Secure Routing Protocols: Affections on

MANETs Performance
Shervin Ehrampoosh and Ali Khayatzadeh Mahani

Abstract— In mobile ad hoc networks, the absence of legitimate users and malicious nodes so various forms of
infrastructure and the consequent absence of authorization attacks such as passive eavesdropping, active signal
facilities impede the usual practice of establishing a practical interference, and jamming could be hampered the network.
criterion to distinguishing nodes as trusted and distrusted. The cooperative nature of Ad Hoc routing protocols makes it
Since all nodes in the MANETs would be used as router during
more vulnerable to data tampering, impersonation, and denial
working period, secure routing protocols have vital rule in the
security of the network. So evaluating the performance metrics
of services (DoS). MANETs architecture causes the
of secure routing protocols and calculating the overhead of inefficiency of some conventional security solutions, such as
security design is very important in network engineering. In a public key infrastructure (PKI), and the intrusion detection
this paper the effects of Black hole attack on two different types system [3]-[4].
of routing, OLSR as proactive routing protocol and AODV as Some of the standard terms about security concepts of
reactive routing protocol are considered. In addition, the networks are message non-repudiation, Isolation and Trust.
performance metrics of ordinary routing protocols (AODV and message non-repudiation means that the sender cannot deny
OLSR) are compared with a new secure routing protocol transmitting a packet. Isolation requires that protocol be able
MAODV, in which neighbors of node help him to distinguish a to identify misbehaving nodes and render them unable to
trusted node. Finally, the simulation results show that the end-
interfere with routing. Alternatively, the routing protocol
to-end delay of a network under black hole attack with
different type of routing protocols, AODV, OLSR and should be designed to be immune to malicious nodes.
MAODV. The simulation results illustrate good comparison of One aspect of mobile ad hoc networks that complicates the
network performance parameters for different conditions such design of a secure routing protocol is in-network aggregation.
as node mobility and number of nodes. In networks that are more conventional a secure routing
protocol is typically only required to guarantee message
availability. Message integrity and confidentiality are handled
Index Terms— Mobile Ad Hoc network (MANET), at a higher layer by an end-to-end security mechanism such as
wireless network, security, OLSR, AODV. SSH or SSL. In an ideal world, we would like to guarantee
the confidentiality, integrity, authenticity and availability of
all messages in the presence of resourceful adversaries. Since
I. INTRODUCTION MANET does not have a centralized piece of machinery such
as a name server, which could lead to a single point of failure
M OBILE Ad Hoc networks (MANET) are self
configuring networks in which mobile devices
connected by wireless links. These networks classify into
and thus make the network that much more vulnerable.
In such an environment, there is no guarantee that a path
infrastructure less networks, where the network between two nodes would be free of malicious nodes that
communication is established without any fixed would not comply with the employed protocol and would
infrastructure, such as battlefields, military applications and attempt to harm the network operation. The mechanisms
other emergency disaster situations. Obviously, security is a currently incorporated in routing protocols cannot cope with
critical issue in such areas [1]-[2]. Network architecture, disruptions due to malicious behavior. For example, any node
shared wireless medium, resource constraints, and highly could claim that it is one hop away from the sought
dynamic topology, lead to various challenges of the security destination, causing all routes to the destination to pass
design. Since the wireless channel is accessible to both through itself. Constructing security for the first time may not
be so difficult, maintaining trust and handling dynamic
changes seem to need more effort. So many routing protocols
in mobile ad hoc network to guarantee the security of network
increase the control packets [12]-[23]. Hence, the overhead of
Shervin Ehrampoosh is an M.Sc. student at Faculty of Electrical and
Computer Engineering, Kerman Graduate University of Technology packet transmission, and the end-to-end delay would be
(KGUT), Kerman, Iran (e-mail: [email protected]). increased.
Ali Khayatzadeh Mahani is with the Department of Electrical In this paper the performance metrics of routing protocols
Engineering, Shahid Bahonar University of Kerman, Kerman, Iran. are simulated. The results are depicted for different protocols
(e-mail: [email protected])
77
and finally a secure protocol is compared with AODV and
OLSR. The rest of the paper is organized as follows. In
Section II, routing protocols are reviewed. The effects of
black hole attack in AODV, OLSR, and new approach called
MAODV are discussed in section III. Simulation results of
performance metrics are described in section IV and the
paper is concluded in section V.

