close

Вход

Забыли?

вход по аккаунту

?

j.compeleceng.2017.10.011

код для вставкиСкачать
ARTICLE IN PRESS
JID: CAEE
[m3Gsc;October 24, 2017;14:50]
Computers and Electrical Engineering 0 0 0 (2017) 1–10
Contents lists available at ScienceDirect
Computers and Electrical Engineering
journal homepage: www.elsevier.com/locate/compeleceng
FCM technique for efficient intrusion detection system for
wireless networks in cloud environmentR
Mingming Chen a,b,c, Ning Wang a,b,d,e,∗, Haibo Zhou a,b, Yuzhi Chen a,b
a
Dept. of Information and Mechatronics Engineering, Xiamen Huaxia University, Xiamen, Fujian 361026, China
Fujian Province Engineering Research Center on New Generation of Information and Communication Technology and Wisdom Education,
China
c
University of Illinois at Springfiel, United States
d
Dept. of Automation, Xiamen University, Xiamen, Fujian, China
e
School of Computing and Information Sciences at Florida International University, FL 33172, USA
b
a r t i c l e
i n f o
Article history:
Received 24 February 2017
Revised 15 October 2017
Accepted 16 October 2017
Available online xxx
Keywords:
Cloud storage
Intrusion detection systems
MANET
Fuzzy clustering techniques
a b s t r a c t
With the emergence of ad-hoc networks, the communication methods in the field of wireless have greatly developed. A great deal of research has been conducted, especially in
mobile nodes or mobile ad hoc networks (MANETs). The greatest advantage of MANETs is
that they don’t require predefined infrastructures, such as router link, hub etc. The nodes
in a MANET are capable of forming a network which is time-dependent but not permanent
to establish communication based on the allocated path. Routing attacks can cause great
damage to MANETs. In the past few years, the research work has proposed a special intrusion detection mechanism to detect these attacks, resulting in eliminating malicious nodes
in the network. The proposed work in this research paper addresses an efficient fuzzy
clustering based algorithm for intrusion detection of a MANET implementation in a cloud
storage environment. This paper has presented a model and experimental justifications for
improving the efficiency.
© 2017 Published by Elsevier Ltd.
1. Introduction
With the development of advanced communication technologies and protocols, the field of information technology has
undergone a great revolution, especially in dealing with large amounts of data and its storage in the cloud environment. This
has become an inevitable trend, especially in today’s environment where all communications and transactions are digital.
Promotions in the global IT industry are heavily depended on cloud environment and its effective management [1–4]. The
efficiency of the cloud is largely determined by the rapid retrieval of system storage and demand. Another key requirement
of the cloud is that it requires a fool security environment to protect the systems from unauthorized data accesses. The
security of the cloud is affected by intrusion attacks in many forms, which may cause great damage to the corresponding
layers such as transport layer and application layer. Hence,in order to improve the security of cloud storage system, it is urgent to develop an intrusion detection scheme or algorithm. In the cloud storage environment, no infrastructure is required
and predefined, thus we use the Ad Hoc network. The Ad hoc network (MANET) model and nodes play important roles in
R
∗
Reviews processed and recommended for publication to the Editor-in-Chief by Associate Editor Dr. S. Smys.
Corresponding Author.
E-mail address: [email protected] (N. Wang).
https://doi.org/10.1016/j.compeleceng.2017.10.011
0045-7906/© 2017 Published by Elsevier Ltd.
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
JID: CAEE
2
ARTICLE IN PRESS
[m3Gsc;October 24, 2017;14:50]
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
Fig. 1. Illustration of Cloud environment.
Fig. 2. MANET architecture for communication for four users.
the network, and packet communicate from the sources to the destination [5–7]. A typical cloud storage environment is
depicted in Fig. 1.
As shown in Fig. 1, the efficiency of the cloud environment depends on several attributes, such as data protection and its
authorized mechanism, the communication of a secured channel from source to destination, protection of nodes from unauthorized access and usage, effective intrusion detection mechanism which protects the nodes or if infected, and a mechanism
which isolates the malicious node from rest nodes. A basic advantage of cloud storage based on traditional storage mechanisms is that data can be provided whenever and anywhere and it can be accessed by the internet. It eliminates the lack of
mobility in all areas where the storage device is carried. In large capacity data pools, it can provide resources on the network for users as needed. Prominent cloud infrastructures could be found in application engines such as Google, Amazon,
Microsoft Azure etc., As mentioned in previous sections, MANETs are analogous to cloud computing networks of proposed
application in this paper. Cloud does not require any predefined infrastructures similar to mobile ad hoc networks [8–12].
