A Study on Security in Wireless Sensor Networks

@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 2 | Issue – 4 | May-Jun 2018 Page: 477
ISSN No: 2456 - 6470 | www.ijtsrd.com | Volume - 2 | Issue – 4
International Journal of Trend in Scientific
Research and Development (IJTSRD)
International Open Access Journal
A Study on Security in Wireless Sensor Networks
1
Rashmi, 2
Sumit Dalal, 3
Shabnam Kumari
1
M.Tech Scholar, 2,3
A.P
1,2
Dept of ECE, 3
Department of CSE,
1,2,3
Sat Kabir Institute of Technology & Management, Bahadurgarh, Haryana, India
Abstract- Wireless Sensor Networks (WSNs) present
myriad application opportunities for several
applications such as precision agriculture,
environmental and habitat monitoring, traffic control,
industrial process monitoring and control, home
automation and mission-critical surveillance
applications such as military surveillance, healthcare
(elderly, home monitoring) applications, disaster
relief and management, fire detection applications
among others. Since WSNs are used in mission-
critical tasks, security is an essential requirement.
Sensor nodes can easily be compromised by an
adversary due to unique constraints inherent in WSNs
such as limited sensor node energy, limited
computation and communication capabilities and the
hostile deployment environments.
Keywords- WSN, WTE, SN, FN, MAC
I. INTRODUCTION
Wireless sensor network (WSN) consists of a large
number of spatially distributed autonomous sensor
nodes working cooperatively to monitor the
surrounding physical phenomena or environmental
conditions (monitored target) and then communicate
the gathered data to the main central location through
wireless links. A sensor node, also known as mote is
defined as a small, low-powered, wireless device,
capable of gathering sensory information, perform
limited data processing and transmit the gathered
information to other nodes in the network via optical
communication (laser), radio frequencies (RF) or
infrared transmission media. A sensor node senses
physical phenomena like light, temperature, humidity,
pressure, chemical concentrations and any other
phenomenon capable of cau-sing the transducer
respond to it. Once the phenomena is sensed, the data
collected (measurement) is converted into signals for
further processing to reveal some characteristics
pertaining the phenomenon from the target area
(Hussain, et al., April, 2013). Several schemes for
malicious nodes detection and isolation in WSN have
been proposed. This research explores and improves
one of them, the Weighted Trust Evaluation (WTE)
Scheme. WTE is a lightweight algorithm use in a
three-layer hierarchical network architecture
consisting of low-powered Sensor Nodes (SN) having
limited capabilities, higher-powered Forwarding
Nodes (FN) which collect data from the lower layer
(SNs) and the Base Stations (BS) or Access Points
(AP) layer that route information between the wireless
sensor network (WSN) and the wired infrastructure.
Weighted Trust Evaluation Scheme is based on
several assumptions i.e. both Forwarding Nodes
(FNs) and Base station (BS) are trusted and won’t be
compromised and that the number of normal working
nodes exceeds the compromised nodes. (Sumathi &
Venkatesan, 2014) [1] Once an adversary gains
control over the BS then it leads to create any possible
attacks in the network. The threat of Forwarding
Nodes being compromised is not considered, a
compromised FN gives an adversary control of all the
sensor nodes under it. In this research, we propose an
enhanced WTE based detection algorithm that aims to
address the drawback of the WTE scheme by
employing STL. The STL will come in handy to
address the threat of the compromised forwarding
nodes and since there are few, issues of congestions
and delays in the network are avoided.
II. LITERATURE REVIEW
Several schemes for malicious node detection and
isolation in WSNs have been proposed. It is critical to
detect and isolate the compromised nodes in order to

