NETWORK LAYER BASED SECURE ROUTING PROTOCOL
FOR WIRELESS AD HOC SENSOR NETWORKS IN URBAN
ENVIRONMENTS
Sanjay K. Dhurandher
1
, Mohammad S. Obaidat
2
, Deepank Gupta
1
, Nidhi Gupta
1
and Anupriya Asthana
1
1
Division of Information Technology, Netaji Subhas Institute of Technology, University of Delhi, New Delhi, India
2
Department of Computer Science & Software Engineering, Monmouth University, NJ, U.S.A.
Keywords: Routing protocols, Security, Wireless sensor networks.
Abstract: Security is essential in wireless sensor networks as they are being used in urban environments, life saving
disaster management and rescue operations. Any serious attack at the routing layer can cause serious
damages. Although a lot of security measures have been proposed for application and transport layer, we
have found that there is not enough research geared towards securing the network at the routing layer. In
this paper, we propose a novel solution for securing against external as well as internal attacks. The protocol
maintains a working network by using redundant multiple paths despite attacks at one route. It also
identifies and removes the malicious nodes from the system. Since the system is totally distributed and does
not require a central server as required in some of the other protocols, there is no single point of failure. We
also keep in mind the limited computing resources and network bandwidth of the wireless sensor nodes.
Finally the paper quantifies the protocol's effectiveness against some of the existing secure routing protocols
namely QDV and SNEP using simulation studies.
1 INTRODUCTION
Wireless Sensor Networks (WSNs) (Mainwaring et
al., 2006) are among the popular emerging
technologies that open a wide perspective for future
applications in ubiquitous computing and ambient
intelligence. Wireless sensor nodes form a dense,
large scale network and are expected to function
unattended. Their hardware limitations have led to the
design of new protocols at the lower communication
levels, such as physical, MAC (Demirkol et al., 2006)
and routing. WSNs can be thought of as an extension
to MANETs since special sensor nodes are included
into the wireless network to give further utilities or
increase the services given by it. Sensor nodes are
stations which sense variables about the physical
world around them. Generic nodes in the network
sense network characteristics whereas sensor nodes
sense changes in the environment in which they exist.
Ensuring security is critical in any WSN. If any
sensor node is hacked or information about various
environment variables fails to reach the desired
controllers, the consequences can be dire. Although
sensor networks share many characteristics with
wireless ad hoc networks, one of the major
differences is the energy and computational resources
available at sensor nodes. Therefore, any security
protocol for WSNs must keep in mind, the relatively
constrained energy and computational resources in
mind. With this criterion in mind, we find that many
existing protocols fail to fit in WSNs.
Routing protocols proposed for WSNs cope well
with the dynamically changing topology but many of
them have no mechanisms to defend against
malicious attacks. We are convinced that security
problems cannot be considered separately and must
be taken into account for the specification of all the
functionalities of the network. Since, research on
protocols is still going on, there is no single standard
routing protocol. Therefore, we aim to capture the
common security threats and try to provide a secure
existing routing protocol in WSNs.
In most of the routing protocols, routers exchange
information based on the topology of the network in
23
K. Dhurandher S., S. Obaidat M., Gupta D., Gupta N. and Asthana A. (2010).
NETWORK LAYER BASED SECURE ROUTING PROTOCOL FOR WIRELESS AD HOC SENSOR NETWORKS IN URBAN ENVIRONMENTS.
In Proceedings of the International Conference on Wireless Information Networks and Systems, pages 23-28
DOI: 10.5220/0003105100230028
Copyright
c
SciTePress
order to establish routes between nodes. Such
information could become a target for malicious
adversaries who intend to bring the network down.
There are two sources of threats to routing protocols.
The first comes from external attackers. By injecting
erroneous routing information, replaying old routing
information, or distorting routing information, an
attacker can successfully partition a network or
introduce excessive traffic load into the network by
causing retransmission and inefficient routing. There
are two ways of ensuring security: either by
encrypting the data or by identification and removal
of malicious node from the network. The first
approach using public-key algorithms such as the
Diffie Hellman (Perrig et al., 2001) is not suitable for
a wireless sensor network in which nodes have
limited computing resources. Using such an algorithm
will mean a large amount of computation power being
used to encrypt and decrypt every message making
the network slow.
The second and more severe kind of threat comes
from the compromised nodes, which might advertise
incorrect routing information to other nodes.
Detection of such incorrect information is difficult.
