Professional Documents
Culture Documents
Abdul Mohssen Jaber Abdul hossen, Department of Computer Science, College of Science,
University of Technology, Iraq
Mohamed Najem Abdual , Department of Computer Engineering,College of Engineering,
University of Technology, Iraq(mustafamuna@Yaho.com)
MohammadQassim, Department of Communication Engineering &mobile computing, College of
Engineering, University of Information Technology and Communication,
Iraq(mohammad.qassim2002@uoitc.edu.iq)
Abstract
Background/objectives:The presented study will examine the applications regarding to An Aritifical
Intelligentin to Software Defined Network paradigm that is considered as part of past attempts for
giving the networks of computer the capability of being programmed depending on separation
between controlas well as the forwarding planes. With regard to the Software Defined Network
method, controller will be specifying the central brain regarding network that result in sophisticated
flexibility level as well as network intelligence.
Methods/ Statistical Analysis: A lot of Artifical intelligent-based approaches were utilized for
achieving improved load balance, network security as well as intelligent applications of thenetwork
in the method of Software Defined Network.
Results: The algorithm of ant colony have been of high importance to increase the maximum Quality
of Experience (QoE) by 24.10% in comparison to minimal path routing method. NN-based system of
preventing intrusions indicates good performance with the little false positive rate. Utilizing the
approach which is based on the reinforcement learning in the systems of the adaptive video streaming
in comparison to shortest path routing as well as the greedy-based method decrease the rate from the
loss of frame by 79% and 60%.
Applications/Improvements: The presented work is the initial attempt to apply AI in the concept of
the Software Defined Network. Furthermore, the hybrid intelligent approaches may be of high
importance to achieve improved behavior in the SDN-based networks.
Keywords:OpenFlow,Arartificial Intelligence(AI),Software Defined Networking (SDN).
1. Introduction:
Recently, traditional traffic of internet was altered to be extra complicated, particularly with the
presence of big data, the data centers now are requiring extra scalability as well as flexibility. Along
with the growth of the advanced network applications in addition to the presence regarding a lot of
types regarding devices even in one area, a lot of endpoint devices might be sharing and exchanging
1133
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
various pattering related to the network traffic. This will be indicating that the present infrastructure
of network doesn’t have the ability for meeting such requirements and novel method required. Also,
due to the fact that there isn’t any central control way in network, network device’s configuration
isn’t consistent and will be consuming much time [1] .Furthermore, conventional networks are
dealing with the distributed management as well as the network processing that is related to network
decision making. Thus, this study will deal with a lot of nodes of networks which might be costly in
certain conditions such a Virtual Machine (VM) migration4 [1,2].Also to add a quality of Experience,
a quality of Services, in addition to policies of security for each one of the end pints in complex
network architectures will be showing another drawback into legacy network approach4 .[1,3]
2.SDN ARCHITECTURE
As it has been stated earlier, the Software Defined Networks is promote innovations through
the introduction of the idea of the centralized programmable data plane control that can facilitate
developing a new network protocols then services [4]. The architecture of the SDN has been modelled
onto a basis of the concept of separation between the control and the data planes (as can be seen in
Fig. A).
The initial attempt has been the Network control point (NCP) [5][14]that has utilized the
concept of these paration for the enhancement of the control of the AT&T’s telephone network, while
the latest proposals, like SANE [6][15] and Ethane [7] have implemented an identical concept for the
Ethernet network[7] [1,6]. SDN applications are into the application plane of the architecture of
Software Defined Networks, in which the northbound application programming interface (API) has
provided commutation between the control planes and the application [2,8], allowing the
implementation of a group of the network services like the intrusion detection, traffic engineering,
quality of service (QoS), firewalls and monitoring applications [4,10].
The north-bound APIs allow the developers writing their custom applications with no needing
to have a detailed information about the functions of the controller or understanding the way by which
data plane operates. It should be mentioned that numerous controllers of the SDN, supplying their
particular Northbound Application Programming Interfaces[10,16]. The communications through the
control and the data planes are given with the use of a south-bound API, like the forwarding and
element of the control separation (ForCES) [17], open vSwitch database (OVSDB) [18], protocol
oblivious forwarding (POF) [19], Open State [20], Open Flow (OF) [21] in addition to the opflex [22],
enabling the exchange of the control messages with the forwarding elements (such as the Open-Flow-
enabled switches). As can be seen in Fig1, every one of the Open Flow- enabled switches can adopt
the flow-based decision making logic which is specified by the so-called Software Defined Network
controller, responsible for the preparation of the forwarding tables of every one of the switches [8].
