PERFORMANCE ANALYSIS OF THE NEIGHBOR WEIGHT TRUST DETERMINATION ALGORITHM IN MANETS

International Journal of Network Security & Its Applications (IJNSA) Vol.8, No.4, July 2016
DOI: 10.5121/ijnsa.2016.8403 29
PERFORMANCE ANALYSIS OF THE NEIGHBOR
WEIGHT TRUST DETERMINATION ALGORITHM IN
MANETS
Ali Abu Romman1
and Hussein Al-Bahadili2
1
King Hussein Faculty for Computing Sciences, Princess Sumaya University for
Technology, Amman, Jordan
2
Faculty of Information Technology, University of Petra, Amman, Jordan
ABSTRACT
Mobile ad-hoc networks (MANETs) are susceptible to attacks by malicious nodes that could easily bring
down the whole network. Therefore, it is important to have a reliable mechanism for detecting and isolating
malicious nodes before they can do any harm to the network. One of the possible mechanisms is by using
trust-based routing protocols. One of the main requirements of such protocols is to have a cost-effective
trust determination algorithm. This paper presents the performance analysis of a recently developed trust
determination algorithm, namely, the neighbor-weight trust determination (NWTD) algorithm. The
performance of the algorithm is evaluated through simulation using the MANET simulator (MANSim). The
simulation results demonstrated the reliability and effectiveness of the algorithm in identifying and
isolating any maliciously behaving node(s) in a timely manner.
KEYWORDS
NWTD; trust determination; trust-based routing protocols; malicious node; MANET; MANSim.
1. INTRODUCTION
A MANET is defined as a set of wireless mobile nodes communicate with one another for a
purpose of data exchange without relying on any pre-existing or centralized infrastructure[1].
Early MANET research assumed a friendly and cooperative environment and focused on
problems such as wireless channel access, multi-hop routing, power consumption, while ignoring
any network security issues. Thus, early MANETs are venerable and susceptible to attacks by
malicious nodes that could easily bring the network down. Therefore, network security becomes a
major issue to protect network infrastructure from being attacked by adversary or malicious nodes
and to provide protected communication between nodes in a potentially hostile MANET
environment [2].
MANETs heavily suffer from malicious nodes, which could easily degrade the network stability
by exhibiting one or more of the following behavior: packet drop or delaying, buffer overflow,
battery drained, bandwidth consumption, link break, message tampering or modification or
discarding, route modification, node isolation, stealing information, session capturing, etc.
Therefore, it is important to have a reliable mechanism for detecting and isolating malicious
nodes before they can impair the network. One of the potential mechanisms that has been
developed is the trust-based routing mechanism. In which only trusted nodes are accepted for
forwarding data/control packets, so that each node to be part of the routing table, it should have a
trust above a certain value known as minimum acceptable trust (MAT). The main requirement
and challenge to these protocols is the availability of an appropriate trust determination [3 -5].

30
One of the earliest approaches for trust determination is Marsh’s formalism [6], in which the
outcomes of direct interactions among nodes is used by each node to calculate the trusts of other
nodes in the network. After each interaction, a node considers whether the other node fulfilled its
obligations. If fulfilled, then trust increases otherwise decreases. Based on this formalism, a
number of trust determination algorithms have been developed throughout the years [7,8].
This paper analyzes the performance of a recently developed trust determination algorithm,
namely, the Neighbor-Weight Trust Determination (NWTD) algorithm [9-11], which is based on
the weighted voting concept. In NWTD, each node in the network periodically broadcasts
messages containing the IDs and trusts of its one-hop neighbors. Each node after receiving these
messages from its one-hop neighbors, extracts the IDs and trusts of each node on the message. It
is expected that a node receives different trusts for the same one-hop neighbor from different
nodes. Thus, the receiving node calculates the average of the received trusts using a weighted-
average formula to represent the new trust of the node. The weight here is the weight of the one-
hop neighbors, therefore it is referred to as the NWTD algorithm. The receiving node participates
in the averaging process by giving itself a trust and weight of unity.
The algorithm defines two types of nodes: the master node and the monitoring node. A master
node (hereinafter referred to as master only) is any trusted node in the network that can take the
responsibility of testing new arriving nodes, determines their initial trust, and then broadcasts the
initial trust to its one-hop neighbors. A monitoring node is any node in the network that has the
capacity to detect the malicious behavior of other nodes within its neighborhood, updates and
broadcasts the trust of the detected malicious nodes. The algorithm allows one or more
monitoring nodes to be active at the same time.
The NWTD algorithm is implemented and integrated with the MANET simulator (MANSim)[12]
to evaluate the performance of the algorithm. In particular, this paper uses MANSim to simulate
three different scenarios. The first one simulates a network having no malicious nodes; the second
scenario simulates a network having one monitoring and one malicious node; and the third
scenario simulates a network having one monitoring and two malicious nodes.
This section introduces the general domain of the paper. The rest of this paper is organized as
follows: Section 2 reviews some of the most recent and related on trust determination. Section 3
describes the NWTD algorithm. The implementation of the NWTD algorithm and the simulation
environment used in this paper are described in Section 4. The results and discussions are
presented in Section 5. Finally, in Section 6, based on the simulation results, conclusions are
drawn and a number of recommendations for future work are pointed-out.
2. LITERATURE REVIEW
Many trust determination models have been developed for peer-to-peer systems [13], which are
based on sharing recommendation information to establish trust and reputation. Applying these
models to MANETs faces two main problems; these are significant network overhead due to the
additional information exchanged, and requirement of a trusted third party (or a computationally
expensive public key infrastructure (PKI)), which are against the nature of MANETs. However,
later on, many trust determination models have been developed for MANETs [6, 7].
Ferdous et al. [4] developed a novel scheme for trust management in MANETs, namely, the
Network Trust Management (NTM) scheme that is based on the nodes’ own responsibility of
building their trust level and node-level trust monitoring. Hughes et al. [5] proposed a dynamic
trust-based resources (DyTR) system, which applies a dynamic notion of trust to network
resources. DyTR continuously assesses the trustworthiness of nodes over time based on network
environment. DyTR utilizes a socio-cognitive model of trustfor dynamic trust assessment.