Merely requiring routing information to be signed by
each node would not work, because compromised
nodes are able to generate valid signatures using their
private keys. To defend against the first kind of
threats, nodes can protect routing information in the
same way they protect data traffic, i.e., through the
use of cryptographic schemes [3, 4] such as digital
signature. However, this defense is ineffective against
attacks from compromised servers. Worse yet, as we
have argued, we cannot neglect the possibility of
nodes being compromised in an ad hoc network.
Detection of compromised nodes through routing
information is also difficult in an ad hoc network
because of its dynamically changing topology: when a
piece of routing information is found invalid, the
information could be generated by a compromised
node, or, it could have become invalid as a result of
topology changes. It is difficult to distinguish
between the two cases. On the other hand, we can
exploit certain properties of ad hoc networks to
achieve secure routing. Note that routing protocols for
ad hoc networks must handle outdated routing
information to accommodate the dynamically
changing topology. False routing information
generated by compromised nodes could, to some
extent, be considered outdated information. As long
as there are sufficiently many non-malicious nodes,
the routing protocol should be able to find routes that
go around these compromised nodes. Such capability
of the routing protocols usually relies on the inherent
redundancies‚ multiple, possibly disjoint, routes
between nodes in ad hoc networks.
With the above premise in mind; we propose a
solution to work around malicious nodes and also to
detect and remove the malicious nodes when found
from the trusted population (other nodes to which a
node transmits data). The solution to this has been
proposed in this paper which is an extension of the
proposed energy efficient protocol in (Sanjay et al.,
2010). We have named this protocol as Dynamic
Energy Efficient and Secure Routing (DEESR)
protocol. This can easily be extended to other cluster
based routing protocols. The distributed algorithm
applied in this protocol does not require excessive
computation resources. Also, there is no extra
network cost involved with this protocol. The paper
goes on to compare the proposed DEESR solution
with some of the earlier proposed secure routing
protocols using simulation techniques. The paper has
been divided as follows. Section 2 discusses existing
methods of achieving secure routing by discussing
some of the well known secure routing protocols.
Section 3 describes the terms and terminologies used
throughout the paper. In section 4, we provide a
detailed explanation of the proposed DEESR
protocol. Simulation results comparing DEESR
protocol with the existing approaches are presented in
Section 5. Lastly, Section 6 summarizes our finding
and provides insights into the future works.
2 PREVIOUS WORKS
There are various security based protocols (Perrig et
al., 2001), (Sanjay et al., 2009), (Yi et al., 2001),
(Karlof et al., 2004) for ad hoc networks. Yi (Yi et al.,
2001) have discussed in their paper that if the routing
protocol is compromised by changing the messages in
the transit, then no security at higher layers can help.
To address this problem they have proposed Security
Aware Ad-hoc Routing (SAR). It makes sure that data
is routed through a secure route composed of trusted
nodes and the security of the information in the
routing protocol. Apart from this, security has been
implemented at link layer in TinySec protocol given
in (Karlof et al., 2004).
SPINS (Perrig et al., 2001) was proposed keeping
in mind the resource limitations. This protocol
encrypts a message differently each time. SPINS
comprises of two building blocks, SNEP and μ-
TESLA. SNEP protocol was designed for stationary
networks and assumed that the base station is trusted.
It also assumed an access point for the other nodes in
the network. The μ-TESLA protocol is based on key
WINSYS 2010 - International Conference on Wireless Information Networks and Systems
24
chain generation, in which a key is generated from the
next key, in addition to the application of a function
and the last key is generated once in a time-interval.
But the encryption approach still requires storage of
public and private keys. Also, the transmission of
digital signatures consumes more energy. Though, it
ensures data authentication and prevents other nodes
from modifying and reinserting packets into the
network, it does not prevent them from overhearing
the data-packets.
Another security routing protocol, QDV (Sanjay
et al., 2009) is based on quality of the route taken by
the packets to be transmitted between nodes in the
network. The route taken, by the packet, is governed
by different parameters based on the quality of
service and quality of security and also the future
benefits in the transmission.
This protocol (QDV) designed for securing the
wireless sensor networks and is based on Ant colony
optimization (ACO) (Dorigo and Stuetzle, 2004). It
uses quality-of- service (QoS) and reputation of the
node to find out the trust of the neighbor node. Thus,
by monitoring these two parameters the protocol is
able to detect and disable the malicious nodes from
gaining access and participating in the network.
