Wireless Sensor Network, 2011, 3, 125-134
doi:10.4236/wsn.2011.34015 Published Online April 2011 (http://www.SciRP.org/journal/wsn)
Copyright © 2011 SciRes. WSN
A Direct Trust Dependent Link State Routing Protocol
Using Route Trusts for WSNs (DTLSRP)
Shaik Sahil Babu, Arnab Raha, Mrinal Kanti Naskar
Department of El ectroni cs a n d Telecommunicat i on E ngineering , Jadavpur University, Kolkata, India
E-mail: {sksahilbabu419, arnabraha1989}@gmail.com, mrinalnaskar@yahoo.co.in
Received February 19, 2011; revised March 3, 2011; accepted March 23, 2011
Abstract
The traditional cryptographic security techniques are not sufficient for secure routing of message from sour-
ce to destination in Wireless Sensor Networks (WSNs), because it requires sophisticated software, hard-
ware, large memory, high processing speed and communication bandwidth. It is not economic and feasible
because, depending on the application, WSN nodes are high-volume in number (hence, limited resources at
each node), deployment area may be hazardous, unattended and/or hostile and sometimes dangerous. As
WSNs are characterized by severely constrained resources and requirement to operate in an ad-hoc manner,
security functionality implementation to protect nodes from adversary forces and secure routing of message
from source node to base station has become a challenging task. In this paper, we present a direct trust de-
pendent link state routing using route trusts which protects WSNs against routing attacks by eliminating the
untrusted nodes before making routes and finding best trustworthy route among them. We compare our work
with the most prevalent routing protocols and show its benefits over them.
Keywords: Wireless Sensor Network (WSN), Geometric Mean (GM), Direct Trust, Indirect Trust, Route
Trust (RT), Base Station (BS), Benevolent Node, Malicious Node, Homogeneous Nodes, Packet
Latency, Packet Transmission Rate, MATLAB
1. Introduction
Trust establishment among nodes is a must to evaluate
the trustworthiness of other nodes and is one of the most
critical issues in WSNs. Trust is dependent on time; it
can increase or decrease with time based on the available
evidence through direct interactions with the node or
recommendations from other trusted nodes. Trust-mod-
eling is mathematical representation of node’s opinion of
another node in a network. We need mathematical tools
to represent trust and reputation, update these continu-
ously. Maintaining a record of the transactions with other
nodes, directly as well as indirectly, from this record a
‘trust’ value will be established [1].
Security and trust are two tightly interdependent con-
cepts. Generally these terms are used interchangeably
when defining a security system [2]. However, security
is different from trust. Security is more complex and the
overhead is high. In other words, security means no one
is trusted and requires authentication all the time which
leads to high overhead, i.e. encryption and decryption
with secret key [3]. Trust means everybody is trusted
somehow and does not require any authentication (less
overhead). It tells the degree of reliability. Every node
finds the trust of all other nodes, based on previous ex-
perience and recommendations in fulfilling its promises.
Trust management system for wireless sensor net-
works (WSNs) is a mechanism that can be used to sup-
port the decision-making processes of the network [4]. It
aids the members of WSN (trustors) to deal with uncer-
tainty about the future actions of other participants (trus-
tees). As WSNs are highly application oriented, these va-
rious applications bring various security needs. In WSN,
sensor nodes have limited communication bandwidth, pro-
cessing resources, memory space and battery capacity [5].
Hence, the trust management system should enable the
WSN to be secure while significantly reducing comput-
ing and communication overheads. The WSNs can be
established without any existing infrastructure, which is
a major feature exploited in most applications, they rely
on the mutual cooperation among nodes to route traffic
towards sink or base station. Hence, trust establishment
among the nodes is a must to evaluate the trustworthiness
of other nodes and is one of the most critical issues in
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
126
WSNs. Survival of a WSN is dependent upon the coop-
erative and trusting nature of its nodes. Hence, the trust
establishment between nodes is must. Trust is dependent