31
Gong et al. [8] presented a routing protocol called secure and energy aware routing protocol
(ETARP) designed for energy efficiency and security for wireless sensor networks (WSN). The
key part of ETARP is to simultaneously consider energy efficiency and trustworthiness of routes.
Simulation results show that ETARP keeps the same security level while achieving more energy
efficiency for packet delivery as compared to previously developed routing protocols.
Sabater & Sierra [14] and Ramchurn et al. [15] developed approaches that are based on Marsh
formalism[6], namely, ReGreT and FIRE, which add reputation information provided by third
parties and knowledge of social structures to arrive at overall trust assessments. However, whilst
powerful, such sophisticated models are not appropriate for MANETs, where resources are
limited and knowledge of social relationships between nodes is unlikely to be accessible. Pirzada
& McDonald [16] developed a trust determination mechanism, where nodes calculate situational
trust according to observed events and then use an aggregated general trust for routing
decisions.Sun et al. [17] presented an information theoretic framework to quantitatively measure
trust. They developed four axioms and based on these axioms, they presented two trust models:
entropy-based and probability-based models, which satisfy all the axioms. Simulations showed
that these models could significantly improve the network throughput as well as effectively detect
malicious behaviors in MANETs.
Sylvia et al. [3] proposed a trust based routing scheme for MANETs under adverse environment
and compare the performance of the proposed scheme against existing schemes. The simulation
results demonstrate that in adverse environment, the proposed scheme improves network
throughput and delay;however, it incurs higher overhead.Cordasco & Wetzel [7] compared the
performance of two MANET routing protocols; these are: the Secure AODV (SAODV) and
Trusted AODV (TAODV), which address routing security through cryptographic and trust-based
means respectively.
A dynamic trust prediction models to evaluate the trustworthiness of nodes were also developed
by K. Haldaret al. [18], Park et al. [19], Saini and Gautam [20], Xia et al. [21], Gowda and
Hiremath [22], and Patil et al. [23].A trust establishment and management framework for
hierarchical wireless sensor networks was developed by Zhang et al. [24], and also a trust
management framework (TMF) for MANETs was developed by Guo et al. [25].
3. THE NWTD ALGORITHM
Trust determination algorithms should resolve three main issues[6]:
a. Determine an initial trust value for any new arriving node, so that it will be either trusted
by other nodes in the network or discarded.
b. Periodically determine/update the trust of all nodes in the network.
c. Broadcast the newly determined trusts to other nodes in the network with minimum
overheads.
Trust determination models are usually used in dynamic and multivariable environment, so that it
is important to define some configuration parameters that should be carefully adjusted at each
node to suit the environment and to optimize the performance of the trust determination model.
These parameters include[6, 10, 11]:
a. Minimum Acceptable Trust (MAT). The minimum trust a node should have in order to be
trusted by other nodes, which is usually between 0 and 1.
b. Trust Update Time (TUT).The minimum duration before updating current trust of any
node on the network.

