
 False Positive (FP): Here wepredict‘Yes’,andactual
result is ‘No’. This is also called Type I error.
 False Negative (FN): Here wepredict‘No’andactual
result is ‘Yes’. This is also called Type II error.
The findings are shown in Table 2 which use Mininet
Emulatorforsimulatingvirtualsoftware-definednetwork,for
TCM-KNN algorithm which improves featureselection,KNN-
ACO algorithm which combines K-Nearest Neighbor and Ant
Colony Optimization, and DPTCM-KNN algorithm which
improves the precision even further.
Table -2: Results based on KDD99 Dataset
Method
True Positive
Rate
False Positive
Rate
TCM-KNN Low (89%) Medium (11%)
KNN-ACO Medium (92%) Medium (11%)
DPTCM-KNN High (97%) Low (6.25 %)
 True Positive Rate (TPR) = TP/(TP + FN)
 False Positive Rate (FPR) = FP/(TP + FN)
5. CONCLUSION
In this paper, we have studied various methodologies which
are used for detection of Distributed Denial-of-Service
(DDoS) Attacks on Software Defined Network (SDN), based
on the findings and results we have concluded that the
Double P-value of Transductive Confidence Machines for K-
Nearest Neighbors (DPTCM-KNN) method is more feasible
and efficient to find out anomalous flow in Software Defined
Network.
REFERENCES
[1] Benzekki, K., El Fergougui, A., & Elbelrhiti Elalaoui, A.
(2016). Software‐defined networking (SDN): a
survey. Security and communication networks, 9(18),
5803-5833.M. Young, The Technical Writer’sHandbook.
Mill Valley, CA: University Science, 1989.
[2] Khan MFI (2017) Software-Defined Networking
Reviewed Model. Int J Adv Technol 8: 177.
doi:10.4172/0976-4860.1000177
[3] S. P. Bendale, J. R. Prasad, “Security threats and
challenges in Future Mobile Wireless Networks”, IEEE
International Conference proceeding GCWCN, 2018-19.
[4] Al-Mafrachi, B. H. A. (2017). Detection of DDoS Attacks
against the SDN Controller using Statistical Approaches
(Doctoral dissertation, Wright State University).
[5] Kotani, D., & Okabe, Y. (2014, October). A packet-in
message filtering mechanism for protection of control
plane in openflow networks. In Proceedings ofthetenth
Predicted
Values
Actual Values
Positive (1) Negative (0)
Positive (1) TP FP
Negative (0) FN TN

ACM/IEEE symposium on Architectures for networking
and communications systems (pp. 29-40). ACM.
[6] Mousavi, S. M. (2014). Early detectionofDDoSattacksin
software defined networks controller (Doctoral
dissertation, Carleton University).
[7] Dong, P., Du, X., Zhang, H., & Xu, T. (2016, May). A
detection method for a novel DDoS attack against SDN
controllers by vast new low-traffic flows.
In Communications (ICC), 2016 IEEE International
Conference on (pp. 1-6). IEEE.
[8] Li, Y., & Guo, L. (2008, March). TCM-KNN scheme for
network anomaly detection using feature-based
optimizations. In Proceedings of the 2008 ACM
symposium on Applied computing (pp. 2103-2109).
ACM.
[9] Jaiswal, S., Saxena, K., Mishra, A., & Sahu, S. K. (2016,
March). A KNN-ACO approach for intrusion detection
using KDDCUP'99 dataset. In ComputingforSustainable
Global Development (INDIACom), 2016 3rd
International Conference on (pp. 628-633). IEEE.
[10] Peng, H., Sun, Z., Zhao, X., Tan, S., & Sun, Z. (2018). A
detection method for anomaly flow in software defined
network. IEEE Access.
[11] Özgür, A., & Erdem, H. (2016). A review of KDD99
dataset usage in intrusion detection and machine
learning between 2010 and 2015. PeerJ PrePrints, 4,
e1954v1.

IRJET- Detection of Distributed Denial-of-Service (DDos) Attack on Software Defined Network (SDN)

More Related Content

What's hot (20)

Similar to IRJET- Detection of Distributed Denial-of-Service (DDos) Attack on Software Defined Network (SDN) (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Detection of Distributed Denial-of-Service (DDos) Attack on Software Defined Network (SDN)