International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
avoid being misled by the falsified information
injected by the adversary. Luo et al. [19] have pointed
out that infrastructureless ad hoc networks rarely have
a real defense mechanism against most of the attacks,
including both outsider and insider attacks such as
compromised node attacks. They suggested a system
design like this – if one node is named trusted by
certain number of its neighboring nodes, that
particular node is trusted both locally and globally.
However, since the system uses a minimum number
of trusted nodes it is not so applicable to sensor
networks where the nodes are randomly spread out. In
other words, it is possible that under certain
conditions nodes cannot find the minimum number of
neighboring nodes in order to be named trusted.
Sung & Choi, (2013) [2] Proposed a Dual Threshold
technique for malicious node detection that employs
two thresholds to minimize false alarm rate as well as
improve the detection accuracy. All deployed sensor
nodes do have transmission ranges, ’tr’, and any other
sensor node in close proximity i.e. within the node
transmission range is considered its neighbor. Each
individual sensor node maintains its neighbors’ trust
values to designate their trustworthiness. The sensor
node makes a localized decision based on its own
readings and those of its neighbors taking into account
their trust values. Trust values lie between 0 and 1. If
Tik=0 means node Ni does not trust Nk at all. A node
also has its own trust value, once Tii=0 means the
node is faulty.
(Curiac, et al., 2007) [3] Proposed Auto regression
Technique which is a mechanism that relies on
past/present sensor node values. The sensor node
present value is compared with an estimated value
computed from its own previous values by an
autoregressive predictor placed at the base station.
The two values are compared to check if node
behavior is normal or abnormal. If the variance
between these two values is higher than a set
threshold, the node is regarded malicious.
(Yang, et al., 2007) [4] Proposed SoftWare-based
ATTestation (SWATT) mechanism to authenticate the
embedded device (sensor nodes) memory contents
and detect any falsification or maliciously altered or
inserted code in memory. The verifier send to the
embedded device a randomly generated MAC key,
which then calculates Message Authentication Code
(MAC) value on the whole memory using the
received key and returns the MAC value. The verifier
uses the checksum to verify the memory contents. If
the memory has been maliciously altered by the
adversary then the checksum is false.
(Bao, et al., 2011) [5] Proposed a Trust-Based
Intrusion Detection approach which considers a
composite trust metric derived from both social trust
and quality of service (QoS) trust to identify
malicious nodes in the wireless sensor network. The
cluster head apply intrusion detection in the sensor
nodes to assess the trust worthiness and maliciousness
of the nodes in its cluster. This is achieved by
statistically examining peer-to-peer trust evaluation
results gathered from the different sensor nodes
(Sumathi & Venkatesan, 2014).
(Nidharshini & Janani, December 2012.) [6]
Proposed a Sequential Probability Ratio Testing
(SPRT) to detect duplicate nodes made by an
adversary in the WSN. The attacker can easily capture
and make replicas of unattended nodes and then use
them to take control of the entire network. The base
station is responsible for identifying compromised
nodes by computing the speed of observed sample
nodes and decides which nodes’ speed exceeds the
decided threshold speed, these ones are regarded
malicious.
1. Security Goals for Wireless Sensor Networks
The main objectives of Wireless Sensor Networks
(WSNs) security are as follows:
A. Data Confidentiality
Confidentiality refers to the ability to conceal vital
messages’ content from being disclosed to
unauthorized party or protect the messages against
unintended access. Sensor nodes may exchange or
pass highly sensitive information such as
cryptographic key distribution and it must therefore
remain confidential. This means that it is very crucial
to build a secure communication channel in a sensor
network. Data encryption should also be used to
secure the data being transmitted across the sensor
network.
B. Data Integrity
Data integrity is referred as the ability to assert that
the message was not altered, tampered with or
improperly modified in transit by an adversary. It is
essential to guarantee data reliability.
The sensor network integrity will be compromised
when (Padmavathi & Shanmugapriya, 2009) [7]: A
malicious node in the network injects incorrect and
misleading data. Unstable and turbulent conditions