II. ROUTING PROTOCOLS

The routing protocols of ad hoc networks are classified Fig.2 Route Error Message in AODV
into two main categories, proactive and reactive. In a
proactive (sometimes-referred to as table-driven) routing B. OLSR protocol
protocol, nodes periodically exchange routing information OSLR protocol is a proactive protocol used in mobile ad-
with other nodes to update their routing information. The hoc networks. It is often called table-driven protocol as it
Optimized Link State Routing (OLSR) [8] protocol is a well- maintains and updates its routing table frequently. OLSR has
known proactive routing protocol. In a reactive (sometimes- also three types of control messages that are describe bellow.
referred to as source-initiated) protocol, a route from source
to destination would be established only when the source 1) Hello
node has a packet to send to the destination [5]. Dynamic 2) Topology Control (TC)
Source Routing (DSR) [6] and Ad Hoc On-Demand 3) Multiple Interface Declaration (MID)
Distance Vector (AODV) [7] are two important examples of
reactive routing protocols. Nevertheless, unfortunately there Hello message is transmitted for sensing the neighbor and
is no common standard routing protocol in MANETs. multi-point distribution relays (MPR) calculation. Topology
control is link state signaling that is performed by OLSR.
A. AODV Protocol MPRs are used to optimize theses messaging. MID messages
contains the list of all IP addresses used by any node in the
When a node “A” as a source node try to initiates a network. All the nodes running OLSR transmit these
connection to destination node “D”, it will generate a route messages on more than one interface.
request message (RREQ). This message is transmitted OLSR exchanges the topology information always with
through a limited flooding to their neighbors. In the second other nodes. Few nodes are selected as MPRs (Multi point
hop the message is forwarded to the neighbors of neighbors relays). MPRs are responsible for transmission of broadcast
and would be continued till to finding destination node or messages during flooding and generating link state
finding a node that has a fresh route to the destination. Then information. MPRs technique used in OLSR protocol will
a new control message, route reply message (RREP), is reduce the message overhead and even minimize the number
replied to the source node. When RREP reaches the source of control messages flooded in the network (Fig 3).
node, a route is established between the source node “A” and Nodes maintain the information of neighbors and MPR's,
destination node “D”. Once the route is established between by sending and receiving HELLO messages from its
“A” and “D”, the communication would be started. Fig. 1 neighbors. This will help in determining the link formation
depicts the exchange of control messages between source illustrated in Fig 4.
node and destination node.
1) Node X transmits the HELLO message to node Z and then the
message received by node Z from node X that can be called
asymmetric link.
2) Even if The node Z transmits the HELLO message to node X
then the resulting link retransmits this HELLO message called
asymmetric link.
3) Finally, the resulted bidirectional link is known as a symmetric
link.
4) Symmetric link formation will help the nodes to choose MPRs.
5) MPRs will send the topology control (TC) messages
containing the information about link status and MRP node
information [9].
Fig. 1 AODV Route Discovery
If the route between source and destination is broke, the
RERR message is sent to the source and destination nodes
separately. The scheme of sending RERR message in a
network s shown in the Fig. 2.
78
 Fig.3 Flooding Packets Using MPR Fig.5 Security Threats
In Fig 6, source node S wants to send data packets to a
destination node D in the network. Node M is a malicious
node that acts as a black hole. The attacker replies with false
reply RREP having higher modified sequence number. So
data communication initiates from S towards M instead of D.
In OLSR black hole attack, a malicious node forcefully
selects itself as MPR. Malicious node keep its willingness
field to Will always constantly in its HELLO message.
Therefore, in this case, neighbors of malicious node will
Fig.4 OLSR Symmetric link formation (Hello Message Exchange)
always select it as MPR. Hence, the malicious node earns a
privileged position in the network that it exploits to carry out
III. SECURE ROUTING PROTOCOLS
the denial of service attack. The effect of this attack is much
harmful when more than one malicious node is present near
As mentioned above MANETs often suffer from security
the sender and destination nodes.
attacks because of their specification such as open medium,
dynamic topology, lack of central monitoring and
management, cooperative algorithms and no clear defense
mechanism. These factors have changed the battlefield
situation for the MANET against the security threats [10].
The attacks could be classified based on:

- The behavior of the attack (Passive vs. Active)

- The source of the attacks (Internal vs. External)
- The processing capacity of the attackers (Wired vs. Mobile)
- The number of the attackers (Single vs. Multiple)
Fig.6 Black hole attacks in MANETs
Also according to the position of attacker, the attacks are
divided into four categories, which are shown in figure 5. Therefore, black hole attacks are non-single attacks; since
These categories are Interception, Interruption, Modification multiple malicious nodes could be acting as a group of
and Fabrication. attacker.
In black hole attack as an active attack, the malicious node
waits for the neighbors to initiate a RREQ packet. As the A. Mitigation Techniques against Black Hole Attack
node receives the RREQ packet, it will immediately send a
false RREP packet with a modified higher sequence number. In MANET, attacks that modify routing messages can be
So, that the source node assumes that node is having the provoked by the use of source authentication. Digital
fresh route towards the destination. The source node ignores signature, message authentication code (MAC), hashed MAC
the RREP packet received from other nodes and begins to (HMAC) can be used. Up to certain level of security can be
send the data packets over malicious node. A malicious node attained at network layer in internet by the use of IPSec.
takes all the routes towards itself. It does notallow Authenticated Routing for Ad-Hoc Networks (ARAN) is
forwarding any packet anywhere. This attack is called a another routing protocol that provides the protection from
black hole as it swallows all objects; data packets [11]. Black Hole attack where there is threat to the changes in
sequence number, hop count modification, source routing
changes and spoofing of destination addresses [12].
The protocol implement in [13] proposed Secure Ad-Hoc
On-Demand Distance Vector Routing (SAODV), which
79
verifies the destination node by exchanging random Rule 1: If a node delivers many data packets to
numbers. SAODV effectively prevents Black Hole attack in destinations, it is assumed as an honest node.
Mobile Ad-hoc network and it is better than AODV in terms Rule 2: If a node receives many packets but do not
of security and routing efficiency. sent same data packets, it is possible that the
Authors of [14] are focused on the requirement of a source current node is a misbehavior node.
node to wait unless the arrival of RREP packet from more Rule 3: When the rule2 is correct about a node, if the
than two nodes. When it receives multiple RREPs, the source current node has sent number RREP packets;
node check that there is any share hops or not. The source therefore surely the current node is
node will consider the routed safe if it finds the share hops. misbehavior.
Its drawback is the introduction of time delay that it has to Rule 4: When the rule2 is correct about a node, if the
wait for the arrival of multiple RREPs before authenticating
current node has not sent any RREP packets;
a node.
therefore the current node is a failed node.
In [15], the authors proposed route confirmation request
message (CREQ) and route confirmation reply (CREP) in
order to avoid Black Hole attack. So when an intermediate IV. SIMULATION RESULTS
node sends RREPs to the source node also it send CREQ to
its next hop node in direction of destination node. After Our simulation model was carried out using the OPNET
receiving CREQ, the next hop look for route in its Modeler software platform. It is a useful research tool for
destination in cache. If a CREP is received during this time it achieving good simulation results. Mobility scenarios are
will confirm the validity of path in RREP and in CREP. generated by using a Random waypoint model by varying 20
Upon matching the source node will recognize the route to 80 nodes moving in a terrain area of 1000m x 1000m. Each
being correct. Its drawback is that it cannot detect multiple node independently repeats this behavior and mobility is
Black Hole attacks. changed by making each node stationary for a short period.
In [16], the author showed that malicious node should The simulation parameters are summarized in Table III.
increase the sequence number of destination to assure the The simulation is done to analyze the performance metrics
source node of its route. The author proposed a statistics of the network. The metrics used to evaluate the performance
based detection for Black Hole that is based on the difference are given below:
between destination sequence numbers of received RREP’s. 1) Packet Delivery Ratio: The ratio of the data delivered
Its drawback is the false positives approach because of the to the destination to the data sent out by the source.
nature of anomaly detection. 2) Average End-to-End delay: The difference in the time
it takes for a sent packet to reach the destination. It
B. MAODV Protocol includes all the delays, in the source and each
intermediate host, caused by the routing discovery,
As discussed in previous section, such malicious nodes queuing at the interface queue etc.
can also create new routing messages and advertise
nonexistence links provide incorrect link state information Main network variables, which are considered to simulate
and flood other nodes with routing traffic thus inflicting the effects of security on the performance metrics, are listed
failures on the system. below.
In this section, we use an approach that has been proposed
in [17] to combat black hole attack in AODV routing Network size: variation in the number of mobile nodes.
protocol. In this approach, numbers of rules are used to Traffic load: variation in the number of sources.
inference about honesty of reply’s sender. The proposed Mobility: variation in the maximum speed.
method is based on this principles the activity of a node in a
network shows its honesty. Each node for participating in Figure 7(a) and 7(b) show that under black hole attack the
data transfer process, must be demonstrate its honesty. Early PDR (Packet Delivery Ratio) of MAODV is improved by 40-
of simulation, all nodes are able to transfer data; therefore, 60% than AODV under attack with Average-End-to-end
they have enough time to show its truth. If a node is the first delay almost same as normal AODV. In addition, we find that
receiver of a RREP packet, forwards that packets to source the difference between OLSR under attack and MAODV is
and initiates judgment process on about replier. The 30-60%.
judgment process is based on opinion of network’s nodes Figure 8(a) and 8(b) conclude the simulation based on the
about replier. Neighbors of each node store the activities of effect of mobility on the MAODV compared to normal
that node. So during the judgment process the neighbors AODV. The PDR stays within acceptable limits almost
send their opinion about a node. When the node collects all 5-20% lower than it should normally be with minimum
opinions of neighbors, it decides about honesty of reply’s overhead with Average-End-to-End delay almost same as
sender node. The decision is based on the following rules normal AODV.
which are used to judge about honesty of a node.