32
c. Minimum Trustable Participants (MTPs).The minimum number of trustable neighbors
that should participate in determining the trust of any other node in the network. For
example if MTP=3, then for Node i(Ni) to determine the trust of Nj (Ti,j), Ni should get
trusts for Nj from at least 2 other trustable nodes as it already has a trust for Nj.
3.1. Trust Determination
In the NWTD algorithm, each node in the network periodically broadcasts an extended HELLO
(EHELLO) message to its one-hop neighbors. The EHELLO message has the same format of
standard HELLO message usually broadcasted by the network routing protocol and some
additional data appended at the end of the message; these are the number of one-hop neighbors
and the ID and trust of each of the one-hop neighbors.
On the other hand, each receiving node then performs the following tasks:
• Compare the trust of the broadcasting node with the pre-defined MAT;if it is ≥MAT, then
accept message, otherwise discard the message.
• Extract the IDs and trusts of the one-hop neighbors of the broadcasting node that pass the
above test.
• Construct a table listing the one-hop neighbor(s) and the trusts they have for each other.
The node is fully trusted by itself (i.e., Tx,x=1)
• Determine the trust of the one-hop neighbors using the mathematical model described
below.
• Remove from the routing table any node for which the determined trust is <MAT.
• Broadcast the newly determined trusts to all one-hop neighbors and waits for broadcasts
from its neighbors.
A node s(0) calculates the trust of each of its one-hop neighbors as follows:
a. Calculates the average trust of the one-hop neighbor xas follows:
, = , (j=1 to n) (kj≥MTP) (1)
Where Ts(i),s(j) is the trust of s(j) as determined by s(i), where s(i) and s(j) are one-hop
neighbors of s(0); , is the average trust of s(j) as calculated by s(0); kj is the
number of nodes that have trust for s(j), n is the number of the one-hop neighbors of s(j);
and MTP is the minimum trustable participants.
b. Calculates the new trust of s(j) as the sum of the product of the trust of s(j) as determined
by s(i) and the weight of s(i) as determined by s(0), which is mathematically expressed
as:
, = , ∙ , (j=1 to n) (kj≥MTP) (2)
Where , is the weight of s(i) as determined by s(0) for a particular s(j), and it is
calculated as:
, =
,
∑ ,
(i=0 to ki, j=1 to n) (kj≥MTP) (3)

33
The average trusts assist to compute the weights of the nodes that determines the new trust of
their one-hop neighbors, which means that the one-hop neighbors share their ideas before
deciding the trust of any of their neighbors, and the contribution of each neighbor depends on its
weight as determined by the receiving node.
3.2. Master Node
A master node is responsible for testing a new arriving node to determine its initial trust; and pass
the result of this test to its one-hop neighbors. Each node in the network must determine a
masterfor itself by applying a certain criteria on its one-hop neighbors and itself. If a node could
not determine a master for itself, it can be a master for itself. The main properties of a master are
it is one-hop neighbor for both the new node and node searching for a master, and the new node
must trust it.If there is more than one node having the same properties, then the one with the
smallest ID is selected as the master [9].
Let us consider a scenario with four nodes (A, B, C and D) trust each other and trusted by their
one-hop neighbors. They are all in same range except for D, which is in range only with C. New
Node x arrives to the network. It is one-hop neighbor for A, B, C and D as shown in Fig.1. All
four nodes can detect that there is a new node arrive to the network. When detected, each node
will compare its table of one-hop neighbors with the neighbor table received from xin order to
choose a master [10, 11].
Fig.1. Master node selection scenario.
For example, A will choose itself as a master, B and C will choose A as a master and D will
choose C as a master. The reason why D chooses C as a master is that it doesnot see A or B.
However, C will not become master, as it already knows that A will test x. Therefore, it will just
wait for the result and then forward it to D.
In the current version of NWTD, the master performs two types of test on any new arriving node,
namely, the FirstTest and the SecondTest. In the FirstTest, the master randomly chooses a
destination node with a known and reachable route. It informs the destination that test will start.
Then, it sends a route request (RREQ) message for that destination through x. The first RREQ has
intentionally increased sequence number, so that the new node should not reply. If it replies it
indicates a malicious behavior (Test1()). While, the second RREQ has a normal sequence number
and the new node should reply back telling the master that it has a route to the destination.
However, if it doesnot reply, this indicates a malicious behavior (Test2()).
This test replays with different destinations for a number of times, and the success and fail ratios
are calculated. The success ratio (S) is ratio between the number of successful test divided by the
total number of tests, and the fail ratio (F) is the ratio between the number of failed tests divided
by the total number of tests; so that S+F=1. If S of the new node is less than a pre-set value, then
it is announced as a non-trusted node. Fig. 2 outlines the pseudocode for the procedures
FirstTest(), Test1(), and Test2().