The OF-enabled switch has a flow tables’pipe line, consisting of the flow entries, every one them has
3 parts: (a) the matching rules utilized to match the incoming packets (b) the counters, maintaining
the matched flow statistics (c) instructions or actions that may be reactively or proactively configured
for being managed upon match [6, 14].
About elements forwarding (in other words, the Open-Flow which an enabled switches) may
be implemented in software as well as hardware. Some of the software switches like the Open vSwitch
have a massive possibility to provide a solution for the data centres and the virtual networks [16].
However, other API types [10, 23] have been suggested for a particular in (for example, the
applications of the VOIP and the inter-domain routing, besidesa variety of the programming languages
1134
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
of the Software Defined Networks, like a procera [24], Net Core[25] and Frenetic [26], providing
high-level Application Programming Interfaces which may be utilised into developing various
Software Defined Networking's applications into more flexible then functional manner.
1135
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
1136
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
switches as well as enhanced operations for saving time spent for communicating between
controller and switch in the conditions of flow table is full9 . Network system’s decoupling
will offer the capability of managing it via high abstraction level. Software Defined Network,
that is considered as current paradigm to the programmable networks, will be facilitating the
operations of network like routing or even add rules to devices of forwarding through single
central controller. Which indicates the objects of forwarding are going to be implementing
decisions provided via such controller. Comparison between the architecture of SDN as well
as the traditional architecture of network will be seen in the Figure 2. The major abstraction
concept specified into Software Defined Networkingare6:
1. Forwarding, 2. Distribution, and 3. Specification. Forwarding abstraction will enable
conducting actions of forwarding through controller whereas it will be hiding a low-level
handling with the devices of switching. The abstraction of the distribution consists of the
replacement of conventional distributed control planes through logically centralized plane. The
abstraction of the specification will allow developers writing the applications of network
through specifying the needed flow actions in addition to the configurations without managing
with a low-level or physical configurations. Logical centralized control plane has been
indicated in the Fig2 will be providing general view with regard to network that will open the
door for enhanced control of forwarding elements. It might be accomplished through
distributed or single controllers10 .Also, FlowVisor16 that is considered as proxy controller,
will offer logical de-centralization with regard to the network virtualization.A Short
comparison regarding platforms of controller will be provided in the Table1. The major
simulators and emulators of Software Defined Network like Mininet, Network Simulation3
and Estinetwill be compared and specified in the Table2.
1137
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
version
university
Networks
Networks
Foundation
Simulator/ Open
Language Platform OpenFlow version
emulator source
1138
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
EstiNel
(emulator/ No - - OF 1.30 and 1
simulator)
The capability of programming forwarding plane of Software Defined Networking through Software
Defined Network controller might be saving the costs needed to add specialized devices like IDSs,
load balancers, as well as firewalls (IDS)25 . Also, SDN will be providing extra dynamic as well as
cost effective solutions for approaches utilized in WAN like path computation technologies, ranging
such solutions as the applications of Software Defined Network allow operators of network for
directly programming Path Computation Element (PCE) 26 .Thus, the controller will achieve
adaptive behavior in various environments like to be a VM manager into the private cloud27 .
1139
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
∑𝑲
𝒋=𝟏 𝑿[𝒋] 𝟐
√(∑𝒌𝒋=𝟏 𝑿[𝒋]𝟐 )−( 𝑲
)
𝑲
𝑴𝒊𝒏 ∑𝒌
(1)
𝒋=𝟏 𝑿[𝒋]
𝑲
1140
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
SDN
Security Use
Characteristics
— Network Forensics.
The algorithm of TRW-CB that identify SYN Flooding, resulted from host depending on the
approach that benign host is going to acquired high effective connection probability in comparison to
malicious one46 . Input with regard to module of fuzzy logic acquired through the indicated
algorithms and attack degree acquired as output. Decision making system has been conducted as a
software defined network controller application within a short-term module of learning so as can be
seen into the Fig4. The suggested system showed enhanced results in comparison to non fuzzy logic
method. Through benefiting from global view, SDN concept BPNN based collaborative intrusion
prevention system conducted in47. Every one of the Open vSwitchs has been responsible to collect
data for performing inputs for a lot of ANNs. The system has been trained offline through MATLAB.