The intrusion detection and management systems proposed in this paper are implemented in the MANET architecture. A
basic MANET architecture is depicted in Fig. 2.
As already known, MANET is a self-organizing, self-configuring capability of bringing together mobile nodes without
wires and hence making it an anti-infrastructure implementation [13]. A MANET initiates the transfer of information in the
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
JID: CAEE
ARTICLE IN PRESS
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
[m3Gsc;October 24, 2017;14:50]
3
form of packets from sources to destination. Each node in the MANET must ensure that it is configured perfectly to start
forwarding the packet to the next node. While they are characterized by numerous merits, MANET implementations are
constrained by changing configurations because the nodes are mobile, and each node in the network has limited power and
memory management capabilities. The architecture shown in Fig. 2 has three major layers, that are, the middleware and
application layers, transport protocol layer and the control layer. The control layer is basically a technology enabling layer
which is divided into local area (LAN), metropolitan (MAN) and wide area (WAN) networks. There is another category known
as the personal area network or PAN which is active to the distance of 8–11 m. Variations in LAN include the Wireless LAN
active from 10 0 0 – 170 0 m. The next layer is the Networking layer which is responsible for defining the self-organizing and
self-configuring protocols. In this layer, an efficient and fast single hop and multi-hop strategy for node are also defined to
node transmission. The last layer is the intermediate layer associated with group communication, memory allocation, and
sharing. MANETs incorporate almost all of state of art technologies like Blue tooth, WIMAX, IEEE 802.11 and Hyper LAN
[14–16].
Among other issues in MANETs, a major challenge is the security of mobile nodes because each node is subjected to
limited physical protection, which is a more obvious issue. Security is also a challenge because a centralized control unit is
lacking due to the infrastructureless environment of the MANET. Because of the above factors, intrusion detection technology cannot guarantee long time secure communication even in the network, especially new hacker technology will appear
every day. Hence, we need an effective and fool proof intrusion detection mechanism for ensuring maximum efficiency and
transmission security of packets in the path between nodes [17]. In order to effectively implement a fool like IDS system,
we must understand different kinds of threats or intrusions of knowledge discussed below. External and internal attacks are
the main categories of the incident intrusions on the network or nodes in the network. As the name suggests, an external
attack is caused by a part of a network that has not been studied. They introduce traffic, which results in overload and
congestion of communication paths, delays the transmission of data from the sources to the destination. Internal attacks are
caused by nodes within the current network, which are called malicious nodes, resulting in disruption or interruption of
normal communication in the channel causing rescheduled path and delay. In depth, the threats or attacks could be classified as break in attack which tries to directly break down the security of the system [18–21]. A penetration attack is an
attempt to obtain data from a cloud through a security mechanism, and a leak attack can cause unauthorized information
or data to be transmitted from the network mechanism. Viruses are usually known attacks, and they try to infect files in the
cloud. Denial of Service (DoS) attacks are quite critical attacks which deny any access or usage of system resources to the
authorized user. Apart from these classification, proactive attacks are factors affecting confidentiality and integrity of data,
while passive attacks pose a threat to data confidentiality. The malicious nodes in the passive attacks extract information
from the network and utilize them in future exploitations or infestations which may cause network damage. The rest of the
paper is organized into section II which presents the findings of the literature survey, section III illustrates the proposed
work followed by the results and discussions of the findings. The proposed work aims to implement an effective fool proof
security mechanism for IDS using a fuzzy set theory of decision making for the cloud storage systems.
2. Related work
With the rapid development of wireless communication technology and protocol, many research results have been found
in the investigation [22,23]. Since the proposed work focuses on development of intrusion detection mechanism for a MANET
in a cloud environment, the literature survey has been limited to contributions in IDS and security issues. The works conducted by Bhosale et al had presented a review on the different types of intrusions, nature of attacks and the evidence
responses could be seen from nodes. The paper has presented a routing misbehavior attack in which attackers exploit perfect cooperation between nodes in the network. A watchdog IDS technique has been presented in the paper from the works
conducted by Marti et al where the influence of the attack is reduced to a minimum amount by enhancing the throughput
through the channel even in presence of the malicious node. Watchdogs observe the nodes of any deviation from normal
behaviors and study the hop of the data from nodes within detecting misbehavior. However, the results show that watchdog
IDS will fail in case of false positives, collisions, and data loss. The results show that the proposed scheme effectively solves
the above shortcomings by using the validation scheme IDS[4] [9]. An improvement in above scheme has been brought
about by adaptive acknowledgement technique proposed by Sheltami et al [7].