34
If the new node made a success then a second phase of testing will begin (SecondTest()). In the
SecondTest, the master sends data packet to some destination via the new node. At the same time,
the master sends the hash value of the transmitted data packet to the destination by using its old
route (not going through the new node).
The SecondTest replays for a number of times. Each time, the destination calculates the hash of
the received data packet and compares it with the hash received from the master. If they match
then it is a successful test, otherwise, it is a failed test. If both phases have completed successfully
then the master announces that the new node is fully trusted. Fig. 3 outlines the pseudocode for
the procedures SecondTest() and Testee(). These tests should be performed periodically between
nodes to ensure stability of the network.
FirstTest()
Failed=0
For (trials<nTests)
If (Test1() == False)
Failed++
End If
If (Test2() == False)
Failed++
End If
If (Failed ≥ 25% of Trials)
BlockNode ();
Else
SecondTest()
End If
Test1()
SendRREQ(IncrSeqNo)
ReceiveRREQ()
If (RREP Received)
Return False;
Else
Return True;
End If
Test2()
SendRREQ(NormSeqNo)
ReceiveRREQ()
If (RREP received)
Return True;
Else
Return False;
End If
Fig. 2. The pseudo code for procedures FirstTest(), Test1(), and Test2().
SecondTest()
packet = GenerateDataPacket()
hash = CalculateHash(packet)
Destination = Testee
OriginalRoute = RoutingTable(Testee)
TestedRoute = RetrieveRouteFrom(RREP)
Send(hash, OriginalRoute)
Send(data, TestedRoute)
Response = ReceiveResponse(Testee)
If (Response == False)
Return False;
Else
Return True;
End If
Testee()
hash = ReceiveHash() // Receive hash from master node
pkt = ReceivePacket()// Receive packet from tested node
If (pkt.IsReceived)
newHash = CalculateHash(pkt)
If (hash == newHash)
Send(True)// Send response to the master node
Else
Send(False)
End If
Else
Send (False)
End If
Fig. 3. The pseudocode for procedures SecondTest() and Testee().
3.3. Monitoring Node
In addition to the master discussed above, NWTD designates as many nodes as possible as
monitoring nodes. The monitoring nodes monitor and classify the behavior of its one-hop
neighbors into positive and malicious behaviors, update the trust of its one-hop neighbors by
upgrading the trust of the positively behaving neighbors and degrading the trust of the
maliciously behaving neighbors, and use the updated trust of its one-hop neighbors in the
forthcoming trust determination procedure described above.

35
The monitoring node updates the trust of its one-hop neighbors using the following simple linear
equation:
Tupdated = α · Tcurrent (4)
Where TcurrentandTupdatedare the trust of the one-hop neighbor before and after the update; and αis
the update factor (α>0, α>1for positively behaving nodes, and α<1for maliciously behaving
nodes). The value of α can be determined dynamically by the monitoring node for each of its one-
hop neighbors separately; which means different values ofαcan be determined for different nodes
at the same time depending on their behavior.
It is important to recognize that any node in the network can act a monitor node as long as it has
the capabilities to monitor the behavior of its neighbors, classify their behaviors into positive and
malicious behaviors and update their trust accordingly. Furthermore, it can be easily seen that the
trust of the node depends on the number of one-hop neighbors, the trusts of the node as
determined by its one-hop neighbors, and the weight of the one-hop neighbors.
4.IMPLEMENTATION AND SIMULATION ENVIRONMENT
In order to evaluate the performance of the NWTD algorithm, it is implemented and integrated
with MANSim, which is a network simulator written with C++ programming language for
evaluating the performance of various MANETs protocols[12]. In particular, three main functions
are developed and integrated with MANSim, these are:
a. TrustDetermination(), where the trust determination model described in Section 3.1 is
implemented.
b. InitialTrust. As it has been discussed in Section 3.2 that the master is responsible for
testing and determining the initial trust of any new arriving node, and broadcast this
initial trust to its one-hop neighbors. This function simulates the function of the masters,
where it determines the trust of the one-hop neighbors for each of the nodes in the
network using the following initial trust distribution function:Ti,j = 0.5 + 0.5ξ, whereTi,jis
the trust of Njas determined by Ni, andξis a randomnumber between 0 and 1.
Accordingly, the initial trust lies between 0.5 and 1. However, Ti,jcan be determined using
any other linear or non-linear function.
c. TrustUpdate(), where the trust update procedure described in Section 3.3 is implemented.
4.1.Simulation Environment
The simulation environment that will be used throughout this paper simulates a network area of
150x150m with 49 nodes distributed across the network using the semi-regular node distribution
of MANSim [12]. Each node has a transmission radius (R) of 30m. The actual nodes locations
across the network are shown in Fig. 4. This distribution is chosen because it is easy to predict the
NWTD performance variation according to the nodes behavior. Furthermore, using the same
nodes distribution keeps the focus on the effect of nodes behavior. Furthermore, the nodes are
assumed to be fixed (non-mobile with speed u=0) throughout the simulations.
The main attributes of the nodes (e.g., location, R, u, initial trust, and simulation time
(Tsim))are determined. The simulation time is divided into discrete intervals (loops). During each
loop, the nodes calculate the new trusts for its one-hop neighbors. In addition, the monitoring
nodes may update (degrade or upgrade) the trust of its one-hop neighbors based on their behavior.
Therefore, we refer to the number of discrete intervals as number of updates (U). In this work,
each simulation carries 25 updates (U=25), which is enough to demonstrate the trust variation.

