Вы находитесь на странице: 1из 6

2016 2nd International Conference on Next Generation Computing Technologies (NGCT-2016)

Dehradun, India 14-16 October 2016

Comprehensive Analysis of Security Attacks


and Intrusion Detection System in Wireless
Sensor Networks
Shanthi.S E.G.Rajan
Department of CSE PRC Private Limited
MallaReddy College of Engineering and Technology Hyderabad
Hyderabad, India India
shanu_shivak@yahoo.com egrajan@gmail.com

AbstractWireless sensor networks consists of tiny sensor preventing the attacks. The paper has been structured as
nodes deployed in various geographic conditions to gather the follows: Section II gives an overview of security in WSN.
information about the environment. The distributed and Section III provides existing security attacks in WSN. Brief
unattended environment makes the network more prone to introduction of IDS and its techniques is given in Section IV.
security breaches. Rather than many studies discuss many Classification of intrusion detection system and its comparison
potential issues of WSN security and detection mechanisms, is given in Section V and section VI describes the IDS
most of these are discussed in isolation. An attempt has been architecture followed by the challenges in WSN.
made to provide a summary of the major security attacks and
present a comprehensive analysis of various Intrusion II. OVERVIEW OF SECURITY PROBLEMS IN WSN
Detection approaches. In addition, we present the comparison WSNs are more vulnerable to security attacks due its open
of Intrusion Detection systems. environment. Some of the issues involved in security are listed
below [2]:
KeywordsWireless sensor networks;security attacks;Intrusion
Detection;Types of Intrusion detection; A. Limited Hardware
The sensor nodes are very tiny and in the recent trends
there is requirement to increase the lifetime of the nodes by
I. INTRODUCTION decreasing the bandwidth consumed, memory etc.Due to this
The wireless sensor networks consist of tiny sensor nodes limited resources, establishing security among these nodes is a
which are deployed randomly in various environments. These quite challenging tasks.
sensor nodes have limited resources like memory,
B. Wireless Communication
computational capacity and energy. The function of sensor
The communication medium is more expensive and it is
nodes is to assemble the data and send the collected more susceptible to threats like eavesdropping, inserting
information to the base station. These sensor nodes are malicious nodes into the network, flooding etc.Due to the
deployed in an hostile and unattended environment, where the wireless medium, we cant opt for complicated protocols that
nodes are always prone to security attacks. WSN is more require exchange of more information or messages.
susceptible to security breaches due its inherent nature, open C.Hostile Environment
environment and unattended hostile environment, limited Since the sensor nodes are deployed in unattended areas,
resources. Among all the other aspects, Security is the most the hackers can able access the nodes and change the contents.
important threat to the networks. The existing security The nodes are not tamper resistant due to its increasing cost
techniques are infeasible due to its limitations like memory, which also provides an easy means to the attacker to access
energy and access of nodes after deployment. Subsequently, the nodes.
the security aspect is the most challenging issue that deserves D.Aggregation Processing
more attention in the wireless sensor networks [1]. Sensor nodes generally obtain the information from each
Many solutions have been provided to the security sensor and transfer the information to the destination. The
issues such as authentication, key exchange, routing protocols lifetime of the sensor nodes can be increased by reducing the
etc. They could able to prevent the attacks to some extent and communication between the nodes. But this cant be
not eliminate the security attacks totally. One of the probable implemented since the sensor nodes have to communicate to
solutions to deal with the security related issues in wireless perform data processing of sensor nodes.
E. Large Scale Deployment
sensor networks is to make use of the IDS - Intrusion
The present sensor networks use 100s to 1000s of sensors
Detection System. It can play a major role in detecting and in applications. So scalability is an important factor to be

978-1-5090-3257-0/16/$31.00 2016 IEEE 426


2016 2nd International Conference on Next Generation Computing Technologies (NGCT-2016)
Dehradun, India 14-16 October 2016

considered in the future networks. Protocols or the security Figure 2: Sinkhole Attack
algorithms designed should keep in mind the concept of In the Sinkhole attack, the main intention of the attacker
scalability otherwise it is of no use [4]. is to attract all the traffic. The compromised node may listen
for the request for routes in the case of flooding and it tries to
III. SECURITY ATTACKS IN WSN give a false route informing the nodes about the shortest path.
There are many types of attacks possible in WSN. d. Sybil Attack:
Wireless sensor network are more vulnerable to attacks due to The attacker is having multiple identities and having
its hostile environment and broadcast nature. The attacks are duplicates and present in multiple locations. It mainly targets
classified generally as active attacks and passive attacks. the fault tolerant systems like storage, network topology,
Active attacks are dangerous and passive attacks dont modify multipath routing, data aggregation, voting, misbehaviour
the data/information and are passive in nature [3]. detection and fair resource allocation, etc[8].