on time; it can increase or decrease with time based on
the available evidence through direct interactions with
the same entity or recommendations from other trusted
entities [2].
Trust aware routing framework for WSNs is proposed
by [6], to secure multi-hop routing in WSNs against in-
truders exploiting the replay of routing information. With
the idea of trust management, their proposal enables a
node to keep track of the trustworthiness of its neigh-
bours and thus to select a reliable route. Their proposal
can also be implemented for large-scale WSNs deployed
in wild environments. Many security attacks have been
presented in ([7,8]) with a significant subset targeting the
routing process [9]. If an adversary force manages to cap-
ture the node, it participates in the network, and it can
damage the routing process by simply dropping the pack-
ets it receives for forwarding. Another attack easy to im-
plement is packet modification. In [10] an approach that
the human society follows proposed to defend against the
majority of routing attacks. Although the design of
mechanisms to enhance security at all layers of the net-
working protocol stack has attracted the interest of the
research community (e.g. [11,12]), very limited imple-
mentation effort has been reported. In [13], the imple-
mentation of a link-layer security architecture is pre-
sented, while in [14] experience regarding the implemen-
tation of hash-based encryption schemes in tinyos oper-
ated sensor nodes is reported. In [15], the efficiency of a
set of routing protocols is compared based on real test-
bed experiments. In [16], very limited information regard-
ing the implementation of a trust model is provided. Fi-
nally, in [17] presented results and experience gained
through the implementation of a location-based trust-
aware routing solution. A distributed trust model is in-
corporated in the routing solution which relies on both
direct and indirect trust information.
In this paper, we present simulated results of a new
link state routing protocol based on trust by eliminating
the malicious nodes from the network (lsrp based on
trust). This protocol incorporates a trust computational
model with direct and indirect experiences based on tra-
ditional weighting approach of the qos characteristics such
as packet forward, data rate, power consumption reliabil-
ity, etc. using beta probability distribution [2]. The trust
management system at the node computes trust table for
network nodes, and then using a threshold find out the be-
nevolent nodes of the network. Then, using link state rout-
ing it finds all available paths by eliminating the mali-
cious nodes.
The rest of this paper is organized as follows: first in
section 2 we present the related work on WSN routing
protocols based on trust and in section 3 the designed lsr
protocol based on direct trust while in section 4 per-
formance evaluation. In section 5 the simulation results
and in section 6 conclusions.
2. Related Work
Trust Based Routing methods: Enhancements in the rout-
ing related protocols based on the trust have been widely
addressed in the literature. The following are the most im-
portant research results in this direction:
2.1. ARIADNE
It is a very efficient protocol, using highly efficient sy-
mmetric cryptographic primitives and per-hop hashing
function [18]. It prevents the attackers or compromised
nodes from tampering with uncompromised routes con-
sisting of uncompromised nodes, and also prevents a
large number of types of Denial-of-Service attacks.
2.2. ATSR (Ambient Trust Sensor Routing)
A fully distributed Trust Management System is realized
in ATSR [5] in order to evaluate the reliability of the
nodes. Using this approach, nodes monitor the behavior
of their neighbors in respect to different trust metrics and
finds direct trust value per neighbor. It also, takes into
account indirect trust information, i.e. trust information
from its neighbors, also called reputation. Direct and in-
direct trust information is combined to reach the Total
Trust information. Finally, the routing decisions are based
on geographical information (distance to the base-station)
and Total Trust information. The trust model presented
has been integrated with a location-based routing proto-
col. If no malicious node exists in the network, i.e. the
Total Trust is almost equal to 1, the ATSR behaves sim-
ply the Greedy Perimeter Stateless Routing (GPSR) pro-
tocol.
2.3. Trusted AODV
