AUTHENTICATION EMPLOYING NEIGHBOR TABLES IN
MOBILE AD HOC NETWORKS
Dorsaf Azzabi, Kunihito Tanaka, Ushio Yamamoto and Yoshikuni Onozato
Graduate School of Engineering, Gunma University 1-5-1 Tenjin-cho, Kiryu, Gunma, 376-8515, Japan
Keywords: MANET, transmission range, mobility, neighbor authentication.
Abstract: In this paper we investigate the authentication mechanism in mobile ad hoc networks (MANET).
Developing a trustworthy authentication mechanism in a dynamic and volatile ad hoc mobile setting is a
complex task. The goal of this paper is to propose an authentication mechanism to validate the existence of
neighborhood in MANET. We focus on common adjacent nodes which can authenticate and trust a mobile
node. Simulation results show the existence of common adjacent nodes that are able to establish trust and
obtain the effective range of neighbor authentication in terms of transmission range and node density.
1 INTRODUCTION
Ad-hoc networks are based on naive “trust-your-
neighbor” relationships. Such relationships
originate, develop and expire on the fly and usually
have short life spans. As the overall environment in
such a network is cooperative by default, these trust
relationships are extremely susceptible to attacks.
For a number of reasons, including better service,
selfishness, monetary benefits or malicious intent,
some nodes can easily mould these relationships to
extract desired goals. Also, the absence of fixed trust
infrastructure, limited resources, ephemeral
connectivity and availability, shared wireless
medium and physical vulnerability, make trust
establishment virtually impossible. To overcome
these problems, trust has been established in ad-hoc
networks using a number of assumptions including
pre-configuration of nodes with secret keys, or
presence of an omnipresent central trust authority
Authentication mechanisms in mobile ad hoc
networks (MANET) are very important since the
role of MANET is to get information through nodes
by ad hoc mode in the network field. In MANET the
communication among nodes to relay information is
exposed to various security attacks. There have been
many works for authentication. However
authentication for MANET especially for moving
nodes is usually left to future research. L. Lazos et
al. (Lazos, 2005) describe typical security threats
such as the wormhole attack and the Sybil attack.
They propose the SeRLoc for the secure localization
in the presence of these threats. They show, based on
their performance evaluation, that it is robust against
various sources of error. But they do not consider
mobile nodes.
LITEWORP (Khalil, 2005) allows detection of
the wormhole by isolating the malicious nodes. The
detection is based on a local monitoring using
neighbor lists. A node will not receive a packet from
a node that is not a neighbor nor forward to a node
which is not a neighbor. It is not assessed whether
the LITEWORP is applicable to the mobile
situations. Azzabi and Uchihara et al. (Azzabi, 2007)
proposed their neighbor authentication mechanism
for neighboring nodes that can cope with nodes
mobility and hostile environments.
Importance of secure neighbor discovery in
wireless networks is explained by Poturalski et al.
(Poturalski, 2008). Time-based protocols and time-
and location-based protocols to achieve secure
neighborhood discovery are proposed and proved in
a formal model. Their protocols are applicable even
in low-density networks.
In this paper we focus on the neighbor
authentication mechanism (Azzabi, 2007) utilizing
neighbor tables which might be practical in ad hoc
mobile environments, since developing a
trustworthy authentication mechanism in a dynamic
and volatile ad hoc mobile setting is a complex task.
We will demonstrate the proposed neighbor
authentication mechanism in this paper operate in
such hostile environment. The goal of this paper is
to extend effective range of the neighbor
47
Azzabi D., Tanaka K., Yamamoto U. and Onozato Y. (2008).
AUTHENTICATION EMPLOYING NEIGHBOR TABLES IN MOBILE AD HOC NETWORKS.
In Proceedings of the International Conference on Wireless Information Networks and Systems, pages 47-52
DOI: 10.5220/0002027300470052
Copyright
c
SciTePress
authentication mechanism in MANET. We
investigate how to increase common adjacent nodes
which can authenticate and trust a mobile node.
Simulation results show the existence of common
adjacent nodes that are able to establish trust and
obtain the effective range of neighbor authentication
in terms of transmission range and node density.
2 NEIGHBOR
AUTHENTICATION
MECHANISM
2.1 Network Model
We assume a MANET shown in Fig.1 where mobile
nodes are randomly allocated and move according to
the random waypoint model (Bettstetter, 2004)
where the length of each leg is limited within a
mobility range M. Transmission range of each node
is fixed as r. Packets are transmitted through nodes
in an ad hoc mode. All packets are gathered at the
Base Station (BS). The network topology changes
dynamically since the nodes are mobile.