resulting from the wireless communication channel
causing data damage or loss. (Akykildiz, et al., 2002)
[8].
C. Data Authenticity
Authentication ensures the reliability of the received
message through source identity verification. An
attacker can alter the data packet or even modify the
whole packet stream by introducing extra bogus
packets. Data authentication is therefore needed so
that the recipient node can confirm that the data
actually originates from the claimed sender (correct
source).
D. Data Availability
Availability seeks to ensure that the required network
services are functioning at a desired level of
performance and work promptly in normal situations
as well as in the event of attacks or environmental
mishaps. It implies that the sensor node has the ability
to access and utilize the available resources and that
the network is operational and ready for use to
transmit messages.
E. Data Freshness
This ensures that the transmitted messages are current
and old content (expired packets) are not replayed by
an adversary to either mislead the network or keep the
network resources busy thereby reducing the sensor
network vitality. It is essential especially in shared-
key design strategies that require the keys be changed
over time. (CHELLI, 2015) [9].
F. Secure Localization
Sensors may get displaced during their deployment,
after a certain length of time or after a critical
displacement incident. WSN operations depends on
its ability to automatically and accurately locate each
sensor node in the network after the displacement.
(CHELLI, 2015) [9].
G. Self-Organization
WSN being an ad-hoc network and lacking a fixed
infrastructure for network management requires that
each node be independent and versatile so as to be
able to self-organize and self-heal depending on the
various situations, topology and deployment strategy.
This inherent feature of the sensor network is a great
challenge to WSN security. If self-organization is
absent in a wireless sensor network, an attack or the
risky deployment environment may have dire
consequences. (Padmavathi & Shanmugapriya, 2009)
[7]
G. Time Synchronization
Time synchronization is required by many WSN
applications, it is essential in multi-hop
communication, conservation of node energy
(periodic time sleep) and node localization. Sensor
nodes may wish to determine the network latency of a
packet as it transits between a pair of sensor nodes
(sender-receiver) (Padmavathi & Shanmugapriya,
2009) [7]. Collaborative time synchronization may be
needed by wireless sensor network for tracking
applications.
III. ATTACKS LAUNCH FROM
MALICIOUS SENSOR NODES
Since the wireless sensor networks are set up in
hostile environments, sensor nodes can be
compromised easily by the adversary due to the
resource constraints such as limited memory space,
battery lifetime and computing capability. Detection
and isolation of these compromised nodes is crucial to
avoid being deceived and misguided by falsified data
injected by the attacker.
An adversary can easily launch a range of attacks
against the wireless sensor network through the
compromised (malicious) nodes Some of the attacks
that can emanate from malicious nodes include
sinkhole attacks, black hole attack, wormhole attack,
Sybil attack, HELLO flooding attacks and Denial-of-
Service attacks. (Atakli, et al., 2008)
A. Denial-of-Service attacks
Denial of Service (DoS) attack refers to an explicit
attempt by the adversary to deny the victim
(legitimate user) use or access to all or part of their
network resources (Soomro, et al., 2008) [10]. In a
DoS attack an adversary may destroy or disrupt a
network, the attacker can also overload the network
with bogus requests thereby diminishing the
network’s ability to provide a service (Virmani, et al.,
2014) [11] . These attacks make the sensor node
depletes the battery power and degrade the overall
sensor network performance.
B. Black Hole attack
A malicious node take advantage of routing protocol’s
packet route discovery process vulnerabilities to
advertise itself to other nodes in the sensor network as
having the shortest valid route to the packets
destination node (Y-C & Perrig, 2004) [12]. The
attack modifies the routing protocol so as to channel
traffic through a particular node (malicious node)
controlled by the adversary.