It is an extended AODV routing protocol to perform rout-
ing by taking trust metrics into account [19]. First, a trust
recommendation mechanism introduced and then the rout-
ing decision rules of AODV are modified to take trust
into account. A set of policies is derived for a node to up-
date its opinions towards others, because it is necessary
to design a trust information exchange mechanism when
applying the trust models into network applications.
2.4. Trusted GPSR
The Greedy Perimeter Stateless Routing [20] is modified
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
127
to take trust levels of node into account. Each time a
node sends out a packet it waits until it overhears its
neighboring node forwarding it. Based on this correct
and prompt forwarding information it maintains a trust
value for its neighbors. This information is then taken
into account in the routing decisions.
2.5. SPINS (A Suite of Security Protocols
Optimized for Sensor Networks)
This [21] has been designed to provide data authentica-
tion, data confidentiality and evidence of data freshness.
In this protocol two security blocks SNEP and µTESLA
are involved. The first block introduces overhead of 8
bytes and maintains a counter fro achieving semantic se-
curity. µTESLA provides authentication for data broad-
casting. Though SPINS claim to provide trusted routing
ensuring data authentication and confidentiality, but it
does not deal with Denial of Service Attacks.
2.6. Trust- Aware DSR
The watchdog and Pathrater modules has been designed
and incorporated in the Dynamic Source Routing proto-
col for security [22]. The watchdog module is responsi-
ble for detecting selfish nodes that do not forward pack-
ets. For this, each node in the network buffers every trans-
mitted packet for a limited period. During this period
each node enters into promiscuous mode in order to over-
hear whether the next node has forwarded the packet or
not. And based on the feedback that Pathrater receives
from the watchdog, it assigns different ratings to the nodes.
These ratings are then used to select routes consisting of
nodes with the highest forwarding rate.
2.7. CONFIDANT (Cooperation Of Nodes,
Fairness In Dynamic Ad-hoc Networks)
This [23] protocol adds reputation system and a trust ma-
nager to the Watchdog and Pathrater scheme. The trust
manager evaluates the events reported by the Watchdog
and issues signals to other nodes regarding malicious
nodes. The signal recipients are maintained in a friends-
list. The reputation system maintains a black-list of nodes
at each node and shares them with friends-list nodes. In
one way it is a punishment based scheme by not forward-
ing packets of nodes whose trust level drops below the
certain threshold.
2.8. TRANS (Trust Routing for Location Aware
Sensor Networks)
TRANS [24] routing protocol selects routes based on
trust information not on hop count to avoid the insecure
locations. This protocol assumes that the sensors know
their locations and that geographic routing is used. A sink
sends a message only to its trusted neighbors for the des-
tined location. Those corresponding neighbors forward
the packet to their trusted neighbors that have the nearest
location to destination. Thus the packet reaches the des-
tination along a path of trusted sensors. Here the impor-
tant feature of TRANS, the sink identifies misbehavior
by observing replies, probes potential misbehaving loca-
tions and isolates insecure locations. On discovery of
such locations, the sink records and advertises to the
neighboring nodes.
2.9. Traditional Weighting Approach Using Beta
Probability [1]
Momani introduced one algorithm for trust calculation
and risk assessment based on trust factors and dynamic
aspects of trust. As shown in Figure 1, he assumed that
trust is computed using traditional weighting approach of
the QoS characteristics such as packet forward, data rate,
error rate, power consumption, reliability, competence,
etc.
A is direct trust (experience), B is indirect trust (rec-
ommendations), C is total trust.
T is required trust, R is risk, Total trust
,CFAB.
Y
TX means trust at Y on X i.e. (trust that Y is hav-
ing on X).
i
Y
TX means trust at Y on X for ith category.
1,2, ,i
Y
in
A
TX
. Sum of trust values at Y on X
for n different events.
Figure 1. General trust computational model in brief (Tra-
ditional weighting approach using Beta probability distri-
bution).
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
128