All mobile nodes NN in the MANET have two
kinds of their neighborhood tables: Physical
existence table (P-table) and Authentication table
(A-table). P-table lists all nodes within the
transmission range of the node. A-table lists all
authenticated nodes within the transmission range of
the node. Furthermore a mobile node constructs
Enlarged authentication table (E-table) based on the
A-table. E-table shows all authenticated nodes of the
node and its authenticated nodes within the
transmission range of the node. Therefore E-table is
constructed by combining A-table of the node and
the ones of the authenticated nodes. When a node
moves, the neighborhood tables are updated. A
received packet is authenticated based on the node
which has sent it. A node does not receive a packet
from a node that is un-authenticated nor forward to a
node which is not a neighbor (Khalil, 2005). Any
node in the neighborhood tables has to be
authenticated in order to keep a secure
communication. That is why, for a malicious node to
attack the MANET, its only opportunity is to enter
in the neighborhood table.
Nodes in a MANET are classified as the
following three nodes: authenticated nodes,
unapproved nodes and isolated nodes. Authenticated
nodes have been recognized in the network by their
neighborhood relationships. Unapproved nodes are
not approved by the neighbor authentication
mechanism. Isolated nodes have not been in
communication with other nodes.
We assume that all nodes initially distributed are
not malicious nodes. We investigate authentication
mechanism for new comers after initial settings so
that we can keep secure communications for mobile
nodes in hostile environments.
Figure 1: A MANET where packets are transmitted
through nodes in an ad hoc mode. All packets are gathered
at the BS.
2.2 Neighbor Authentication
Mechanism for MANET
In order to authenticate a node, it is necessary to
check the existence of common adjacent nodes and
examine all nodes within its transmission range.
This authentication process is executed employing
two neighborhood tables; P- and A-tables, and
enlarged authentication table E-table. Theses tables
are initially constructed when all nodes are allocated
at initial position. In this process, P-table is
constructed at first by exchanging beacon including
the node ID. Then, each node copies its own P-table
to A-table, because all nodes are initially reliable by
our assumption. Next, the node asks all nodes in its
A-table to reply its A-table. Combining all received
A-tables, the node construct E-table.
When a node moves into the transmission range
of another node, three tables need to be updated.
Especially, updating A-table needs the existence of
common adjacent nodes. Common adjacent nodes
only can authenticate newly detected node.
The authentication mechanism is explained in
Fig.2. Assume the two nodes n
x
and n
y
where n
y
is
moving into the transmission range of n
x
as shown in
Fig.2. We explain how n
y
constructs three tables: P-,
A- and E-tables. Both n
x
and the neighboring nodes
of n
y
always monitor the way n
y
behaves. They may
only give the authenticity of n
y
. When n
y
moves,
there are two common adjacent nodes of n
x
and n
y
,
WINSYS 2008 - International Conference on Wireless Information Networks and Systems
48
i.e. n
1
and n
2
as shown in Fig.2. They mediate
between n
x
and n
y
. n
x
searches E-table and finds that
n
1
and n
2
are common adjacent nodes, namely, they
authenticates both n
x
and n
y
.
Figure 2: Common adjacent nodes after the movement of
n
y
.
The authentication algorithm is given as follows:
Step1) n
y
(moving node) broadcasts a randomly
generated nonce value (Lazos, 2005).
Step2) n
x
receiving the broadcast from n
y
responds
with the ID of n
x
and the received nonce value.
Step3) n
y
realizes that it is entering the transmission
region of n
x
. And n
y
searches for the common
adjacent nodes from E-table. If the common
adjacent node exists, n
y
asks a confirmation of
authenticating n
x
to the common adjacent nodes.
In this case, n
x
also asks a confirmation of
authenticating n
y
to the common adjacent nodes.
When the common adjacent node receives
confirmation packets from both n
y
and n
x
, it
recognizes that they are justified nodes, then it
replies to them. After receiving this reply, n
y
appends n
x
to the A-table. If n
y
can’t receive such
reply, it does not add n
x
to the A-table.
Step4) n
y
asks all nodes in its A-Table to send back
their A-tables to ny.
Step5) After receiving all A-tables, n
y
construct E-
table by combining all A-tables so far received.
In order to extend effective range of the neighbor
authentication mechanism, in Step 4) in the above
algorithm, we enlarge asking range of n
y
to
neighbors of authenticated nodes, which we call
one-step neighbors. Furthermore we enlarge asking
range of n
y
to one by next neighbors, which we call
two-step neighbors. When authenticated nodes are
sparsely located, this extension of authentication
does not increase authenticated nodes so much.
However in rather dense node distributions, we will
show the effect of the extension in the next section.
We also introduce the following two control
messages to isolate the malicious nodes when we
detect attacks such as the wormhole: an accusation
message and a suspicion message (Azzabi, 2007).
The accusation message is sent to the BS when a
node finds a malicious node in the neighborhood list.
BS deletes the malicious node if BS receives a
number of accusation messages beyond the
threshold. The suspicion message is sent to the BS
when a node doubts the neighborhood relationship.
BS asks, in that case, all nodes to delete the
suspicious node from their neighborhood lists.