Though it takes into account the malicious activity of
packet drop and reinsertion, it does not secure the data
in the data packet. Also it is not energy efficient for a
resource constrained environment.
3 TERMS AND NOTATIONS
The DEESR protocol has been designed so that a
sensor ad hoc network can be set up for mobile
devices securely and in an energy efficient manner.
To facilitate this, certain parameters have been
focused upon. These parameters are then used to
reach to a routing decision. The following subsections
present the terms and parameters used in the protocol
as it is necessary to understand them before
proceeding.
Population
Every node in the network has a set of nodes which
are in its transmission range and with which it decides
it can communicate securely. This set of nodes with
which a node communicates with is called the
population of that node. Only the nodes satisfying a
certain criteria are included in the population of a
node. In this protocol, as security is of prime
importance, the parameter used for narrowing down a
node's population is the Dynamic Trust Factor (DTF).
To limit the database requirements and to
minimize network traffic for population maintenance,
a node maintains a population restricted up to a
maximum population size. The population size is
dependent on the computing resources and network
bandwidth of a node.
Dynamic Trust Factor (DTF)
This is given by the ratio of the sum of the packets
received and generated subtracted by the number of
packets transmitted to the sum of the packets received
and generated. Thus, DTF gives the ratio of the
number of packets dropped to the total number of
packets received and generated. In DEESR protocol a
particular node is trusted if its packet dropping ratio is
less. Thus, trust is inversely proportional to the
number of packets dropped. We do not consider or
handle the situation where the packet is tampered
with at the next-hop before retransmission. This value
is also extracted from the routing table. Equation (1)
shwon below represents how the DTF is calculated:
δ = ( p
r
+ p
g
- p
d
) / ( p
r
+ p
g
) (1)
Where:
δ denotes the Dynamic Trust Factor (DTF)
Pr = number of packets received by that node
p
g
= number of packets generated by that node
p
d
= number of packets dropped by that node
Dynamic Trust Factor (DTF) is given by the ratio
of the packets transmitted by a node to the total
number of packets received and generated at that
particular node. Subtracting the number of
packets dropped (p
d
) from the sum of the number
of packets received (p
r
) and packets generated
(p
g)
results in the number of packets transmitted.
Similarly the sum of p
r
and p
g
results in the total
number of packets at a particular node. At the
source and the destination nodes, the packets for
which the node is the source/destination are not
counted in the DTF calculation, which provide
security against active attack, specifically,
malicious packet injection.
4 PROTOCOL
AND EXPERIMENT
In this section, the functioning of the proposed
protocol has been explained. There are four major
parts of the protocol: route initiation, population
maintenance, updating of next-hops and route-error
respectively. The route initiation and data packet
NETWORK LAYER BASED SECURE ROUTING PROTOCOL FOR WIRELESS AD HOC SENSOR NETWORKS IN
URBAN ENVIRONMENTS
25
transmission are not explained in major detail in this
section. Most of the focus is on population
maintenance and updating of next-hops.
Route Initiation
Let us consider a scenario in which a node A acts as a
source and requires sending data to the destination
node E according to the topology shown in Figure 1.
A broadcasts the route-request packet. This packet is
received by, say, node B which updates the
information about A in its routing table and
broadcasts the packet again. Similarly, this goes on
till the packet reaches the destination node E with the
information of the previous node. Since E is the
destination node, it initiates a route-reply packet and
sends it to all the previous nodes from which it has
received the request since there maybe more than one
route between the source and destination. The first
hop in the reply route updates the information of the
destination node and again replaces that information
with its own information and unicasts the packet back
to previous nodes in the route. When the reply packet
finally reaches the source node, it makes an entry in
its routing table for the routes.
In Figure 1, let us assume that a malicious node F
injects packets into the network disguising itself as a
trustworthy node, say, node A. Since F is not the
actual source, so the disguised packet will be tracked
down and the DTF will be calculated which will
indicate its untrustworthiness. Hence, DEESR
protocol will be able to isolate such a node from the
network.
Attack during route initiation
Figure 1: Illustration of how a malicious node F is detected
in the system.
Population Maintenance
The number of nodes in the population of a particular
node is not equal to the number of routing table
entries, but equal to the number of distinct next-hops
in the routing table. This is because, there might be a
single node which is the next-hop for two
destinations, which results in two routing table
entries, but only one distinct node in the population.