1,2, ,Y
Ym
BTXm
. Average of sum of indirect
trust values at Y on m nodes. Trust reported from all the
surrounding nodes that have had previous experience
with the node.
Total trust

,CFAB and it can be


1,2, ,
1,2, ,
i
AB
YA
In
YB
Ym
CAW BW
TXW
TX Wm
 


The weights WA and WB can be assigned using differ-
ent approaches.
CASE 1: Some nodes may be given more weightage in
direct trust; others may be given more weight in recent
indirect trust, i.e WA > WB or WA < WB.
CASE 2: Weights to the direct trusts of some events
may be given more importance, and others are less im-
portance. Similarly, for indirect trusts nearby nodes may
be given more importance and others is less importance
as shown in figure.
 
 
 

12
34
12
34
123
4
+
YY
YY
aab
b
CTXWTXW
TXWTXW
TX WTXWTX W
TX W


 

Risk can be calculated as the difference between trusts
required and total trust.
, and AB
RTCC AWBW
A traditional weighing approach to calculate Trust and
asses Risk (Risk assessment algorithm) is introduced.
These weights WA, WB can be assigned using different
approaches. Some nodes might give more weight to di-
rect trust, others might give more weight to recent indi-
rect trust.
3. A Direct Trust Dependent Link State
Routing Protocol Using Route Trusts
for WSNs (DTLSRP)
Our model uses the Momani’s model of assigning dis-
crete trust values to the sensor nodes. Most of the work
present in the literature establishes different techniques
to calculate trust of individual nodes with respect to its
one hop neighbor (direct trust) or with respect to the one-
hop neighbors of the one-hop neighbor used initially to
measure the direct trust (known as the indirect trust). We
present this concept in brief.
Let us first present a network topography depicting the
location of nodes and the path through which the packet
must be transmitted from one node(which may be called
the source) to the Base Station or another node(which
may be called the destination) as is shown in Figure 2.
This diagram represents a network of 25 nodes deployed
in the form of a square area. Now suppose the source
wants to send a packet to the destination. Let us for our
ease of understanding we first assign some trust values to
the nodes color marked in this diagram. We represent the
same in the form of a table. In this paper we are only
interested the transmission and reception of data packets
and hence our concern will be only in the direct commu-
nication trusts measured particularly on the basis of first-
hand experience of percentage of successful reception of
packets. Suppose node A wants to send packets of data to
B. The trust map is shown in Figure 2. Now in this case,
we are only interested in the direct trust parameters which
node B is having on node A such as percentage of suc-
cessful reception of data packets, latency of transmission,
relative power levels etc. First of all here WB = 0 as we
are not concerned with indirect trusts. So WA = 1 and the
calculation of TB(A) which is the trust value that B is
having on A will only be based upon the direct trust pa-
rameters which is also known as the one-hop trust met-
ric.
The non-inclusion of indirect trusts is obvious from
the fact that as we are only interested in communication
trust that is transmission of data through the network,
such emphasis on indirect trust is redundant more so be-
cause the paths are determined beforehand and a receiv-
ing node is only concerned whether the data it is receiv-
ing from the sender is at all trustworthy or not irrespec-
tive of what other nodes think about it. It also decreases
the memory and processing capability of the nodes.
In the following table we represent some practically
calculated trust values assigned to the 10 nodes as shown
in the network topology described earlier. It is to be noted
in this case that for this table the direct trust value as-
signed to a single node is with respect to its direct one
hop neighbor in the routing path. If a node belongs to
more than one routing path then individual trust values
must be taken in to account. For example, in case of node
6 if two separate paths exist and its 2 next hop neighbors
are respectively 7 and 10 then T7(6) and T10(6) both are
to be calculated for calculation of route trusts of the two
paths. We divide our routing protocol into several steps
as shown in the following Figure 3.
Step1: Calculation of direct trust of the indivudual
nodes
As mentioned previously we are only concerned with
the direct trusts in this case. So if our topology consists
A B
Figure 2. Simple trust map.
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
129
Figure 3. Proposed protocol structure.
of N nodes then all of TmXi(K) i.e. the trusts of one-hop
neighbors (Xi) to the node K, are calculated where i = 1
to N, and m is the QoS parameters for different metrics.
This is shown in Figure 4. The parameters used for such
trust calculation are a) T1Xi(K) probability of successful
reception of packets of node Xi from node K (here m = 1);
b) T2Xi(K) ratio of minimum latency possible and mean
latency of packets sent from i to Xi (here m = 2); c)
T3Xi(K) ratio of the power level (or battery life) of node K
and the maximum power level (or battery life) possible
for that WSN node (here m = 3), etc. More number of
direct trust parameters can be added such as TmXi(K) etc.
We calculate overall direct trust as
 

1
123
iiiii
m
XXXXmX
TKT KTKTKTK

 

which is the Geometric Mean(GM) of all the parameters.
For m different metrics, Trust of node K at X1, X2 and X3
may be
 


1
111 1
1
123
π,,,,
X
m
XXX mX
TK
TKTKTK TK



 


2
222 2
1
123
π,,,,
X
m
XXX mX
TK
TKTKTKTK



 