36
Fig.4. Nodes distributed in 150x150m network area for all scenarios.
5.RESULTS AND DISCUSSIONS
In order to evaluate the performance of the NWTD algorithm in identifying and isolating a
malicious node, the MANSim simulator has been used to simulate three different scenarios
considering the simulation environment described above.
Scenario #1 simulates a neighborhood that confined no malicious nodes. It assumes all nodes
across the network are behaving positively (i.e., αi,j=1). The trust of the one-hop neighbors of N25
(T25,x) (e.g., Nodes: 17, 18, 19, 24, 26, 31, 32, 33) are estimated and plotted in Fig. 5.The results
demonstrate that after an initial fluctuation for few updates, the trust of the nodes almost
stabilizes and none of the nodes introduces any changes in the trusts of its one-hop neighbors.
Fig.5. Trusts of some neighbors of N25 (T25,x) (αi,j=1).
1
2
3
4
5
6 7
8 9 10
11
12
13
14
15 16
17 18
19 20
21
22
23 24
25 26
27 28
29 30
31 32
33 34 35
36 37 38
39 40 41 42
43 44 45
46 47
48
49
0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1 3 5 7 9 11 13 15 17 19 21 23 25
Trust
Trial
17 18 19 24 25 26 31 32 33

37
Scenario #2 simulates a neighborhood that confines one monitoring node (N25) and one malicious
node (N31). It assumes all nodes across the network are behaving positively (αi,j=1) except for
N31, which presents some malicious behavior as recognized by the monitoring node (N25). The
trust of some neighbors of N25 (T25,x) are estimated considering two different update factors
(α25,31) of 0.9and 0.7. The results obtained for these two different cases are shown in Figs. 6 and
7.
In this scenario, after an initial fluctuation for few updates, the trusts of the nodes almost stabilize
with some decreasing rates for all nodes. The reduction rate depends on the nodes locations with
respect to N25 and N31, and the update factor. For example, after 25 updates, N18has a trust of
78.6%, 74.8%, and 68.6% for α25,31 1.0, 0.9, and 0.7, respectively. This means that the trusts of
N18 decreases as α25,31 increases. This is because the reduction in the trust of N31 reduces its
weight and consequently the trust of any nodes received through it. Furthermore, it has been
found that if MAT=0.5, N25 requires around 50 updates to isolate N31 with α25,31=0.6 and around
40 updates with α25,31=0.5.
Fig.6. Trusts of one-hop neighbors of N25 (T25,x) (α25,31=0.9)
Fig. 7. Trusts of some neighbors of N25 (T25,x) (α25,31=0.7).
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1 3 5 7 9 11 13 15 17 19 21 23 25
Trust
Trial
17 18 19 24 25 26 31 32 33
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1 3 5 7 9 11 13 15 17 19 21 23 25
Trust
Trial
17 18 19 24 25 26 31 32 33

38
Scenario #3 simulates the same network configuration in Scenario #1. However, it assumes
different nodes behavior in which two of the first hop-neighbors of N25 are acting maliciously,
namely, N31 and N26as detected by N25. Due to their malicious behavior, N25 reduces their trusts
by a certain percentage. In particular, we shall investigate two different cases, in the first, the trust
of N31=N26=0.8 (α25,31=α25,26=0.8)and, in the second,0.7 (α25,31=α25,26=0.7).The variation of T25,31
and T25,26 for the first and second cases are shown in Figs.8 and 9. The figures also show the
variation of T25,31 and T25,26 for a25,31=0.8 and a25,31=0.7 (one malicious node) for the sake of
comparison between Scenarios #2 and #3.
Scenario #3 demonstrates that with two malicious nodesless converges time or number of updates
is required to isolate malicious nodes. In this scenario,N25 reduces the trust of N31 and N26 each
update, and consequently reduce the trust and weight of their neighbors (N32). Because of that,
these neighbors contribute to further reduction in the trust of the N31 and N26. For example, after
25 updates, T25,31 is equal to 0.626 when only N31 acting maliciously (α25,31=0.7), and equal to
0.551 when both N31 and N26 are acting maliciously (α25,31=α25,26=0.7). Also, T25,26 is equal to
0.712 when only N31 acting maliciously (α25,31=0.7), and equal to 0.524 when both N31 and N26
are acting maliciously (α25,31=α25,26=0.7).
Fig. 8. Variation of T25,31 for various α25,31 and α25,26
Fig. 9. Variation of T25,26 for various α25,31 and α25,26.
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
1 3 5 7 9 11 13 15 17 19 21 23 25
Trust(T25,31)
Number of updates (U)
α25,31=0.8
α25,31=0.7
α25,31= α25,26=0.8
α25,31=α25,26=0.7
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1 3 5 7 9 11 13 15 17 19 21 23 25
Trust(T25,26)
Number of updates (U)
α25,31=0.8
α25,31=0.7
α25,31= α25,26=0.8
α25,31=α25,26=0.7