Figure 3: Sybil Attack


e. Wormhole Attack:
The wormhole attack requires the presence of atleast
malicious nodes. They establish a wormhole link between the
malicious nodes. Then the packets received at one location of
the network are routed to the other location of the network. In
this given figure.4, X and Y are malicious nodes and the
communication is established between the two nodes [8].
Figure 1: Attacks on wireless sensor networks
a. Jamming Attack:
Jamming Attack is caused by interfering the radio
frequency of attacker nodes with the other nodes. This attack
is carried out by the transmission of radio signals. It mainly
causes the Denial of service attack and all the nodes are not
communicating because of the jamming attack and mainly
caused by jammer.
b. Collision Attack:
In this collision attack, whenever the legitimate node is
transmitting data, the attacker hears the transmission and Figure 4: Wormhole Attack
transmits its own signal for producing interferences. Even a f. Hello flood Attack:
collision of single byte can produce error and damage entire The HELLO packet is a beacon usually sent by the new
message. This collision attack is better than jamming attack nodes to the other sensor nodes informing the new route. This
with respect to consumption of power and detection ability. attack uses HELLO messages as a weapon to carry out the
This attack intends at draining the communication channel and task. In this attack, the attacker having high transmission
deprivation of network services. power tramsnits the HELLO messages to the sensor nodes
c. Sinkhole Attack: which are in an isolated large area and the nodes are
influenced that the attacker is a neighbor and starts
transmitting the information. The message is passed via the
attacker and the attacker leaks the message before transmitting
the information to the base station. In this type of attack, the
attacker will usually have good signal strength. The hello
flood attacker can find out by checking the average signal
strength of all the nodes and the node which is having more
signal strength than the average neighbors signal strength is
termed to be the attacker[8].

427
2016 2nd International Conference on Next Generation Computing Technologies (NGCT-2016)
Dehradun, India 14-16 October 2016

time; intrusion type; intention of the intruder; position of the


intruder; nature of the intruder etc.IDS can only be able to
detect the attacks it cannot prevent the attacks. Hence it only
used for detecting the attacks [10][11]. In this paper, various
kinds of Intrusion Detection systems are discussed.
The components of the IDS are generally classified as:
Monitoring Component Analyze the traffic and keep track of
it; Analysis and Detection Tries to detect the strange
behavior in the network; Alarm Component Once threats
identified it raises the alarm [6][7].
Figure 5: Hello Flood Attack
g. Battery Drainage:
In this attack, attacker forces the sensors to remain awake
so that they waste their energy. Because of this large power is
consumed by limited power sensor nodes. After the energy is
exhausted, these sensors stop working and causes Denial of
Services through Denial of Sleep [5].