3
333 3
1
123
π,,,,
X
m
XXX mX
TK
TKTKTK TK


In our case we assume WA = 1 and WB = 0. So TXi(K) is
the set of trust values assigned to the node K by one-hop
neighbors (Xi) .
This is different from the Mohammad Momani’s model
as he has calculated the trust as the arithmetic sum or
mean of the different parametric probabilities which can
lead to some serious false values. Suppose at least one pa-
rameter (say % successful packet transmission or packet
latency) gives a trust value of 0 but others have high va-
lues. So a high value of trust may be assigned even th-
K
X
1
X
2
X
3
X
4
Figure 4. Trust relationship with individual nodes.
ough packet transmission is 0 or latency is infinite. This
can be avoided if we calculate the product or the geo-
metric mean of the trusts as suggested in this model. This
is also proved by an example later.
It is to be noted that since trust is a probabilistic value
its range must always remain within 0 to 1 with a higher
value of indicating higher trust in the positive sense.
Hence it is to be noted that while calculating the trust if
there is some parameter which is better if less, then it has
to be multiplied after subtracting it from 1.
Step 2: Calculation of threshold (tth) direct trust value
and subsequent selection of benevolent nodes
The value of TTH is application-dependent and needs to
be determined according the accuracy, precision, relia-
bility, risk acceptable to be operated in the present net-
work. It should be calculated upon the consumer or ap-
plication needs. In course of this paper we assume the
TTH = 0.5 to be the balanced and appropriate value.
Hence all the subsequent formulations, deductions and
derivations involve TTH to be equal to 0.5. Now if
TXi(K) TTH then node K is trusted and it is assigned to
be a benevolent node with respect to the th
i
X
node.
Else if TXi(K) < TTH then node K is not trustworthy and
it is assigned to be a malicious node with respect to the
th
i
X
node. This process is repeated for i = 1 to N.
Step 3: Application of link state routing protocols
(lsrp) for finding all available paths using the benevo-
lent nodes only
Link state routing protocols are the most widely used
static routing protocols. We are only interested in the ba-
sic features of the LSRP and are not mentioning the wide
details of it or whether OPSF, IS-IS, MOPSF, MLSRP
etc. are used in this case. Applying anyone of these
LSRPs are possible depending upon other network
needs.
The basic features of LSRP in brief are:
a) Discovery of the neighbors of the nodes and learn-
ing their network addresses.
b) Measurement of the delay or cost to each of its
neighbors.
c) Construction of a packet telling all the information
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
130
learnt by it.
d) Transmission of this packet to all the router nodes.
One of the main advantages in our algorithm is that it
doesn’t require the LSRP to apply Dijsktra’s algorithm
or any other algorithm to find the shortest path from the
source to the sink. It gets automatically evaluated from
determination of Route Trust Values.
The above 4 steps help in completion of the adjacency
database in which each node stores all the information
regarding its neighbours. The only addition for our trust
based routing protocol is that an extra field will be added
to this database in our case that is of the Trust values.
Data packets or acknowledgement packets may be modi-
fied for inclusion of this field. For example, according to
the given topology if the DESTINATION node has 3 one
hop neighbours 3, 5 and 9 then it should have a table
embedded in its memory as shown in Table 1. Now such
a table will be formed only in those nodes and only those
nodes will be used for LSRP (i.e. included in the routing
table) which are determined to be benevolent nodes with
the method given in the previous step. Hence valid rout-
ing paths will only be decided by LSRP involving only
the benevolent nodes eliminating the malicious nodes.
Step 4: Evaluation of multiplicative route trusts (rt)
Suppose that upon the implementation of the LSRP
protocol the paths that are found out are shown by bold
continuous arrows in the following Figure 5.
Table 1. Different Trusts in node’s memory.
N_ID Trust Value
3 0.9
5 0.75
9 0.7
DE
S
TI
N
AT I
ON
SOURCE
1
4
10
8
2 5
9
3
6 7
Figure 5. Evaluation of multiple routes in WSN.
Assuming the all the colored nodes have trust valuesa-