At some point, the members in the population may
increase beyond the population size. At this juncture,
the node has to eliminate certain nodes from its
population, which it does by evaluating the DTF of
the nodes present in the routing table and keeping the
nodes with the maximum DTF out of the nodes
present.
Let us illustrate this with the help of an example
where the Population size is 5. In Figure 2, frame 1
shows the scenario when node B enters into the
network and has no nodes in its population. Slowly,
as it needs to communicate more, it starts finding
more nodes as next-hops to include in its population.
This is depicted in frames 2,3,4,5 and 6 of Figure 2.
Node B does not check for any parameter before
taking a node into its population till the population
size is reached.
Figure 2: Shows how other nodes get included into
population of node B till Population Size is reached.
Figure 3: Shows the scenario after an extra node competes
to be included into node B's population.
WINSYS 2010 - International Conference on Wireless Information Networks and Systems
26
Now in Figure 3, see frame 7 on top of figure,
node G is also discovered by node B, but it already
has the maximum number of members in its
population, but does not know if node G should be
included or not. To resolve this, it calculates the DTF
of all the nodes in its population along with the DTF
for node G. After calculation, node B finds that the
node D has the minimum DTF and thus decides to
eliminate node D and give membership to node G in
its population; thus, maintaining its population size.
Updating of Next-Hops
Route maintenance is required to minimize the
occurrences of route errors. In this subroutine, when a
node receives a reply packet for a destination, it
checks in its routing table if the current next-hop is
the same as the previously stored next hop.
An updating subroutine is also used in the
protocol for updating the nodes in the population of a
certain node about the activity of that node. An
UPDATE packet is created and transmitted to the
next-hop nodes in the routing table of a node
whenever there is a change in its characteristic
parameters namely: packets generated, packets
received and packets forwarded. This is uni-cast to
only those nodes which feature as a next-hop in its
routing table. It is also to be noted here, that the node
sending the UPDATE packet does not send it to that
entry in the routing table which it used as a next-hop
for data transmission. This is beneficial since all the
other nodes will have updated information regarding
the node and hence a fresh value of the DTF. In this
process, it is to be noted that there is no reply for the
UPDATE packet in order to prevent the unnecessary
usage of battery power of a node.
To prevent the network from getting flooded with
update packets and also to reduce the energy
consumption at the sending and receiving nodes of
these packets, the update packets are sent after
specific intervals rather than after every data packet.
This interval is dictated by the amount of change in
battery power left, number of packets transmitted,
packets received and packets generated.
Route Error
This subroutine is invoked when a route existing in
the routing table turns out to be broken. Once a source
node needs to send data to a destination node, it looks
up its routing table. If a route exists then, the packet is
forwarded to the next-hop for the given destination.
Similarly, the packet is forwarded till it reaches the
destination. In case, the route found in the routing
table is outdated or broken, this subroutine starts,
where, the route-error packet traces its path back to
the source, which removes the entry from its routing
table and starts forward the route discovery procedure
(Sanjay et al., 2010). While the route-error packet is
going back to the source node from the node where
the broken path was discovered, all the intermediate
nodes also delete the entry for the specific destination
from their routing tables.
5 SIMULATION ANALYSIS
AND RESULTS
The simulation experiments conducted were
evaluated using the Global Mobile Information
System Simulator (GloMoSim version 2.03) (Bajaj et
al., 1997). It has been designed using the parallel
discrete-event simulation capability provided by
PARSEC (Bagrodia et al., 1998). PARSEC is a C-
based simulation language, developed by the Parallel
Computing Laboratory at UCLA. It can also be used
as a parallel programming language. GloMoSim
currently supports protocols for a purely wireless
network. In this section, we shall be comparing the
DEESR protocol with the SNEP and the QDV
protocols to gauge the security aspect of the DEESR
protocol. A comparative study is done on the basis of
their time of detection of malicious nodes and the
number of malicious nodes detected out of the total
malicious nodes present.
Figure 4 shows the effect of malicious nodes on
the packet delivery ratio in the network. In Figure 4,
the x-axis represents the percentage of malicious
nodes in the network. The y-axis represents the packet
delivery ratio for the three protocols. As can be seen
from the Figure, the packet delivery ratio of DEESR
protocol is considerably better than both the QDV
protocol and the SNEP protocol. But, in general, with
the increase in the percentage of malicious nodes, the
Figure 4: Comparison of Packet Drop ratio against two
secure protocols.