Table 1: Comparison of Attacks in WSN

Figure 6: Components of IDS


IDS are basically classified into two types based on the
audit data: Host based and Network based. Host based
depends on the application logs for analyzing the attacks and
the network based IDS tries to detect the packets in the
network [11] [12]. The IDS can further be classified into
various types based on the detection techniques like:
Anomaly Based IDS
Signature Based IDS
Specification Based IDS
Cross Layer IDS
Hybrid IDS
Signature Based Detection Systems
IV. INTRUSION DETECTION SYSTEM The signature based IDS system is also known as rule
Security is a major issue in WSN due to its restricted based IDS systems [13]. These IDS have pre-defined rules for
resources like limited resources and vulnerable to physical different security attacks. Any deviation of the network
attacks. Some of the techniques used for security issues behaviour from the pre-defined rules is classified as an attack.
include key management, routing protocols, cryptography and One major advantage of these types of techniques is low false
security mechanisms for specific attacks and various IDS.The positive rate. Signature based Detection systems works well
above listed security mechanisms are not sufficient to identify for known intrusions and it cant able to detect the attacks
various attacks in WSNs. IDS provide efficient and effective having no predefined rules and the recent attacks. Routing
methods to detect the various attacks in WSN. Intrusion attacks and sinkhole attacks are detected by these signature
detection system is used to detect the intruders in a network. based detection systems. Every node observes and works
Intrusion is a second line of defense to save the network together with neighbours. Signature-based IDS are best suited
Intrusion can be defined as an unauthorized activity which is for large sized WSNs in which the operations of network are
performed in a network. Intrusion detection System tries to compromised by added security threats and other attacks.
collect the data from the network and analyses the data Signature-based IDS needs additional resources and more
gathered for the abnormal behavior. Intrusion can be achieved computation power as evaluated against anomaly-based IDS.
collect the data from the network and analyses the data The main drawback of these systems is that they cannot
gathered for the abnormal behavior. Intrusion can be achieved identify the new attacks. The database should be continuously
in two ways either statically or dynamically. IDS provide updated for the attack signatures. Another drawback is that the
much valuable information in the network like the intrusion

428
2016 2nd International Conference on Next Generation Computing Technologies (NGCT-2016)
Dehradun, India 14-16 October 2016

signatures should include all possible discrepancies of the Hybrid based Detection Systems
pertinent attack and it should avoid the intrusive activity. This is composed of anomaly based along with signature
based IDS system. It inherits the basic properties from
Anomaly Based Detection Systems anomaly based as well as signature based IDS system. One
In Anomaly based IDS system [14], identification of detection module verifies the known attacks using signatures
Intrusion can be achieved by analyzing the history of the test and other module monitors the overall network behaviour
signal which is called unsupervised data or by collecting the deviation from normal behaviour. It is the most accurate
training data which is called semi supervised data. The data set detection system with less number of false positive. The major
can be discrete, continues, or multivariable. Intrusions are drawback of the hybrid system is it requires more energy and
identified by providing threshold values; that is, any activity resources. These IDS are mostly deployed in cluster based or
carried out within a threshold value is normal, while any to some extent in hierarchical WSNs; some are used to carry
activity having a value more than threshold is known as an out signature based detection while others used to perform
intrusion. The normal behavior is established by automated anomaly detection in order to reduce the utilization of
training and in this way it could possibly able to detect the resources. Hybrid IDS are appropriate for large and
novel attacks as Intrusions. Anomaly-based IDS are more sustainable Wireless Sensor Networks. It makes use of the
suitable for small-sized WSNs in which only a small number benefits of both the approaches and offers simplicity, high
of nodes exchange information with the base station. In small- safety, and low utilization of energy. The Hybrid IDS could
sized WSN, the change in normal traffic or behavioral change able to detect more attacks and the ratio of detection of false
can be considered as intrusion. The problems of misuse attacks is comparatively less.
detection are overcome by focusing on normal behaviors than
attack behaviors. The problems related with these methods Cross layer IDS
are, such as more false alarms are generated by IDS and it Cross layer IDS can be applied at any layer of TCP/IP
could not be able to identify well-known intrusions. protocol stack. It uses cross layer interface to detect intrusion
at each layer of TCP/IP stack by monitoring, communicating
Specification based Detection Systems and exchanging the information. The main drawbacks of this
Specification based system is a form of anomaly based type of IDS it is expensive in terms of cross layer interface
detection system with slight deviation and it combines the [13] [14].
strengths of misuse and anomaly detection. The normal
behaviour is defined manually with specifications and it tries Reputation (Trust) based IDS
to monitor the behaviour with respect to these constraints. It Wang et al. [15] planned an intrusion detection system for
looks for any abnormal behaviour at the system level. These WSNs that uses the concept of packet marking and heuristic
types of IDS define a legitimate behaviour and when the ranking algorithms is used to identify the possible dangerous
system departs from this model, it can detect an intrusion [13]. nodes within the network. The identity of source packet is
In this way, legitimate but previously unseen behaviours will hidden by encrypting and padding. Packet mark is added to
not cause a high false alarm rate, as in the anomaly detection each packet and the data sink will identify the source packet. It
approach. Also, since it is based on deviations from legitimate could able to measure the dropping ratio of the packets of each
behaviours, it can still detect previously unknown attacks. On sensor node. The heuristic ranking algorithms is used in order
the other hand, the development of detailed specifications by to detect the intruders.
humans can be time-consuming and bare the inherent risk that
certain attacks may pass undetected. Statistical detection based IDS
Ngai et al. [16] proposed a proficient approach for detecting
intrusion in sinkhole assault. This approach initially identifies
the traces of suspected nodes. The proposed system examines
through a network flow graph and subsequently finds the
intruder efficiently. This approach makes use of multivariate
method which is based on chi-square test.