bove the threshold value the paths we get are the follow-
ing:
Route 1 = S->1->2->3->D,
Route 2 = S->6->7->8->9->D,
Route 3 = S->4->5->D,
Route 4 = S->6->10->9->D, with the source(S) and the
destination (D) be included implicitly.
Now the Route Trusts are calculated as follows:
Trust for Route 1 i.e. RT1 = π (T1(S), T2(1), T3(2),
TD(3))
Trust for Route 2 i.e. RT2 = π (T6(S), T7(6), T8(7),
TD(9))
Trust for Route 3 i.e. RT3 = π (T4(S), T5(4), TD(5))
Trust for Route 4 i.e. RT4 = π (T6(S), T10(6), T9(10),
TD(9))
So the Route Trusts (RTs) are calculated by multiply-
ing the direct trusts of all the nodes belonging to the path
with each other. Such a method provides plenty of ad-
vantages.
Step5: Selection of the most appropriate path and
subsequent transmission of data through it
In the fifth and final step, data will be routed only th-
rough that path whose RT value is the highest.
So if in the previous case RT3 > RT1 > RT 4 > RT2, then
data under normal conditions will be transmitted through
the 3rd route or through S->4->5->D. The priority order
will be same as the decreasing values of RT. In case RTL
= RTM, where L and M are different routes. Then both the
routes will be given the same priority and transmission
can take place through any of them or may be through
some prior tie-breaking rule as per network needs like
less number of hops route.
This method provides many advantages as compared
to the existing ones: Firstly, it allows us to find the short-
est path without applying Dijsktra’s algorithm. It also
allows us to find out the correct path even when the TTH
value cannot be decided or evaluated and hence the sepa-
ration of benevolent and malicious nodes is not possible.
Even if it is done this method provides more accu- rate
and precise choice of the suitable path. Examples of
these cases are given next:
In the Table 2 given below we represent 4 distinct,
random but important cases which will show the benefit
of using our protocol for routing purposes compared to
the conventions followed presently although no clear pro-
cess of routing using trust values exist in the available
literature.
In accordance with the network topology given, we as-
sign the following unique trust values to each node. For
simplicity, if multiple instances of a single node exist in
each routing path then instead of assigning 2 different
trust values we have given a single value, e.g., T7(6) =
T10(6).
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
131
Table 2. Trust levels for different cases.
Node
trust
Trust level
Case I Case II Case III Case IV
T(1) 0.3 0.8 0.85 0.75
T(2) 0.85 0.5 0.4 0.75
T(3) 0.9 0.6 0.3 0.75
T(4) 0.6 0.75 0.75 0.75
T(5) 0.75 0.2 0.5 0.75
T(6) 0.9 0.4 0.55 0.75
T(7) 0.35 0.9 0.6 0.75
T(8) 0.5 0.9 0.4 0.75
T(9) 0 0.7 0.2 0.75
T(10) 0.30 0.9 0.9 0.75
These cases show one of the major advantages of our
protocol over the existing ones. In case of the existing pro-
tocols there is no existence of routing path trusts. They
suffer in different aspects:
Calculation of direct trust of a single node with respect
to another node based on number of parameters is ac-
complished by taking the average of the individual single
parametric trusts as shown in Momani’s model ([1,2]).
So if say the trust due to successful transmission of pack-
ets is 0 and the rest have a high value of it, there will not
be any successful transmission of packet although the
overall trust of the node will be quite high which denotes
definitely a false value.
Another deficiency of the protocols is that as the rout-
ing is based upon the trust values of the nodes based only
upon the one-hop neighbours they remain largely unaware
of the rest of the network topology. Decisions on routing
being taken largely and only upon the one-hop neigh-
bours the probability of choosing the best path is very
low.
If all the nodes are homogeneous and they have the
same levels of trust probabilities assigned then the selec-
tion of routes may be random and the routing may not
take place through the shortest path. It will be evident
from the following cases that our algorithm does not suf-
fer from these 3 difficulties but instead provide solu-
tions of these. We take each case and give the different
choices of routing paths with respect to the existing
method and our mode as follows:
Case 1 Case 2 Case 3 Case 4
Other
methods Route 2 Route 1 Route 1 Any
Route
Our
method Route 3 Route 4 Route 3 Route 3
4. Explanations