In the route finding process, the source node relays
RREQ (Route Request) packets to intermediate
forwarding nodes to find the best valid route to the
intended packet destination node. Since malicious
nodes do not consult the routing table, they reply
immediately to the source node. ( Das, et al., 2002)
[13] . The source node then assumes that the route
finding process is over, ignores other nodes’ RREP
(Route Reply) messages and selects the route through
the malicious node as the best route to transmit the
data packets to the intended destination. The
malicious node is able to accomplish this by
allocating a high sequence number to the RREP
packet. The source node starts forwarding its packets
to the black hole trusting that they will be relayed to
the destination. The adversary controlling the black
hole may now discard these packets instead of
relaying them to the destination node as stipulated by
the protocol.
C. HELLO Flood attack
A laptop-class adversary with a higher radio
transmission power and range relays routing protocol
HELLO packets to a number of other sensor nodes
within a WSN making them assume the attacker is
their neighbor (Padmavathi & Shanmugapriya, 2009)
[7]. The hello packets recipient sensor nodes are
influenced that the compromised node (adversary) is
within their radio range. These node during data
transmission to the base station may forward packets
to the adversary since they assume it is their neighbor
and are eventually spoofed by the adversary.
Hello flood attack could be mitigated using pairwise
authentication of nodes or by employing geographic
routing protocols. Pairwise authentication enable
sensor nodes verify bi-directionality of a link before
they can construct routes to forward traffic received
over the link. Geographic routing protocols like
Geographic and Energy-Aware Routing allow nodes
discard hello messages received from nodes not
within their communication range in terms of
locations, which nodes broadcast to each other
(Raymond & Midkiff, 2008).
D. Sink hole attack
In a sinkhole attack, the attacker’s main goal is to
allure the traffic from nodes in its close proximity
(neighboring nodes) through a compromised sensor
node. These attacks make the compromised attacking
node look enticing and ideal to be used by the
surrounding neighboring nodes to forward traffic.
(Padmavathi & Shanmugapriya, 2009) [7]
Figure 1: Sinkhole Attack (Alajmi, July 2014)
E. Sybil attack
Sybil attack is an identity-based attack in which an
attacker infects a single node with malicious code that
duplicates the node; presenting multiple identities in
multiple locations to other nodes in the sensor
network. The multiple identities of node degrades the
integrity of data as well as straining the network’s
resources. The Sybil attack decreases the efficiency of
fault tolerant schemes like multipath routing,
distributed storage and topology maintenance
(Padmavathi & Shanmugapriya, 2009).
Figure 2: Sybil Attack (Alajmi, July 2014)
Authentication and encryption schemes can protect a
sensor network from Sybil attacks.
F. Worm Hole attacks
This is an attack in which the packets or their
individual bits are captured at one part of the sensor
network, tunneled over a low latency link to another
location and are then replayed at their destination
location (Hu, et al., 2003). This is usually
accomplished by two distant colluding nodes which
create an impression that the two locations involved
are directly connected even though they are genuinely
distant (Virmani, et., 2014)[11].

Figure 3: Wormhole attack (Abdullah, et al., 2015)
[14]
A wormhole attack involves two distant malevolent
nodes conspiring to understate their inter-node
distance by relaying packets through an out-of-bound
channel which is only available to the adversary.
Some defense strategies against wormhole attacks are
packet leashes which ensures that packets are not
accepted “too far” from their source. Geographical
leashes uses GPS information embedded into the
packet being send whereas temporal leashes uses the
nodes’ clocks timestamps added to the packet.
IV. MALICIOUS NODES DETECTION
TECHNIQUES
Several schemes for malicious node detection and
isolation in WSNs have been proposed.
A. Weighted Trust Evaluation Scheme.
Weighted-Trust Evaluation (WTE) based scheme is a
light-weighted algorithm used to detect and
subsequently isolate compromised (malicious) nodes
by monitoring their reported data in a hierarchical
WSN architecture. (Zhao, et al., March 2013) [15]
(Atakli, et al., 2008) [17] Employed and demonstrated
this method using a three-layer hierarchical sensor
network. The components of the three-layer
hierarchical network architecture are:
[2] Low-power Sensor Nodes (SN) whose
functionalities are limited. SN is in the lowest tier
and does not offer multi-hop routing capacity as in
a traditional flat sensor network. SNs report the
data to its Forwarding Node.
[3] Higher-power Forwarding Nodes (FN) which
collect data from the lower layer (SNs), verify its
correctness, aggregate and forward it to other FNs
or to the upper layer (Base Station).
[4] Base Stations (BS) or Access Points (AP) which
verifies data reported by the FNs as well as
routing data between the wireless sensor network
and the wired infrastructure.
Figure 4: Architecture of the hierarchical WSN
(Atakli, et al., 2008)[17]
This scheme is based on two assumptions; first, the
FNs and Base station are trusted nodes that cannot be
compromised by an attacker since once an adversary
seize control of the BS then they can launch any
possible attack in the sensor network (Sumathi &
Venkatesan, 2014) [1] (Hu, et al., 2009) [16] (Atakli,
et al., 2008) [17]. Another critical assumption is that
the normal nodes (working in proper condition) in the
sensor network exceeds in number the compromised
nodes. Otherwise, the scheme may misidentify normal
node as compromised nodes increasing false
positives. The proposed enhanced WTE intends to
detect and isolate malicious FNs in the sensor network
instead of assuming they won’t be compromised by
adversaries. This aims to cautions all the SNs under a
FN which the attacker can control and manipulate
once it take control of a particular FN.
a. Malicious Nodes Detection
A compromised sensor node provides falsified
information that may wrongly mislead the senor
network. This problem is referred as the Byzantine
problem. A compromised/malicious sensor node can
continuously forward wrong information to the upper
layers. The aggregator (AP or FN) in the upper layer
may compute an incorrect aggregation result due to
the misleading information emanating from the
malicious nodes. This may have disastrous effects to
the decision making process.
WTE scheme models malicious node detection and
isolation in 2 steps;
First, an initial weight Wn is assigned to every sensor
node (SN) in the sensor network. The Forwarding
Node (FN) gathers all the reported data from all the
SNs under it and computes an aggregated result taking
into account each SN weight.
E =∑ =1/∑ =1