Clustering (Hierarchical) based IDS


In this Hierarchical IDS, the entire network is classified into
clusters with each cluster having a cluster-heads. In each
cluster, cluster-head acts as the head and it behaves as a
central point. The members of the cluster reports to the cluster
head about the malicious nodes and it performs the routing in
the group. Every node has IDS agent running within the node
Figure 7: IDS Models and its duty is to locally monitor and detect all the intrusions.
However, clustered is responsible for both local and global
node for its cluster. It examines the network packet traffic and

429
2016 2nd International Conference on Next Generation Computing Technologies (NGCT-2016)
Dehradun, India 14-16 October 2016

controls the network whenever intrusion in the network has focused on determining the weakest node in the network and
been detected [17].Similarly, the attacks against the cluster then providing strategies to defend that node. The problem
heads should be detected by other cluster-heads. All the with this approach was that there might be multiple intrusions
cluster heads communicate with the central base station to to the WSN and only one of them would be caught by the IDS
form global IDS. while leaving others undetected.
Su, et al. [18] has proposed two approaches inorder to improve
the security aspect for clusters in sensor networks using IDS. V.IDS Architectures in WSN
The first approach is based on the concept of authentication, The wireless sensor networks are susceptible to vulnerable
which avoids the external attacks. Its uses the concept of attacks because of its distributed nature. Since the nodes are
message authentication code (MAC) and is added to each distributed in a random fashion, the chances of tampering the
message. Message authentication code is usually produced as node or attacking the nodes from any direction are easier. So
a key-pair value. Along with the message, node adds a they are more vulnerable to attacks. In order to handle these
timestamp and a MAC are generated by a key-pair or challenges, many possible IDS architectures are available
individually depending on the sender. The security mechanism including standalone IDS, distributed and cooperative IDS and
LEAP is being used and the sender can be verified. The hierarchical IDS [23].
second technique is called Energy-Saving. This method tries
to identify the nodes which are deviating from its actual Standalone IDS
behaviour mainly the Member nodes (MN) and cluster-head In this approach, each node operates as independent IDS and
nodes (CH). Whenever an abnormal behaviour is detected in is responsible for the detection of various attacks. As the name
the network, an encrypted warning message along with the says, standalone IDS, they work separately and do not share
cluster key is broadcasted by CH to restrain this specific node. information with each other. In this approach, the nodes are
The cluster-based model provides more security and consumes capable of executing and running its own IDS.In this network
Less Energy. Data delivery is guaranteed in this approach due infrastructure, malicious activities occurring in the victim
to Centralized routing. node are detected by running the IDS independently on each
and every node[24][25]. Moreover, different nodes collect
Watchdog based IDS different information and each node separately makes the
Roman et al. [19] explains the application of IDSs to static decision based on the information collected on its own , since
WSNs. spontaneous watchdogs have been recommended its a standalone IDS,there is no sharing or updating between
IDS for WSNs. In this approach, the neighbors are monitored the nodes in the network . Even though the nodes belong to the
optimally and few nodes wish to monitor their neighborhood same network information regarding the condition on other
communication independently. nodes will not be known to other nodes. The infrastructure is
not worthy because of its restrictions, since all the nodes need
Distributed and collaborative IDS to be configured with IDS and this network would be more
Krontiris et al. [20] proposed distributed IDS for WSNs based efficient if all the nodes are not configured with IDS. The
on the concept of collaborative neighborhood watching. The intrusion could not be detected based on the information
proposed IDS works better against the black hole and selective collected from every node and the information will not be
forwarding attacks The author[23] presented a solution for the sufficient enough to determine whether an intrusion occurred
problem of cooperative intrusion detection in wireless sensor or not. Therefore, this architecture is best suited only for flat
network. The necessary conditions for exposing the attacker infrastructure network. To conclude with, this network was
have been presented by the author in an efficient way. Nodes not a feasible solution to design the IDS architecture for WSN.
could able to recognize the intruder in a distributed manner by
using the local detector modules. The detector module triggers Distributed and Cooperative IDS
uncertainties about an interruption in the sensors region In Zhang et al., 2003[26], each node detects intrusion by running
[21], the proposed IDS used a specification based detection an IDS agent individually. In general, the job of the IDS agent
algorithm. They used the decentralized approach in which is to identify and collect local events. IDS can monitor only
intrusion detectors were randomly distributed among the their own communication and data to detect possible
network. The information gathering and processing were intrusions as well as it initiates response independently. This
carried out in distributed fashion. Due to scalability and robust IDS is more suitable for a flat network configuration than a
nature, the distributed approach was more useful than the cluster based one.
centralized approach.
VI.CHALLENGES OF IDS IN WSN
Game theory based IDS The IDS have been designed for wired networks and
According to Agah et[22] al.the Game theory IDS consists of it cant fit a wireless sensor networks. Some of the challenges
two participants one is the attacker and the other one the involved in WSN were as follows:
detecting intrusion. He formulated strategies for both parties.
This IDS is based on non-cooperative and non-zero game
model to improve the detection probability. Both schemes