39
6. CONCLUSIONS
The NWTD developed algorithm is an efficient and effective algorithm that can be reliably used
to determine the trust of mobile nodes in MANETs, and consequently identify and isolate any
malicious node. The simulation results demonstrate that for fully trusted network (no malicious
node within the network), the nodes determine a certain trust for each other and the trust between
the nodes remain steady as long as the monitoring node does not detect any malicious behavior
from any node within the network (Scenario #1). However, if a malicious behavior is detected by
the monitoring node for one (Scenario #2) or more nodes(Scenario #3) within the network, then
the trust of all interactive nodes are negatively affected (reduced) and those of malicious behavior
will be affected more, so that can be quickly detected and isolated.
The simulation results of Scenario #3 show that the algorithm can handle efficiently more than
one malicious node at the same time by the same monitoring node, and can use different update
factors for each malicious node. Furthermore, as we discussed above that smaller update factor
can help with quicker identification and isolation of a malicious node.
REFERENCES
[1] C. S. R. Murthy and B. S. Manoj. Ad Hoc Wireless Networks: Architectures and Protocols. Prentice
Hall PTR, 2004.
[2] H. Yang, H. Luo, F. Ye, S. Lu, and L. Zhang. Security in Mobile Ad Hoc Networks: Challenges and
Solutions.IEEE Transactions on Wireless Communications, Vol. 11, No. 1, pp. 38–47, 2004.
[3] D. Sylvia, J. Katiravan, and D. Srinivasa Rao. Trust based Routing in Wireless Ad Hoc Networks
under Adverse Environment. International Journal of Computer Applications (IJCA), Vol. 136,
No.10, pp. 23-28, 2016.
[4] R. Ferdous, V. Muthukkumarasamy, and A. Sattar. A Node-based Trust Management Scheme for
Mobile Ad-Hoc Networks. Proceedings of the 4thInternational Conference on Network and System
Security (NSS), pp. 275–280, 2010.
[5] T. Hughes, P. D, J. Denny, P. A. Muckelbauer, P. D, and J. Etzl. Dynamic Trust Applied to Ad Hoc
Network Resources.Proceedings of Autonomous Agents & Multi-Agent Systems Conference, 2003.
[6] S. P. Marsh. Formalising Trust as a Computational Concept.PhD Thesis, Department of Computing
Science and Mathematics, University of StirlingUSA, 1994.
[7] J. Cordasco and S. Wetzel. Cryptographic Versus Trust-based Methods for MANET Routing
Security. Electron. Electronic Notes in Theoretical Computer Science, Vol. 197, No. 2, pp. 131–140,
2008.
[8] P. Gong, T. M. Chen, and Q. Xu. ETARP: An Energy Efficient Trust-Aware Routing Protocol for
Wireless Sensor Networks. Journal of Sensors, Vol. 2015, Article ID 469793, 10 pages, 2015.
[9] K. El-Zayyat, H. Al-Bahadili, and T. Zobai. A Novel Neighbor Weight-Based Trust Determination
Model for Wireless Ad Hoc Networks. Proceedings of the International Conferenceon Theoretical and
Mathematical Foundationof Computer Science (TMFCS-10), No. 10, pp. 116–126, Jul. 2010.
[10] A. Abu Romman. Evaluating the Performance of the Novel Neighbor Weight-Based Trust
Determination Algorithm in Wireless Ad Hoc Networks. Master Thesis, Princes Sumaya University
of Science and Technology, Faculty of Information Technology. Amman, Jordan, 2013.
[11] H. Al-Bahadili. Trust Determination in Wireless Ad Hoc Networks. Book Chapter in Handbook of
Research on Threat Detection and Countermeasures in Network Security (Editors: A. H. Al-Hamami
and Gh. M. W. al-Saadoon), Chapter 18, pp. 330-348, 2015.
[12] H. Al-Bahadili. On the Use of Discrete-Event Simulation in Computer Networks Analysis and
Design.In Handbook of Research on Discrete-Event Simulation Environments: Technologies and
Applications (Eds.: Evon M. O. Abu-Taieh and Asim A. El-Sheikh). Information Science Reference,
Chapter 19, pp. 418-442, 2010.
[13] S. Song, K. Hwang, R. Zhou, and Y.-K. Kwok. Trusted P2P Transactions with Fuzzy Reputation
Aggregation. IEEE Internet Computing, Vol. 9, No. 6, pp. 24–34, 2005.