80
 Figure 9(a) and 9(b) depict that as the traffic load is
increased, the PDR (Packet Delivery Ratio) of MAODV
increases by approximately 40% than AODV under attack.

Fig. 8 Impact of Mobility on the performance

Fig.7 Influence of Network Size on the performance

TABLE III
SIMULATION PARAMETERS
Parameter Value
Simulator OPNET 14.5
Routing Protocol AODV, OLSR and MAODV
Simulation Time (sec) 1000
Number of Nodes 20-80
Simulation Area (m × m) 1000 × 1000
Packet Size (bit) Exponential(1024)
Minimum Mobility (m/s) 10
Maximum Mobility (m/s) 60
Transmission Range (m) 250
Traffic Model TCP
MAC Protocol IEEE 802.11
Packet Size (bit) 1024
Mobility Model Random Way Point (RWP)
Pause Time in RWP (sec) 50
Message TTL (sec) 100
Data Rate (Mbs) 11
Transmit Power (mW) 5 Fig.9 Influence of Traffic Load on the performance
No. of Malicious Node 5
No. of Source Node 1-6 V. CONCLUSION
In this paper the effects of Black hole attack in MANET
using both Proactive routing protocol Reactive routing
protocol such as OLSR and AODV are considered. The
impact of Black Hole attack on the performance of MANET
is illustrated finding out which protocol is more resilience
against packet delivery ration, mobility and end-to-end delay.
Additionally, a secure routing protocol, which is proposed in