In the review work, some research have proposed an intrusion detection scheme using digital signature technology to
solve the problems of data loss and false positives found in previous work [11][14][21]. This is achieved by assuming that
malicious nodes lie between the source and destination nodes and act as intermediate nodes. These malicious nodes divide
the packet before passing it to the next node, and link it to the source node through a forging confirmation. A transmission
is said to be successful if the source receives the acknowledgement signal from the destination within a threshold time or
declared unsuccessful on the other condition. The survey also showed that the development of IDS technology either depends on the network or relies on the host [24]. Network based techniques have a network interface for the management
console, while the latter is software monitoring. Research papers have also dealt with knowledge based techniques where
events are recorded in database and intimated to control in order to avoid a similar event occurs in the future. On the other
hand, behavior based IDS methods [3] generate an alarm to the console manager if any deviation from the normal behavior
of the network is observed. Literature presents several recent techniques like IDSX, where X stands for extended to address
Phantom intrusions and elimination of the intrusion in a minimum period of time [8]. The other prominent techniques proPlease cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
JID: CAEE
4
ARTICLE IN PRESS
[m3Gsc;October 24, 2017;14:50]
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
Fig. 3. Intrusion detection mechanism architecture.
posed in the literature include honesty IDS [12] and neural network based techniques [7,19]. Neural network based methods
utilize watermarking techniques for IDS and the intrusion detection technology based on neural network and the hybrid
technology based on SOM also improve the throughput. Another well-known technique is the SCAN algorithm [15] where
every node in the network scans or monitors other corresponding nodes in its neighborhood for any malicious behavior.
Results indicate the accuracy of 94% in identification of misbehaving nodes. In SCAN based methods, the misbehaving node
is isolated by depriving its access to the network [24,25].
Other techniques for efficient IDS implementation include a multi agent system IDS known as MASID, [6,13] where a collection of predefined agents perform the function of detecting abnormal behavior. In a particular case, if there is not enough
reason to prove malicious behavior, then the local agent is combined in a cooperative manner to justify the abnormal behavior by providing additional information. Other techniques proposed in the literature [10] include OCEAN where the state
of the node is maintained and controlled by the neighbor nodes, CNMR [19,24] emphasizes the coordination of monitoring
node activities and the skeleton of the famous SOM model lying in the artificial neural network structure. The works conducted by Bo Sun et al present a Zone based approach for IDS where the node is identified as a gateway node if a node is
physically connected to another node in a different zone. Internally connected nodes are known as Intra zonal nodes. Since
the entire framework is divided into separate regions, the computation time and cost of IDS are greatly reduced.
The other issues addressed in the literature include security in the management console [25]. Unlike traditional mobile
devices connected to wireless networks via access points, Ad Hoc networks, as a research focus, do not have such access
points, so a fully distributed architecture is formed. This affects the defense mechanism of the network and divides the
nodes into trusted and unauthentic nodes. Since the nodes are mobile in MANET, they can freely join or leave the network,
and whether or not they are notified will lead to changes in network configuration. An additional problem related to compromised nodes [24] is the potential Byzantine failures encountered within MANET routing protocols. A Byzantine failure is
not able to directly observe faulty nodes or behavior of them could never be achieved. In the network transmission process,
the nodes caused by these faults can cause the deviation of the new routing messages. Based on the above findings, this paper proposes a fuzzy clustering algorithm to convert the input into a clear set, so as to classify the trusted and unauthentic
nodes.
3. Proposed work
A basic model of intrusion detection mechanism is shown in Fig. 3 starting with the client and the destination, to understand the different types of intrusion into the database during the identification process.
The functional architecture of the proposed solution is based on Google App Engine(GAE) platform. The platform is used
to develop, host, and data centers managed by Google. GAE is a computer based sand server around the world and linked
by a network. The App Engine applications are easily to be built and maintained. They easily withstand the increasing traffic
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
ARTICLE IN PRESS
JID: CAEE
[m3Gsc;October 24, 2017;14:50]
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
5
Table 1
Fuzzy rules set for the proposed work.
No.