NETWORK LAYER BASED SECURE ROUTING PROTOCOL FOR WIRELESS AD HOC SENSOR NETWORKS IN
URBAN ENVIRONMENTS
27
packet delivery ratios drop significantly in the
network; dropping to as low as 0.6 in the case of
SNEP.
In Figure 5, we compare the detection times of the
three protocols while varying the percentage of
malicious nodes in the network. Thus, here also the x-
axis represents the percentage of malicious nodes in
the network. The y-axis represents the time taken to
detect the malicious nodes in the network. It is clear
from the Figure that the time taken for detection of
the malicious nodes is minimum for DEESR protocol
in all the instances. For both QDV and DEESR
protocols, the amount of time taken to detect
malicious nodes keeps decreasing when the
percentage of malicious nodes keeps increasing.
Figure 5: Comparison of time taken to detect malicious
nodes in the network.
6 CONCLUSIONS
In this paper, we have discussed the two types of
security threats viz. attacks by an external attacker
and attacks through a compromised node. Then we
discussed an approach towards secure routing which
involves automatic detection and removal of
malicious nodes in the system to keep the system
secure while making sure that the network keeps
operating despite the attacks. With extensive
simulation studies, we believe that the protocol is safe
for communication in different scenarios, generally
delivering better results in terms of the packet
delivery percentage and detection of malicious nodes
over the QDV and the SNEP security based protocols.
Thus, the protocol can provide reasonable amounts of
services while keeping a good level of security. Since
security is a dynamic field and there might be new
ways of attacking being discovered every-day, the
task does not end with this protocol itself. Future
work involves adapting the protocol to major routing
protocols and testing the protocol in a real
implementation environment.
REFERENCES
Alan Mainwaring, David Culler, Joseph Polastre, Robert
Szewczyk, “Wireless sensor networks for habitat
monitoring”, in Proceedings of the 1
st
ACM
international workshop on Wireless Sensor Networks
& Applications, Atlanta, Georgia, pp. 88-97, 2002.
Ilker Demirkol, Cem Ersoy, and Fatih Alagöz, “MAC
Protocols for Wireless Sensor Networks: A Survey”,
IEEE Communications Magazine, Vol. 44, Issue 4, pp.
115-121, 2006
Adrian Perrig, Robert Szewczyk, J. D. Tygar, Victor Wen
and David E. Culle, “SPINS: Security Protocols for
Sensor Networks”, in Proceedings of 7
th
Annual
International Conference on Mobile Computing and
Networking, pp. 189-199, Rome, Italy, July 2001.
Sanjay K. Dhurandher, S. Misra, M. S. Obaidat and N.
Gupta, “An Ant Colony Optimization Approach for
Reputation and Quality-of-Service-Based Security in
Wireless Sensor Networks”, Security and
Communication Networks, John Wiley & Sons,
Volume 2, Issue 2, pp. 215-224, March/April 2009.
Sanjay K. Dhurandher, Mohammad S. Obaidat, Deepank
Gupta, Nidhi Gupta, Anupriya Asthana, “A Dynamic
Energy Efficient Routing Protocol for Wireless Ad Hoc
Sensor Networks in Urban Environments”, submitted
to IEEE Globecom 2010.
S. Yi, P. Naldurg, and R. Kravets, “Security-aware routing
protocol for wireless ad hoc networks,” in Proceedings
of the 2nd ACM International Symposium on Mobile
Ad Hoc Networking & Computing, pp 299-302, Long
Beach, CA, USA, October 2001.
Chris Karlof, Naveen Sastry, David Wagner, “TinySec: A
Link Layer Security Architecture for Wireless Sensor
Networks”, in Proceedings of the 2
nd
International
Conference on Embedded Networked Sensor Systems,
SenSys 2004, Baltimore, MD, USA, pp. 162-175,
November 3-5, 2004.
M. Dorigo and T. Stuetzle, Ant Colony Optimization,
Prentice Hall, 2004.
L. Bajaj, M. Takai, R. Ahuja, K. Tang, R. Bagrodia, and
M. Gerla. “GlomoSim: A Scalable Network Simulation
Environment”. Technical Report CSD Technical
Report, 990027, UCLA, 1997.
Bagrodia R., Meyer R., Takai M., Chen Y., Zeng X.,
Martin J., Song H.Y. “PARSEC: a parallel simulation
environment for complex systems” IEEE Computer,
Vol. 31, No. 10, pp. 77-85, 1998.
WINSYS 2010 - International Conference on Wireless Information Networks and Systems
28