Where:
E = FN aggregate result.
Wn = SN assigned weight (Ranging between 0 and 1).
Un = SN output information (Un is usually dependent
on the sensor network application. The output value
may be “true” or “false” or continuous numbers like
in a case of temperature readings).
Figure 5: Weight-based hierarchical wireless sensor
network (Atakli, et al., 2008) [17].
Each SN weight is updated based on the accuracy of
the reported information. The SN weight is updated
for two reasons. First, if a compromised sensor node
continuously forward data that is inconsistent with the
final aggregate decision, its weight is likely to be
reduced by a set weight penalty. If the weight
decreases below a given threshold, then it is identified
as a malicious node. Second, the SN weight determine
how much a sensor node report contribute to the final
aggregate decision. This is meant to lower the effect
of incorrect reports from malicious sensor node.
b. Weight Value Recovery
The SN weight is decreased by a certain penalty value
once it is detected to be reporting falsified data.
However, the false report may be a result of a
temporary communication channel interruption and
the SN is neither malicious nor faulty. The weight
values for such SNs needs to be recovered after the
disturbance rather than keeping these values low
permanently. The SNs that behave correctly thereafter
longer than a set recovery time have their weight
value increased.
B. Stop Transmit and Listen (STL)
The STL scheme employs non-transmission time slots
to detect malicious nodes. Each sensor node have an
inbuilt time limit to stop their data transmissions and
listen for traffic. Once the nodes have been deployed
and they have started sensing the target phenomena,
the sensed data is sent to the base station. After every
few seconds or after a set transmission time, each
sensor node halt their data transmission process and
listens for malicious traffic. If a sensor node transmits
data during the non-transmission time (listening time)
, it is caught by its neighbor nodes in the sensor
network and it is regarded malicious as it exhibits
malicious behavior. If a malicious node doesn’t
transmit data during a non-transmission time slot, it
will still be caught in other frequent non-transmission
times. The malevolent behavior of a malicious node is
broadcasted across the entire sensor network.
(Sathyamoorthi, et al., 2014) [18]. Then every other
sensor network node desist from either forwarding
data to the detected malicious node or accepting from
it.
This technique has some weaknesses such that when
the whole network or a major portion of it stopped
their transmission at a time (during non-transmitting
time) and then resume transmission, congestion and
unwanted delay in the network operations arises
(Sumathi & Venkatesan, 2014) [1].
2. CONCLUSION
One of the important problems that are related to the
use of wireless sensor networks in harsh environments
is the gap in their security. This paper provides
information about various security goals in WSN.
Then paper elaborates the Attacks that launch from
Malicious Sensor Nodes and various techniques
available for the above purpose have been discussed.
References
[1] Sumathi , K. & Venkatesan, D. M., 2014. A
Survey on Detecting Compromised Nodes in
Wireless Sensor Networks. (IJCSIT) International
Journal of Computer Science and Information
Technologies, Volume 5, pp. 7720-7722.
[2] Sung, . Y. L. & Choi, Y.-H., 2013. Malicious
Node Detection Using a Dual Threshold in
Wireless Sensor Networks. Journal of Sensor and
Actuator Networks.
[3] Curiac, D.-I.et al., 2007. Malicious Node
Detection in Wireless Sensor Networks Using an
Autoregression Technique. Athens, Greece, s.n.
[4] Yang, Y., Wang, X., Zhu, S. & Cao, G., 2007.
Distributed Software-based Attestation for Node
Compromise Detection in Sensor Networks.
Pennsylvania, s.n.
[5] Bao, F., Chen, I.-R., Chang, M. & Cho, J.-H.,
2011. Trust-Based Intrusion Detection in Wireless
Sensor Networks. Kyoto, Japan, s.n.
[6] Nidharshini, T. & Janani, V., December 2012..
Detection of Duplicate Nodes in Wireless Sensor
Networks Using Sequential Probability Ratio