F
P
S
AL
1
2
3
4
5
6
7
8
L
L
L
L
H
H
H
H
L
L
H
H
L
L
H
H
L
H
L
H
L
H
L
H
L
M
M
H
M
H
H
H
loads with increasing data storage requirements. Other cloud platforms include similar offers such as Amazon Web Services
and Azure Services. Therefore, our proposed solution, deployed in GAE, which does not represent a centralized solution
because it is stored in grid computing and scattered around the world. It is based on a client/server architecture based
on the RPC (remote procedure call) of Google Web Toolkit (GWT). Fig. 1 shows the different modules required to invoke a
service. Each service has a small cluster of utility classes and interfaces. Each service has a small class of utility classes and
interfaces, some of which, such as proxy services, are automatically generated transparently through the user. The model of
the utility class is the same as the implementation service, which lets us quickly become familiar with RPC.
A basic clustering model for faulty node detection could be modeled by assuming a set of clustersS = {s1 , s2 ,s3 ,s4 ….sp }
with the compactness which is defined asC, the clustering schema could be obtained as
Cp = s
C
i=1 (
(1)
μi j 2 ||x−v||
μi j 2
where μ denotes the membership function corresponding to the i clusters. The difference term in the denominator denotes
the Euclidean distance function. Eq. (1) denotes the degree of compactness of the given cluster. On the contrary, with the
same assumptions discussed above, the degree of separation of cluster is given as follows:
s
Sp =
i=1
min||vi − vc ||2
c
(2)
In Eq. (2), the parameter vi denotes the center of the cluster. Based on the above determination of compactness and
separation, the proposed algorithm could be summarized as below:
Input: C = {T, N} Rp , T – threshold and N – no of clients
Output: Fault List {F1 , F2 , ….Fn } Np
Procedure
For i = 1: N, ∈ Rp
Generate key using N-1 shares
Apply FCM for every computed Cp , Sp
Check for convergence for all i ∈ R
If computed cluster center vc !=T
Label F1 = F1 ← vc for corresponding Sp
Compute new cluster and apply FCM for every i ∈ N + 1
Check for convergence
If i=N
End if
End
End procedure
4. Results and discussion
In order to evaluate the efficiency of the proposed work, Iris data set has been selected with four sets of repository data,
that are Iris, Dermatology, Breast cancer and Mammographic masses. The simulation has been carried out with varying
node numbers ranging from 20, 40, 60, 80 to 100. The speed has been taken as 12 m/s with a FDSAR routing protocol. The
observed simulation time is 50seconds. The number of attackers has been varied from 1 down to 6. The entire environment
is implemented and tested in an Intel I5, 1.85 GHz processor. Some certain parameters define the efficiency of IDS. The first
parameter is True Positive Rate (TPR) which is the degree of identification of malicious nodes, followed by False Positive Rate
(FPR) which is a measurement of the number of genuine nodes identified precisely. The number of unrecognized malicious
nodes in the system is measured by true negative rates (TNR). The experimental data set is given as an input to the system
where a set of rules are defined from the entries under various instances of input. The membership function used for the
proposed work utilizes a triangular function which is shown in Fig. 4.
A set of rules for generating clear output for a given target, as shown in Table 1.
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
JID: CAEE
6
ARTICLE IN PRESS
[m3Gsc;October 24, 2017;14:50]
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
Fig. 4. Intrusion detection mechanism architecture.
Fig. 5. Simulation of MANET with 14 nodes with 1 attack.
In the above table, the indices L, H and M which respectively denotes Low, High and Medium probabilistic conditions for
executing the input from the nodes. The implementation of MANET with a one attack in the form of a worm hole with 14
nodes is depicted in Fig. 5.
The performance of packet delivery ratio has also been computed and plotted against the number of attacks as depicted
in Fig. 6. It could be seen that the our proposed method out performs the FGSAR protocol as well as the conventional
technique of Bo Sun et al.
The cluster and its effectiveness indicators is illustrated in Fig. 7. The fuzzy clustering module is used to divide the
training dataset into several small clusters as depicted in the figure. These clustering methods are applied to three layer
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
JID: CAEE
ARTICLE IN PRESS
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
[m3Gsc;October 24, 2017;14:50]
7
Fig. 6. Packet delivery vs no. of attacks.
Fig. 7. Cluster number versus validity index for Iris repository.
feed forward neural networks and fuzzy decision modules. 12 input nodes and 6 output nodes are utilized in this work. The
nodes in the input layer and the hidden layer adopt the Sigmod transfer function, and the output layer node adopts the
linear transfer function. A training error of 0.0014 is observed in the work.
The response time of the system towards the proposed algorithm is also an essential parameter which is compared with
K means and subtractive clustering methods and it is found to be minimized by a margin of over 30%, as seen from Table 2.