430
2016 2nd International Conference on Next Generation Computing Technologies (NGCT-2016)
Dehradun, India 14-16 October 2016

[7]. Nabil Ali Alrajeh, S. Khan, and Bilal Shams,Intrusion Detection Systems
in Wireless Sensor Networks: A Review, International Journal of
Distributed Sensor Networks Volume 2013, Article ID 167575, 7 pages.
[8]. Yassine MALEH and Abdellah Ezzati a review of security attacks
and intrusion detection schemes in wireless sensor network,
International Journal of Wireless & Mobile Networks (IJWMN) Vol. 5,
No. 6, December 2013
[9]. A Survey of Intrusion Detection Systems in Wireless Sensor Networks,
Ismail Butun, Salvatore D. Morgera, and Ravi Sankar,IEEE
COMMUNICATIONS SURVEYS & TUTORIALS, VOL. 16, NO. 1,
FIRST QUARTER 2014
[10] Sonu Duhan, Padmavati khandnor,.Intrusion Detection System in
Wireless Sensor Networks: A Comprehensive Review, , International
Conference on Electrical, Electronics, and Optimization Techniques
(ICEEOT) 2016
[11]. Manali Singh,Khushbu Babbar, Kusum Lata Jain, A Survey on Intrusion
Detection System in Wireless Sensor Networks, International Journal of
Wireless Communications and Networking Technologies, Vol 3, No.3,
April - May 2014
[12]. Ranjit Panigrahi, Kalpana Sharma, M.K. Ghose,wireless sensor networks
architecture, security requirements,security threats and its
Figure 8: Challenges of IDS in WSN countermeasures, , pdcta 2013 pp. 107115, 2013. cs & it-cscp 2013
[13]. Raymond D R, Midkiff S F. Denial-of-service in wireless sensor
networks:attacks and defenses. IEEE
VII.COMPARISION OF IDS [14]. David R. Raymond, Scott. F. Midkiff Denial of Service in WSN:
Attacks and defences, IEEE pervasive computing, Vol. 1, No. 7, pp.
Characteristics Anomaly Signature Hybrid 74-81, 2008.
[15]. C. Wang, T. Feng, J. Kim, G. Wang and W. Zhang, Catching Packet
Based Based Based Droppers and Modifiers in Wireless Sensor Networks, IEEE Trans.
Memory Low Low Medium Parallel Distrib. Syst., vol. 23, num. 5, pp. 835843, 2012.
[16]. E. Ngai, J. Liu and M. Lyu, On the Intruder Detection for Sinkhole
Utilization Attack in Wireless Sensor Networks, ICC06, Istanbul, Turkey, June
Energy Low Low Medium 2006.
[18]. C.-C. Su, K.-M.Chang, Y.-H.Kuo, and M.- F. Horng, The new intrusion
Consumption prevention and detection approaches for clustering-based sensor
Detection Rate Medium Medium High networks, in 2005 IEEE Wireless Communications and Networking
Conference, WCNC 2005: Broadband Wirelss for the Masses - Ready
False Alarm Medium Medium Low for Take-off,2005.
[19].R. Roman, J. Zhou, and J. Lopez, Applying intrusion detection systems
VIII.CONCLUSION to wireless sensor networks, in Proc. IEEE Consumer Communications
and Networking Conference, 2006.
In many applications, WSN has it widespread usage. In the [20].I. Krontiris, Z. Benenson, T. Giannetsos, F. Freiling and T. Dimitriou,