81
[17], is considered and the impacts of security design on the [14] M. Al-Shurman, S-M. Yoo, and S. Park, “Black Hole
network performance metrics are simulated. As we expect Attack in Mobile Ad-Hoc Networks.”, ACM Southeast
the Simulation results show that the secure protocol has Regional Conf. 2004.
more end-to-end delay comparing with the AODV and [15] G. A. Pegueno and J. R. Rivera, “Extension to MAC
802.11 for performance Improvement in MANET.”,
OLSR, but it provides better performance in terms of packet Master Thesis at Karlstads University, Sweden, December
delivery ratio than the conventional routing protocols in 2006.
presence of Black holes attack. In the Future works we will [16] S. Kurosawa et al., “Detecting Black hole Attack on
concentrate on the secure routing design with better resilience AODV-Based Mobile Ad-Hoc Networks by Dynamic”,
against mobility. IEEE Military Communications Conference, Vol. 2,
page(s):1054-1059, Oct 2003.
[17] Mehdi Medadian, M.H. Yektaie, A.M Rahmani, “Combat
REFERENCES with Black hole attack in AODV routing protocol in
[1] Todd R. Andel, Alec Yasinsac, “Surveying Security MANET ”, First Asian Himalayas International
Analysis Techniques in MANET Routing Protocols”, Conference on Internet, page(s) 1-5, 3-5 Nov. 2009 .
IEEE Communications Surveys, 4th Quarter, No.4, 2007. [18] L. Zhou and Z.J. Haas, “Securing Ad Hoc Networks”,
[2] N.H Saeed, M.F Abbod, H.S Al-Raweshidy, “Modeling IEEE Network Journal, No. 6, page(s):24–30, 1999.
MANET Utilizing Artificial intelligence”, Second [19] B. Awerbuch, D. Holmer, C. Nita-Rotaru, and H. Rubens,
UKSIM European Symposium on Computer Modeling “An On-demand Secure Routing Protocol Resilient to
and Simulation, EMS '08, Page(s):117–122, 8–10 Sept. Byzantine Failures”, Proceedings of the ACM Workshop
2008. on Wireless Security, page(s):21–30, 2002.
[3] Y. Hu, A. Perrig, and D. Johnson, “Packet Leashes: A [20] Y. Hu, A. Perrig, and D. Johnson, “Ariadne: a secure on-
Defense Against Wormhole Attacks in Wireless Ad Hoc demand routing protocol for ad hoc networks”, Proc. of 8th
Networks”, Proc. of IEEE INFOCOM, 2002. ACM Mobile Computing and Networking (MobiCom’02),
[4] A. Menezes, P. Oorschot, and S. Vanstone, “Handbook of page(s):12–23, 2002.
Applied Cryptography”, CRC Press, 1996. [21] Z. Haas and M. Pearlman, “The Performance of Query
[5] A. Kush, C.J Hwang, “Proposed Protocol for Secured Control Scheme for The Zone Routing Protocol”,
Routing in Ad Hoc Networks”, International Association ACM/IEEE Transactions on Networking, pages:427–438,
of Computer Science and Information Technology Spring August 2001.
Conference, IACSITSC '09, Page(s):76–81, April 2009. [22] Yang Xiao, Xuemin Shen and Ding-Zhu Du, “Wireless
[6] M. Bouhorma, H. Bentaouit, A. Boudhir, “Performance Network Security.”, Springer, 2007.
comparison of ad-hoc routing protocols AODV and [23] K. Sanzgiri, B. Dahill, B.N. Levine, C. Shields, E.M.
DSR.”, International Conference on Multimedia Belding Royer, “Secure routing protocol for Ad Hoc
Computing and Systems, 2009. ICMCS '09, Page(s):511– networks”,10th IEEE International Conference on
514, April 2009. Network Protocols, Dept. of Computer Sciences,
[7] Y. Hu, A. Perrig, and D. Johnson, Ariadne, “A Secure On- California University, Santa Barbara, CA, USA.
Demand Routing for Ad Hoc Networks.”, Proc. of Page(s):78-87, 2002.
MobiCom 2002, Atlanta, 2002.
[8] Zhan Huawei, Zhou Yun, “Comparison and Analysis
AODV and OLSR Routing Protocols in Ad Hoc
Network”, 4th International Conference on Wireless
Communications, Networking and Mobile computing
2008, WiCOM '08, Page(s):1 – 4, 12-14 Oct. 2008.
[9] P. Jacquet, P. Muhlethaler, T. Clausen, A. Laouiti, A.
Qayyum, L. Viennot, “ Optimized Link State Routing
Protocol”, IEEE INMIC Pakistan 2001.
[10] Irshad Ullah, Shoaib Ur, “Analysis of Black Hole Attack
on MANETs Using Different MANET Routing
Protocols.”, Master Thesis at School of Computing,
Blekinge Institute of Technology, 2009.
[11] Dokurer, Semih. “Simulation of Black hole Attack in
Wireless Ad-hoc Networks.”, Master's thesis, Atılım
University, September 2006.
[12] H. Deng, W. Li, D.P. Agrawal, “Routing security in
wireless Ad-Hoc networks” Cincinnati University, Ohio,
USA, IEEE Communications Magazine, Vol.40,
page(s):70-75, Oct. 2002.
[13] S. Lu, L. Li, K.Y. Lam, L. Jia, “SAODV: A MANET
Routing Protocol that can Withstand Black Hole Attack",
International Conference on Computational Intelligence
and Security, 2009.

82