From Table 2, as we can see, the proposed FCM technology for cloud storage clearly shows a dramatic reduction in the
amount of computation associated with the node, even if the number of nodes increases. The experimental results are given
to the parameter estimator which obtain a true positive rate as high as 98% for a node number of 15 and a false negative
rate of 2%. The increasing node number is with a sample number of 80 of TPR falls at 73% with a FNR of 27%. TPR and FNR
are used further to compute the precision and recall factors for further analysis of accuracy.
It can be concluded that our proposed algorithms implement architectures that follow distributed and collaborative, and
have extensively tested cloud storage and retrieval applications for IRIS data repositories. The malicious nodes have been
identified and computed to obtain the total accuracy as shown in Fig. 8.
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
ARTICLE IN PRESS
JID: CAEE
8
[m3Gsc;October 24, 2017;14:50]
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
Table 2
Performance of response time for various techniques.
No. of Nodes
Algorithm
Response time (s)
15
Proposed
KCM
SCM
7.1
7.8
8.1
50
Proposed
KCM
SCM
11.45
12.65
13.45
80
Proposed
KCM
SCM
21.58
24.85
27.59
100
Proposed
KCM
SCM
44.02
48.52
52.84
Fig. 8. Plot of total accuracy towards various attacks.
With the increasing popularity of wireless communications and the mobility of users from one place to another, temporary network security has become an important part of the research. Ad hoc network security has evolved to be a major
component of research in recent time with increasing utilization of wireless communication. The proposed work has taken
maximum 10 attacks and identification of malicious nodes under light and heavy traffic conditions into consideration which
is simulated and observed. A future scope of the work is focused towards implementing a high number of node structure
and further extensions to VANETS.
5. Conclusion
MANET is a self-organizing and capable network of bringing mobile nodes together without wires and hence making it
a non-infrastructure implementation. An MANET initiates the transfer of information in the form of packets from sources to
destination. Each node in the MANET must ensure that it is configured perfectly to start forwarding the packet to the next
node. While they are characterized by numerous merits, the implementation of MANET is subject to changes of configuration
because the nodes are mobile and every node in the network has limited power and memory management capabilities. This
paper presents the novel FCM technique for an efficient intrusion detection system which is used for wireless networks
in the cloud environment. In our framework, the fuzzy clustering module is used to divide the training dataset into small
clusters, as depicted in above figures. In our proposed algorithm, we apply these clustering methods to three layer feedforward neural networks with fuzzy decision modules. Experimental results show that the proposed work takes 10 attacks
into account, and this paper identifies the simulation and observation of malicious nodes under light and heavy traffic
conditions. Compared with other models, the proposed method in this paper has better performance.
References
[1] Zhang Y, Lee W, Huang Y-A. Intrusion detection techniques for mobile wireless networks. Mobile Netw Appl 2003:1–16.
[2] Cho J, Chen IR. Model based evaluation of distributed intrusion detection protocols for mobile group communication systems. Wireless Pers Commun
2011;60:725–50.
[3] Wu B, Wu J, Dong Y. An efficient group key management scheme for mobile ad hoc networks. Int J Secur Netw 2009;4(1):125–34.
[4] Vydeki D, Bhuvasneswaran RS. Effect of clustering in designing a fuzzy based intrusion detection for mobile adhoc networks. J Comput Sci
2013;9(4):521–5.
[5] Chaudhary PDA. Fuzzy based intrusion detection system in mobile Ad Hoc networks. Int J Comput Appl 2014:45–7.
[6] Zhang S, Wang H, Huang W. Two-stage plant species recognition by local mean clustering and Weighted sparse representation classification. Cluster
Comput, 2017:1–9.
[7] Chopade SS, Mhala NN. A co-operative intrusion detection system in mobile Ad-Hoc network. Int J Comput Appl March 2011;18(6):34–9.
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
JID: CAEE
ARTICLE IN PRESS
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
[m3Gsc;October 24, 2017;14:50]
9
[8] Mitrokotsa A, Komninos N, Douligeris C. Protection of an intrusion detection engine with watermarking in Ad Hoc networks. Int J Netw Secur
2010;10(2):93–106.
[9] Wang Y, Li J, Wang HH. Cluster and cloud computing framework for scientific metrology in flow control. Cluster Comput 2017:1–10.
[10] Huang HC, Chuang YY, Chen CS. Multiple kernel fuzzy clustering. IEEE Trans Fuzzy Syst 2012;20(1):120–34.