absence of various security mechanisms, a variety of attacks is Cooperative intrusion detection in wireless sensor networks, Springer
possible in WSN. The main aim of the paper is to present the J. Wireless Sensor Networks, pp. 263-278, 2009.
[21]. A.P. da Silva, M. Martins, B. Rocha, A. Loureiro, L. Ruiz and H.C.
security attacks in WSN and the Intrusion Detection system in Wong, Decentralized Intrusion Detection in Wireless Sensor
WSN. The security problems in WSN, security attacks and the Networks, in Proc. 1st ACM International Workshop on Quality of
most promising Intrusion Detection System have been Service and Security in Wireless and Mobile Networks (Q2SWinet 05),
discussed with a hope that will assist the researchers to carry ACM Press, October 2005, pp. 16-23.
[22].A. Agah and S.K. Das, Preventing DoS attacks in wireless sensor
out further research. networks: A repeated game theory approach, International Journal of
Network Security, volume 5, number 2, pages 145-153,2007.
References [23]. Andreas A. Strikos, A full approach for Intrusion Detection in Wireless
Sensor Networks,School of Information and Communication
[1]. Y. Wang, G. Attebury, B. Ramamurthy, A survey of security issues in Technology , Stockholm, Sweden ,March 1, 2007.
wireless sensor networks, IEEE CommnsSurveys, vol. 8, pp. 2-23, 2006.
[24]. Rassam, Murad A., M. A. Maarof, and Anazida Zainal. "A Survey of
[2]. Y. Zhou, Y. Fang, and Y. Zhang, Securing Wireless Sensor Networks: A
Intrusion Detection Schemes in Wireless Sensor Networks." American
Survey, IEEE Communications Survey,vol. 10, no. 3, pp. 6-28, 2008.
Journal of Applied Sciences 9, no. 10 (2012): 1636.[19]
[3]. G. Padmavath, D. Shanmugapriya, A Survey of Attacks, Security
[25]. Anantvalee, Tiranuch, and Jie Wu. "A survey on intrusion detection in
Mechanisms and Challenges in Wireless Sensor Networks, (IJCSIS)
mobile ad hoc networks." In Wireless Network Security, pp. 159-180.
International Journal of Computer Science and Information Security,
Springer US, 2007.
Vol. 4, No. 1 & 2, 2009.
[26].Y. Zhang and W. Lee, Intrusion detection in wireless ad-hoc networks,
[4] A.S.K.Pathan,H.-W. Lee, and C.S. Hong,Security in Wireless Sensor
Proc. 6th annual international conference on Mobile computing and
Networks: Issues and Challenges, in 8th International Conference on
networking, pp. 275-283, 2000.
Advanced Communication Technology (IEEE ICACT 2006), Volume II,
20-22 February, Phoenix Park, Korea, 2006, pp. 1043-1048.
[5]. I.Onat and A. Miri, An intrusion detection system for wireless sensor
networks, Wireless and Mobile Computing,Networking And
Communications, vol. 3, 2005, pp.253259.
[6]. I. Krontiris, T. Dimitriou, Th. Giannetsos, and M.Mpasoukos, Intrusion
Detection of Sinkhole Attacks in Wireless Sensor Networks, LNCS, vol.
4837, pp. 150-161, 2008

431

Вам также может понравиться