[14] J. Sabater and C. Sierra. Reputation and Social Network Analysis in Multi
Proceedings of the 1st International Joint Conference on Autonomous Agents and Multi
Systems (AAMAS'02): Part 1, pp. 475
[15] S. D. Ramchurn, D. Huynh, and N. R. Jennings. Trust in Multi
Engineering Review, Vol. 19, No. 1, pp. 1
[16] A. A. Pirzada and C. McDonald. Establishment in Pure Ad
Communications, Vol. 37, No. 1
[17] Y. L. Sun, S. Member, Z. Han, and K. J. R. Liu. Information Theoretic Framework of Trust Modeling
and Evaluation for Ad Hoc Networks. IEEE Journal on Selected Area in Communications, Vol. 24,
pp. 305–317, 2006.
[18] K. Haldar, N. Narayan and B. K. Mishra. Mathematical Model on Selfishness and Malicious Behavior
in Trust based Cooperative Wireless Netwo
Information Security, Vol. 10, pp. 15
[19] S.-S. Park, J.-H. Lee, and T.
Networks.Proceedings of the 3rdInternational Conference on C
Technology (ICCIT ’08),Vol. 1, pp. 526
[20] R. Saini and R. K. Gautam. Establishment of Dynamic Trust among Nodes in Mobile Ad
Network. Proceedings of 2011 International Conference on Computational Intellige
Communication Networks (CICN), pp. 346
[21] H. Xia, Z. Jia, X. Li, L. Ju, and E. Sha. Trust Prediction and Trust
Ad Hoc Networks. Journal of Ad Hoc Network, Vol. 11, Issue 7, pp. 2096
[22] S. R. Gowda and P. S. Hiremath. Secure Routing Schema for MANET with Probabilistic Node
NodeForwarding. International Journal of Computer Science Issues (IJCSI), Vol. 10, No. 3, pp. 51
58, 2013.
[23] K. Patil, P. Bhanodia, and S. Joshi. A Novel Paradigm RIP
MANET.International Journal of Engineering Research & Technology (IJERT), Vol. 2, No. 1, pp. 1
5, 2013.
[24] J. Zhang, R. Shankaran, M. A. Orgun, V. Varadharajan, and A. Sattar. A Dynamic Trust
Establishment and Management F
IEEE/IFIP 8th International Conference on Embedded and Ubiquitous Computing (EUC), pp. 484
491, 2010.
[25] J. Guo, A. Marshall, and B. Zhou. A New Trust Management Framework for Detecting Malic
Selfish Behaviour for Mobile Ad Hoc Networks. Proceddings of 2011 IEEE 10thInternational
Conference on Trust, Security and Privacy in Computing and Communications (TrustCom), pp. 142
149, 2011.
AUTHORS
Ali Abu Romman is a Digital Forensics Investigator at Jordan Anticorruption
Commission since 2014. He received his MSc. degree in Information Systems
Security and Digital Criminology from Princess Sumaya University for Technology
(PSUT) in 2014. He received his BSc. deg
University of Petra in 2008. His professional experience includes Networking,
Telecommunications, ISP and IPTV. He holds several international certificates in
Security and Networking including CEHv9, CCFP, CHFIv8, OSForensi
EnCase, CCNA, and CCNP and ToT. His research interests include network security
and digital forensics.
Hussein Al-Bahadili is an associate professor at
Universityof Petra, Jordan. He received his B.Sc. in Engineering from the University of
Baghdad in 1986.He received his M.Sc.
(Queen Mary College) in 1988 and 19
networks, routing protocols optimiz
cryptography and network security, and data compression
J. Sabater and C. Sierra. Reputation and Social Network Analysis in Multi-Agent Systems.
International Joint Conference on Autonomous Agents and Multi
Systems (AAMAS'02): Part 1, pp. 475–482, 2002.
S. D. Ramchurn, D. Huynh, and N. R. Jennings. Trust in Multi-Agent Systems. The Knowledge
Engineering Review, Vol. 19, No. 1, pp. 1-25, 2005.
A. A. Pirzada and C. McDonald. Establishment in Pure Ad-Hoc Networks. Wireless Personal
Communications, Vol. 37, No. 1-2, pp. 139–168, 2006.
Y. L. Sun, S. Member, Z. Han, and K. J. R. Liu. Information Theoretic Framework of Trust Modeling
K. Haldar, N. Narayan and B. K. Mishra. Mathematical Model on Selfishness and Malicious Behavior
in Trust based Cooperative Wireless Networks. International Journal of Computer Network and
Information Security, Vol. 10, pp. 15-22, 2015.
H. Lee, and T.-M. Chung. Cluster-Based Trust Model against Attacks in Ad