[11] Yang X, Liu P. A new algorithm of the data mining model in cloud computing based on web Fuzzy clustering analysis. J Theor Appl Inf Technol
2013;49(1).
[12] Pandeeswari, Kumar G. Anomaly detection system in cloud environment using Fuzzy clustering based ANN. Mobile Netw Appl 2016;21(494).
[13] Wang H, Wang J. An effective image representation method using kernel classification. In: Tools with artificial intelligence (ICTAI), 2014 IEEE 26th
International conference on; 2014. p. 853–8. IEEE.
[14] Gang W, Jinxing H, Jian M, Lihua H. A new approach to intrusion detection using artificial neural networks and fuzzy clustering. Expert Syst Appl
2010;37(9):6225–32.
[15] Oktay U, Sahingoz. Attack types and intrusion detection systems in cloud computing. In: Proceedings of 6th international information security &
cryptology conference; 2013. p. 71–6.
[16] Zhang Q, Liang H, Xing Y. A parallel task scheduling algorithm based on fuzzy clustering in cloud computing environment. Int J Mach Learn Comput
2014;4(5).
[17] Li B. An improved fuzzy C-Means algorithm. Comput Appl Softw 2009;25:21–5.
[18] Bhosale R, Ambedkar RK. A survey on intrusion detection system for mobile Ad-Hoc systems. Int J Comput Sci Inf Technol 2014;5(6).
[19] Akbani RH, Patel S, Jinwala DC. DoS attacks in mobile ad hoc networks: a survey. In: Proc. 2nd int. meeting ACCT; 2011. p. 535–41.
[20] Kang N, Shakshuki E, Sheltami T. Detecting forged acknowledgements in MANETs. In: Proc. IEEE 25th int. conf. AINA; 2011. p. 488–94.
[21] Kakkar L, Mehta G. A review: hadoop storage and clustering algorithms. IOSR J Comput Eng 2016;18(1):23–9.
[22] Rao VS, Vidyavathi DS. Comparative investigations and performance analysis of FCM and MFPCM algorithms on Iris data. Indian J Comput Sci Eng
2010;1(2):145–51.
[23] Solanke GJ, Chandre PR. Literature survey on ids in mobile Ad–Hoc networks. Int J Sci Eng Technol Res 2015;4(1).
[24] Gondaliya1 TP, Singh M. Intrusion detection system for attack prevention in mobile Ad-hoc network. Int J Adv Res Comput Sci Softw Eng 2013;3(4).
[25] Aljawarneh S, Monther A, Muneer BY. Anomaly-based intrusion detection system through feature selection analysis and building hybrid efficient
model. J Comput Sci 2017, http://www.sciencedirect.com/science/article/pii/S1877750316305099.
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
JID: CAEE
10
ARTICLE IN PRESS
[m3Gsc;October 24, 2017;14:50]
M. Chen et al. / Computers and Electrical Engineering 000 (2017) 1–10
Mingming Chen, is the member of Academic Committee of Xiamen Huaxia University and the recipient of Educational Evaluation Expert of Fujian Province.
She is the dean of the information and mechanical Engineering Department of Xiamen Huaxia University. Her research fields are Information Communication Network System; System Development of Big Data; Communication Network Optimization; Information Storage.
Ning Wang, is a doctoral student of Xiamen University and visiting scholar of School of Computer Science, FIU(2016). He was a recipient of Fujian Provincial Higher Education professional leaders(2014) and Educational Evaluation Expert of Fujian Province(2015). His research fields are Data Mining, System
Development of Big Data, Information System Engineering, Cloud Computing.
Haibo Zhou, is the Principal Investigator of one Technology Funded Project of Fujian Province and one Xiamen Science and Technology Bureau. His research
fields are Computer Network, Automatic control, Electronic Information, System Development of Big Data.
Yuzhi Chen, has experience in many engineering projects. He took part in many Technology Funded Projects of Fujian Province and Xiamen Science and
Technology Bureau. His research fields are Electronic Information, Network Optimization, System Development of Big Data.
Please cite this article as: M. Chen et al., FCM technique for efficient intrusion detection system for wireless networks in
cloud environment, Computers and Electrical Engineering (2017), https://doi.org/10.1016/j.compeleceng.2017.10.011
Документ
Категория
Без категории
Просмотров
2
Размер файла
1 528 Кб
Теги
011, 2017, compeleceng
1/--страниц
Пожаловаться на содержимое документа