Testing. International Journal of Advanced
Research in Computer and Communication
Engineering, 1(10).
[7] Padmavathi, . D. . G. & Shanmugapriya, M. D.,
2009. A Survey of Attacks, Security Mechanisms
and Challenges in Wireless Sensor Networks.
International Journal of Computer Science and
Information Security,, Volume 4.
[8] Akykildiz, . I. . F., Su, W., Sankarasubramaniam,
Y. & Cayirci, E., 2002. A Survey on Sensor
Networks. IEEE Communication Magazine.
[9] CHELLI, K., 2015. Security Issues in Wireless
Sensor Networks:Attacks and Countermeasures.
Proceedings of the World Congress on
Engineering 2015, Volume 1, pp. 1-6.
[10] Soomro, S. A., Soomro, S. A., Memon, A. G.
& Baqi, . A., 2008. Denial of Service Attacks in
Wireless Ad-hoc Networks. Journal of
Information & Communication Technology,
Volume 04, pp. 01-10.
[11] Virmani, D., Soni, A., Chandel, S. &
Hemrajani, M., 2014. Routing Attacks in Wireless
Sensor Networks: A Survey. Bhagwan Parshuram
Institute of Technology, India.
[12] Y-C , H. & Perrig, A., 2004. A Survey of
Secure Wireless Ad Hoc Routing. IEEE Security
and Privacy.
[13] Das, R., Purkayastha, D. B. S. & Das, D. P.,
2002. Security Measures for Black Hole Attack in
MANET: An Approach. Proceedings of
Communications and Computer.
[14] Abdullah, M. I., Rahman, M. M. & Roy, M.
C., 2015. Detecting Sinkhole Attacks in Wireless
Sensor Network using Hop Count. I. J. Computer
Network and Information Security, p. 51.
[15] Zhao, S., Tepe, K., Seskar, I. & Raychaudhuri,
D., March 2013. Routing Protocols for Self-
Organizing Hierarchical Ad Hoc Wireless
Networks. Proceedings of the IEEE Sarnoff
Symposium, Trenton, NJ,
[16] Hu, H. et al., 2009. Weighted trust evaluation-
based malicious node detection for wireless
sensor networks. Int. J. Information and Computer
Security, 3(2), p. 148.
[17] Sathyamoorthi, T., Vijayachakaravarthy, D.,
Divya, R. & Nandhini, M., 2014. A SIMPLE
AND EFFECTIVE SCHEME TO FIND
MALICIOUS NODE IN WIRELESS SENSOR
NETWORK. International Journal of Research in
Engineering and Technology, 03(02)

A Study on Security in Wireless Sensor Networks

More Related Content

What's hot (18)

Similar to A Study on Security in Wireless Sensor Networks (20)

More from ijtsrd (20)

Recently uploaded (20)

A Study on Security in Wireless Sensor Networks