Networks.Proceedings of the 3rdInternational Conference on Convergence and Hybrid Information
Technology (ICCIT ’08),Vol. 1, pp. 526–532, 2008.
R. Saini and R. K. Gautam. Establishment of Dynamic Trust among Nodes in Mobile Ad
Network. Proceedings of 2011 International Conference on Computational Intellige
Communication Networks (CICN), pp. 346–349, 2011.
H. Xia, Z. Jia, X. Li, L. Ju, and E. Sha. Trust Prediction and Trust-Based Source Routing in Mobile
Ad Hoc Networks. Journal of Ad Hoc Network, Vol. 11, Issue 7, pp. 2096-2114, 2013.
Gowda and P. S. Hiremath. Secure Routing Schema for MANET with Probabilistic Node
NodeForwarding. International Journal of Computer Science Issues (IJCSI), Vol. 10, No. 3, pp. 51
K. Patil, P. Bhanodia, and S. Joshi. A Novel Paradigm RIP (Reputation Index Protocol) for
MANET.International Journal of Engineering Research & Technology (IJERT), Vol. 2, No. 1, pp. 1
J. Zhang, R. Shankaran, M. A. Orgun, V. Varadharajan, and A. Sattar. A Dynamic Trust
Establishment and Management Framework for Wireless Sensor Networks. Proceedings of the 2010
IEEE/IFIP 8th International Conference on Embedded and Ubiquitous Computing (EUC), pp. 484
J. Guo, A. Marshall, and B. Zhou. A New Trust Management Framework for Detecting Malic
Conference on Trust, Security and Privacy in Computing and Communications (TrustCom), pp. 142
is a Digital Forensics Investigator at Jordan Anticorruption
Commission since 2014. He received his MSc. degree in Information Systems
Security and Digital Criminology from Princess Sumaya University for Technology
(PSUT) in 2014. He received his BSc. degree in Software Engineering from
University of Petra in 2008. His professional experience includes Networking,
Telecommunications, ISP and IPTV. He holds several international certificates in
Security and Networking including CEHv9, CCFP, CHFIv8, OSForensics, FTK,
ToT. His research interests include network security
is an associate professor at Faculty of Information Technology,
. He received his B.Sc. in Engineering from the University of
M.Sc. and PhD degrees from University of London
and 1991. His research interests include computer
, routing protocols optimizations, parallel and distributed computing,
work security, and data compression.
40
Agent Systems.
International Joint Conference on Autonomous Agents and Multi-Agent
Agent Systems. The Knowledge
Hoc Networks. Wireless Personal
Y. L. Sun, S. Member, Z. Han, and K. J. R. Liu. Information Theoretic Framework of Trust Modeling
K. Haldar, N. Narayan and B. K. Mishra. Mathematical Model on Selfishness and Malicious Behavior
rks. International Journal of Computer Network and
Based Trust Model against Attacks in Ad-Hoc
onvergence and Hybrid Information
R. Saini and R. K. Gautam. Establishment of Dynamic Trust among Nodes in Mobile Ad-Hoc
Network. Proceedings of 2011 International Conference on Computational Intelligence and
Based Source Routing in Mobile
Gowda and P. S. Hiremath. Secure Routing Schema for MANET with Probabilistic Node-to-
NodeForwarding. International Journal of Computer Science Issues (IJCSI), Vol. 10, No. 3, pp. 51-
(Reputation Index Protocol) for
MANET.International Journal of Engineering Research & Technology (IJERT), Vol. 2, No. 1, pp. 1-
J. Zhang, R. Shankaran, M. A. Orgun, V. Varadharajan, and A. Sattar. A Dynamic Trust
ramework for Wireless Sensor Networks. Proceedings of the 2010
IEEE/IFIP 8th International Conference on Embedded and Ubiquitous Computing (EUC), pp. 484–
J. Guo, A. Marshall, and B. Zhou. A New Trust Management Framework for Detecting Malicious and
Conference on Trust, Security and Privacy in Computing and Communications (TrustCom), pp. 142–

PERFORMANCE ANALYSIS OF THE NEIGHBOR WEIGHT TRUST DETERMINATION ALGORITHM IN MANETS

More Related Content

What's hot (15)

Similar to PERFORMANCE ANALYSIS OF THE NEIGHBOR WEIGHT TRUST DETERMINATION ALGORITHM IN MANETS (20)

Recently uploaded (20)

PERFORMANCE ANALYSIS OF THE NEIGHBOR WEIGHT TRUST DETERMINATION ALGORITHM IN MANETS