Due to the fact that it’s collaborative system, Open v-Switches required communication with one
another. Regrettably, with regard to the SDN paradigm, Open v Switches have no ability of talking
to one another. Thus, table of neural forwarding in every one of the Open Vswitchs was achieved,
also the controller could help to build such tables. Figure 5 displays template related to neural message
as well as results of experimentations indicate that the network grows in the rate of the detection
regarding distributed DoS attack will increase and the false positive rate will be decreased. While a
study by 48 provided Self-Organized Maps (SOM) method to detect DDOS attacks. SOM can be
considered as variant regarding ANNs depending onto unsupervised learning. SOM could be applied
as classification approach 49 in the case when dealing within unlabeled input vector. Training in the
SOM is on the basis of set of required features from the flow entries related toOpen v-Switches. The
loop of the detection includes 3 phases: 1. Flow collection that request flow entry from every Open
v-Switch. 2. The extraction of thefeaturesthat take output of flow collection module as well as
extracting the major significant features which from possible DDoS attack. Such features consist of:
Average of Bytes per flow (ABf), Average of Packets per flow (APf), Percentage of Pair-flows (PPf),
Average Duration per flow (ADf), Growth of Different Ports (GDP), and Growth of Single-flows
(GSf) and (c) SOM classification that has been applied as classification approach. Such stages
conducted as modules of application levels in the SDN controller. The suggested method put to
1141
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
comparison with various approaches implemented on recognized KDD-99 dataset showed low over-
head 48,50
Q̅ (St, at) = Q̅ (St, at) + n (rt+1 + y max at Q̅ (St+1, a t+1) – Q (St, at)) (3)
And softmax function displayed in the Equation (4) representing probability regarding choosing
action in state s time t.
𝑄(𝑠,𝑎)
exp[
𝑇
𝑃(𝐴|𝑠) = 𝑄(𝑠,𝑏) (4)
∑𝑏𝜖𝐴 exp[ ]
𝑇
In which T representing random move which is applied in the simulated annealing approach for
escaping from the issue of the local optima. The controller could be changing the existing a path then
adaptively extracting/adding which chosed layers onto a basis of provide bandwidth for increasing
QoE regarding service of the video streaming. The indicated method put to comparison with shortest
1142
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
path routing and greedy-based methods showed decrease in the rate of the frame loss by 80% and
60%.
4. Conclusion
The presented work will be providing summaryregarding integration between AI approaches and
SDN paradigm, also it defined the major architecture of SDN and the important role regarding Open-
Flow protocol in it. After that it provided a summary of the current work contributions for providing
more advanced behavior of network in SDN method. NNs were utilized in various applications like
network security and load balancing. Using the approaches of AI in SDN security aware systems
indicated decrease in false positive detection rate. Furthermore, the results related to adaptive video
streaming system indicated decrease in frame loss rate. Therefore, the studies in this filed are rapidly
growing, also the hybrid intelligent method might bring additional enhancements to the area which is
related to the SDN-based networks.
5. References
1. Open Networking Foundation. Software-defined networking: The new norm for networks.
Available from: https://
www.opennetworking.org/images/stories/downloads/sdnresources/white-papers/wp-sdn-
newnorm.pdf
2. Astuto BN, Mendonca M, Nguyen XN, Obraczka K, Turletti T. A survey of software-defined
networking: past, present, and future of programmable networks. IEEE Communications Surveys
and Tutorials. 2014; 16(3):1617–34.
3. Bakshi K. Considerations for Software Defined Networking (SDN): Approaches and use cases.
IEEE Aerospace Conference, Big Sky; MT. 2013 Mar. p. 1–9.
4. Shinde MB, Tamhankar SG. Review: software defined networking and OpenFlow. International
Journal of Scientific Research in Network Security and Communication. 2013Jun; 1(2):18–20.
5. Feamster N, Zegura E, Rexford J. The road to SDN: An intellectual history of programmable
networks. ACM SIGCOMM Computer Communication Review archive. 2014; 44(2):87–98.
6. Kreutz D, Ramos FMV, Verissimo P, Rothenberg CE, Azodolmolky S, Uhlig S. Software-
defined networking: A comprehensive survey. Proceedings of the IEEE. 2015; 103(1):14–76. 7.
Basta A, Kellerer W, Hoffmann M, Hoffmann K, Schmidt E-D. A virtual SDN-enabled LTE
EPC architecture: A case study for S-/P-gateways functions. Future Networks and Services
(SDN4FNS); Trento. 2013 Nov. p. 1–7.
7. Jammal M, Singh T, Shami A, Asal R, Li Y. Software defined networking: State of the art and
research challenges. Computer Networks. 2014; 72:74–98.
8. Rowshanrad S, Namvarasl S, Abdi V, Hajizadeh M, Keshtgary M. A survey on SDN, the future
of networking. Journal of Advanced Computer Science and Technology. 2014; 3(2): 232–48.
9. Braun W, Menth M. Software-defined networking using OpenFlow: Protocols, applications and
architectural design choices. Future Internet. 2014; 6(2):302–36.
1143
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
1144
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
20,(1),pp.45-56.