CASE 1: The existing method will choose Route 2 which
is evident as T(6) >T(4) > T(1) and T(7) > T(10). But this
decision is totally wrong since all the packets will be
entirely dropped at node 9. But our method will choose
Route 3 which not only provides the highest reliability
but also the shortest path. In the first step as shown be-
low, these nodes are eliminated based upon TTH = 0.5 and
with it the paths and then the RTs are calculated. This is
shown in the following Figure 6.
Even in case the threshold value is indeterminate, RT2
= RT1 = 0. Hence they will not be chosen.
CASE 2: When similar type of process is followed
Momani method chooses Route 1 and our method Route
4 in case of separation of benevolent and malicious nodes
take place on the basis of TTH = 0.5. This method signi-
fies the importance of the threshold trust level in our pro-
tocol because in its absence the path chosen would have
been R4. It would have been an incorrect path as in this
path the first node itself is behaving maliciously although
the RT4 value is the highest. But due to prior elimination
of the node 6 as a malicious node we can overcome the
problem. Setting the TTH value to suit the network’s QoS
needs appropriate path can be selected.
CASE 3: In this case upon application of the first and
second steps of our algorithm, only one path remains
valid so calculation overhead is decreased.
CASE 4: This case is very interesting. When the trust
levels of all the nodes are equivalent existing protocols
will be choosing randomly one of the valid paths. So if
there exists K number of paths the probability that the
most appropriate is chosen is only 1/K. However if we
apply our model, it will always choose the shortest path
DESTINATION
SOURCE
1
4 8
25
9
3
67
10
Figure 6. Appropriate route selection in WSN.
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
132
that is available between the SOURCE node and the
DESTINATION (SINK) node. It can be proved as fol-
lows:
Suppose T(i) = p, where T(i) is the trust level of the
node i and p is the desired probability. If there are three
paths P1, P2, P3 consisting of n1, n2 and n3 number of
nodes respectively then RTP1 = p * n1, RTP2 = p * n2 and
RTP3 = p * n3. If n1 > n2 > n3 then it is evident that RTP3 <
RTP2 < RTP1, as 0 p 1 and vice-versa.
Figure 7 shows the advantage of our protocol as
compared with the ATSR (LSRTP denotes DRTLSRP).
Although in a few cases the performance of both are
quite similar but in others our model scores over the
ATSR one.
Figure 8 shows the plot of transmission latency with
random trust assignment. Although we can’t clearly de-
cide which one is better, it’s possible to conclude that in
the long run our model behaves better than their one es-
pecially in the case of equal trusts and when the number
of nodes in the network is very large.
Figure 7. Packet transmission rate.
Figure 8. Packet delay.
Figure 9. Matlab simulation model.
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
133
5. Conclusion and Future Work
It can be ultimately concluded from this simulation re-
sults (as shown in Figure 9) that our model performs
better with respect to the available protocols such as
ATSR, CONFIDANT etc. Future work includes inclu-
sion of indirect trusts and calculation and assignment of
route trust based on fuzzy logic.
6. References
[1] M. Momani, J. Agbinya, G. P. Navarrete and M. Akache,
“A New Algorithm of Trust Formation in Wireless Sen-
sor Networks,” Proceedings of the 1st IEEE International
Conference on Wireless Broadband and Ultra Wideband
Communications, Sydney, March 2006.
[2] M. Momani, “Bayesian Methods for Modeling and Man-
agement of Trust in Wireless Sensor Networks,” Ph.D.
Thesis, University of Technology, Sydney, July, 2008.
[3] A. Sorniotti, L. Gomez, K. Wrora and L. Odorieo, “Se-
cure and Trusted In-network Data Processing in Wireless
Sensor Networks: A Survey,” Journal of Information As-
surance and Security, Vol. 2, No. 3, 2007, pp. 189-199.
[4] J. Lopez, R. Roman, I. Agudo and C. F. Gago, “Trust
Management Systems for Wireless Sensor Networks:
Best Practices,” Computer Communications, Vol. 33, No.
9, 2010, pp. 1086-1093.
doi:10.1016/j.comcom.2010.02.006
[5] T. Zahariadis, H. C. Leligou, P. Trakadas and Stamatis
Voliotis, “Mobile Networks Trust Management in Wire-
less Sensor Networks,” European Transactions on Tele-
communications, Vol. 21, No. 4, 2010, pp. 386-395.
[6] G. X. Zhan, W. S. Shi and J. Deng, “TARF: A Trust-