28. Heorhiadi, V., Reiter,M.K.,Sekar, V.:‘Simplifying Software-Defined Network Opti- mization
Using SOL’. In: Proceedings of the 13th Usenix Conference on Networked Systems Design
and Implementation, California, USA, March 2016, pp.223-237.
29. Martini, B., Adami, D., Sgambelluri, A., et al.: ‘An SDN orchestrator for resources chaining in
cloud data centers’. In: Proceedings of 2014 European Conference on Networks and
Communications (EuCNC), Bologna, Italy, June 2014, pp.1-5.
30. Akyildiz, I.F., Lee, A., Wang, P., Luo, M., Chou, W.: ‘A roadmap for vation in campus
networks’, ACM SIGCOMM Computer Communication Review, 2008,38,(2), pp.69-74.
31. IETF. ’OpFlex Control Protocol’ [Internet]. 2014. Available from: http://tools.ietf.org/html/draft-
smith-opflex-00. [Accessed:2017-10-16]
32. Ferguson, A.D., Guha, A., Liang, C., Fonseca, R., Krishnamurthi, S.:‘Participatory networking:
An API for application control of SDNs’, ACM SIGCOMM computer communication review,
2013,43,(4),pp.327-338.
33. Voellmy, A., Kim, H., Feamster, N.: ‘Procera: a language for high-level reactive network
control’. In: Proceedings of the first workshop on Hot topics in software defined networks,
Helsinki, Finland, August 2012, pp.43-48.
34. Mittal P, Singh Y. Development of intelligent transportation system for improving average
moving and waiting time with artificial intelligence. Indian Journal of Science and Technology.
2016 Jan; 9(3). DOI: 10.17485/ijst/2016/ v9i3/84156.
35. Davis B. Leveraging the load balancer to fight DDoS. Available from:
http://www.sans.org/reading-room/ whitepapers/firewalls/leveraging-load-balancer-fightddos-
33408
36. Califano A, Dincelli E, Goel S. Using features of cloud computing to defend smart grid against
DDoS attacks. 10th Annual symposium on information assurance (Asia 15), ALBANY; 2015Jun.
p. 44–50.
37. Chen-Xiao C, Ya-Bin X. Research on load balance method in SDN. International Journal of Grid
and Distributed Computing. 2016; 9(1):25–36.
38. Ruelas AMR, Rothenberg CE. Implementation of neural switch using Open Flow as load
balancing method in data center. Campinas, Brasil: University of Campinas; 2015.
39. Chou L-D, Yang Y-T, Hong Y-M, Hu J-K, Jean B. A genetic-based load balancing algorithm in
open flow network. Advanced Technologies, Embedded and Multimedia for Human-centric
Computing. 2013; 260:411–7.
40. Balaguer R. Flow embedding algorithms for software defined audio networks [Master thesis].
Zurich, Switzerland: Swiss Federal Institute of Technology. Available from: http://ftp.
tik.ee.ethz.ch/pub/.../MA-2014-14.pdf
41. Dobrijevic O, Santl M, Matijasevic M. Ant colony optimization for QoE-centric flow routing in
software-defined networks. 2015 11th International Conference on Network and Service
Management (CNSM); Barcelona. 2015 Nov. p. 274–8.
1145
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC
International Journal of Advanced Science and Technology
Vol. 29, No. 5s, (2020), pp. 1133-1146
42. Latah M. Solving multiple TSP problem by K-means and crossover based modified ACO
algorithm. IJERT. 2016 Feb; 5(2):430–4.
43. Akhunzada A, Ahmed E, Gani A, Khan MK, Imran M, Guizani S. Securing the software defined
networks: taxonomy, requirements, and open issues. IEEE Communications Magazine.
2015Apr; 53(4):36 –44.
44. Jankowski D, Amanowicz M. Intrusion detection in software defined networks with self-
organized maps. Journal of Telecommunications and Information Technology. 2015; 4:3–9. 39.
Dabbagh M, Hamdaoui B, Guizaniy M, Rayes A. Softwaredefined networking security: Pros and
cons. IEEE Communications Magazine. 2015; 53(6):73–9.
45. Bai H. A survey on artificial intelligence for network routing problems. NM,USA: University of
New Mexico; 2007.
46. Mustafa U, Masud MM, Trabelsi Z, Wood T, Al Harthi Z. Firewall performance optimization
using data mining techniques. 2013 9th International Wireless Communications and Mobile
Computing Conference (IWCMC); Sardinia. 2013 Jul. p. 934–40.
47. Mukherjee D, Acharyya S. Ant colony optimization technique applied in network routing
problem. International Journal of Computer Applications. 2010; 1(15):66–73.
1146
ISSN: 2005-4238 IJAST
Copyright ⓒ 2020 SERSC