Aware Routing Framework for Wireless Sensor Net-
works,” Proceedings of the European Conference on
Wireless Sensor Networks, Coimbra, 17-19 February
2010, pp. 65-80.
[7] T. Kavitha, D. Sridharan, “Security Vulnerabilities in
Wireless Sensor Networks: A Survey,” Journal of Infor-
mation Assurance and Security, Vol. 5, No. 1, 2010, pp.
31-44.
[8] J. Sen, “A Survey on Wireless Sensor Network Se- cu-
rity,” International Journal of Communication Net- works
and Information Security, Vol. 1, No. 2, August 2009, pp.
59-82.
[9] C. Karlof and D. Wagner, “Secure Routing in WSNs:
Attacks and Countermeasures,” Ad Hoc Networks Jour-
nal, Vol. 1, No. 2-3, September 2003, pp. 293-315.
doi:10.1016/S1570-8705(03)00008-8
[10] A. A. Pirzada, C. McDonald and A. Datta “Performance
Comparison of Trust-Based Reactive Routing Protocols,”
IEEE Transactions on Mobile Computing, Vol. 5, No. 6,
June 2006, pp. 695-710.
[11] S. Zhu, S. Setia and S. Jajodia, “LEAP: Efficient Security
Mechanisms for Large-Scale Distributed Sensor Net-
works,” Proceedings of the 10th ACM Conference on
Computer and Communications Security, Washington
DC, 27-30 October 2003.
[12] A. Perrig, R. Szewczyk, V. Wen, D. Culler and J. D. Ty-
gar, “SPINS: Security Protocols for Sensor Networks,”
Proceedings of the 7th Annual International Conference
on Mobile Computing and Networking, Rome, 16-21 July
2001.
[13] C. Karlof, N. Sastry and D. Wagner, “TinySec: A Link
Layer Security Architecture for Wireless Sensor Net-
works,” Proceedings of the 2nd International Conference
on Embedded Networked Sensor Systems, Baltimore, 3-5
November 2004.
[14] H.-R. Lee, Y.-J. Choi and H.-W. Kim, “Implementation
of TinyHash Based on Hash Algorithm for Sensor Net-
work,” Proceedings of the World Academy of Science,
Engineering and Technology, Saint Louis, Vol. 10, Au-
gust 2005.
[15] M. Becker, S. Schaust and E. Wittmann, “Performance of
Routing Protocols for Real Wireless Sensor Networks,”
Proceedings of the 10th International Symposium on
Performance Evaluation of Computer and Telecommuni-
cation Systems, San Diego, July 2007.
[16] M. J. Probst and S. K. Kasera, “Statistical Trust Estab-
lishment in Wireless Sensor Networks,” Proceedings of
the 13th International Conference on Parallel and Dis-
tributed Systems, Hsinchu, Vol. 1, 5-7 December 2007.
[17] T. Zahariadis, H. Leligou, P. Karkazis, P. Trakadas, I.
Papaefstathiou, C. Vangelatos and L. Besson, “Design
and Implementation of a Trust-Aware Routing Protocol
for Large WSNs,” International Journal of Network Se-
curity & Its Applications, Vol. 2, No. 3, July 2010, pp.
52-68.
[18] Y. C. Hu, A. Perrig and D. B. Johnson, “Ariadne: A Se-
cure On-Demand Routing Protocol for Ad Hoc Net-
works,” Proceedings of the 8th Annual International Con-
ference on Mobile Computing and Networking, Atlanta,
23-28 September 2002, pp. 12-23.
doi:10.1145/570645.570648
[19] X. Q. Li, M. R. Lyu and J. C. Liu, “A Trust Model Based
Routing Protocol for Secure Ad-Hoc Networks,” Pro-
ceedings of the IEEE Conference on Aerospace, Big Sky,
Montana, Vol. 2, 6-13 March 2004.
[20] A. A. Pirzada and C. McDonald, “Trusted Greedy Pe-
rimeter Stateless Routing,” Proceedings of the 15th IEEE
International Conference on Networks, Adelaide, 19-21
November 2007, pp. 19-21.
doi:10.1109/ICON.2007.4444087
[21] A. Perrig, R. Szewczyk, J. D. Tygar, V. Wen and D.
Culler, “SPINS: Security Protocols for Sensor Net-
works,” ACM Journal of Wireless Networks, Vol. 8, No.
5, September 2002, pp. 521-534.
doi:10.1023/A:1016598314198
[22] S. Marti, T. Giuli, K. Lai and M. Baker, “Mitigating
Routing Misbehavior in Mobile Ad-Hoc Networks,” Pro-
ceedings of the 6th Annual International Conference on
Mobile Computing and Networking, ACM Press, Boston,
6-11 August 2000, pp. 255-265.
doi:10.1145/345910.345955
S. S. BABU ET AL.
Copyright © 2011 SciRes. WSN
134
[23] S. Buchegger and J. Boudec, “ Performance Analysis of
the CONFIDANT Protocol: Cooperation of Nodes –
Fairness in Distributed Ad-Hoc Networks,” Proceedings
of the 3rd ACM International Symposium on Mobile
Ad-Hoc Networking and Computing (MobiHoc), ACM
Press, Lausanne, 9-11 June 2002, pp. 226-236.
[24] S. Tanachaiwiwat, P. Dave, R. Bhindwale and A Helmy,
“Location-Centric Isolation of Mis-behavior and Trust
Routing in Energy-Constrained Sensor Networks,” Pro-
ceedings of the IEEE International Conference on Per-
formance, Computing and Communications, Phoenix,
April 2004, pp. 463-469.