In order to detect the malicious node with the
collaborative detection in the above mechanism,
common adjacent nodes are necessary so that
authenticity of the node is verified. We take in
consideration the condition that nodes are not
isolated in mobile situations, where the proposed
authentication mechanism works and the trust is
placed.
2.3 Countermeasure Against Malicious
Nodes
In this paper we assume that all nodes initially
distributed do not include any malicious node so that
all neighbor tables are correct at the initial state. We
investigate authentication mechanism works sound
for updating in case when any malicious nodes
attack after initial settings we can keep secure
communications for mobile nodes in hostile
environments.
If one malicious node attacks by forging P-table,
or A-table, the above mechanism can detect illegal
updating of the tables. Even though unjustly
fabricated authentication node can be detected by
legitimate nodes. When two malicious nodes attack,
if we can separate the two compromised nodes, the
above mechanism can detect. However if the two
nodes are in proximity or collude with such as in a
wormhole link, we cannot detect them (Tanaka,
2008). This is left for future research.
We will describe more about these control
mechanisms in our future work. So far, the scope of
this paper is to verify via simulation the idea of
neighbor authentication, that is if a neighbor exists
or not, because it is a fundamental prerequisite, in
order to implement the before mentioned trust
control messages.
AUTHENTICATION EMPLOYING NEIGHBOR TABLES IN MOBILE AD HOC NETWORKS
49
3 SIMULATION SCENARIO AND
RESULTS
3.1 Simulation Scenario
We consider simulating, as mentioned above, the
authentication mechanism of adjacent nodes. The
proposed authentication mechanism needs
cooperation and coordination scenarios in ad hoc
mobile environments so that isolated nodes should
not be left so. We consider authenticated nodes as
the nodes that have been recognized in the network
by their neighborhood relationships. Isolated nodes
are the nodes that have not been in communication
with other nodes. Unapproved nodes are defined as
nodes which are not authenticated by any node
within its transmission range even though there are
nodes within its transmission range. The number of
isolated nodes and the number of unapproved nodes
are indeed important parameters to evaluate the
performance of the neighbor authentication
mechanism.
We randomly distributed NN nodes within
100x100 rectangular area where we set NN=100,
200, 300, 400, and 500, and the transmission range
r= 1, 2, 3, …, 9. We assume that each node moves
according to the random waypoint model (Bettstetter,
2004) within a mobility range M for each movement,
where M= 1, 2, 3, …, 9. We measure the number of
authenticated nodes, the number of isolated nodes,
and the number of unapproved nodes.
3.2 Simulation Results
The percentage of authenticated nodes among NN
nodes for various values of mobility range when r=9
is shown in Fig. 3. As NN increases, the numbers of
authenticated nodes increase. When mobility range
is much higher, the percentage of authenticated
nodes is very low. It is very understandable that if
the nodes are very mobile, the communication
environment is not stable unless the transmission
range is set proportionally to cope with the high
mobility rate. This is shown through the slice
authentication distribution improvement when r=9.
The percentage of isolated nodes for various
values of mobility range is shown in Fig. 4. As NN
increases, the numbers of isolated nodes decrease
when mobility range is much smaller. In Figs. 5 and
6, we notice that from a transmission range from 1 to
9 the number of isolated nodes tends to decrease for
various mobility ranges varying between 1 and 9.
The percentage of isolated nodes decreases as NN
increases. The bigger the number of nodes is and the
larger the transmission range is, the more common
adjacent nodes are kept in the network where a node
is not fixed in one position.
The numbers of unapproved nodes for various
values of mobility range are shown in Figs.7 and 8.
As r increases, the number of unapproved nodes
increases until the neighbor authentication
mechanism works. In Fig.7, as NN=100, the number
of unapproved nodes increases especially for a
higher mobility range (from 2 and above) because
the nodes can cover more transmission areas so that
chances of authentication occurs but failed. However
when the number of nodes is much higher as shown
in Fig. 8, where NN=500, the number of unapproved
nodes decreases for increasing transmission ranges.
This is because of the authentication mechanism the
number of unapproved nodes decreases. In Fig. 8, all
curves seem to have the similar tendency.
In Fig. 9, isolated, authenticated and unapproved
nodes for the cases of one-step neighbor and two-
step neighbor are depicted in the four sections where
LEFTUP shows position of nodes, RIGHTUP shows
one-step neighbor case, RIGHTDOWN shows two-
step neighbor case, and LEFTDOWN shows
augmented relations because of increased
authentication process where NN=500, r=10 and
M=9. There are 358 authenticated nodes. In the one-
step neighbor case, 101 authenticated nodes increase
because of enlarging to one-step neighbor. There are
still 41 unapproved nodes.
Our algorithm checks for and confirms the idea
that a node is authenticated by its neighbor. Our
graphs show two main parameters in this simulation
that influence the authentication process: node
mobility and transmission range. The higher the
transmission range is, the better the percentage of
authenticated nodes is. When node density increases,
and the transmission range increases, the trust
increases. The parameters in form of authentication
percentage and isolated nodes ratio indicate the
performance measures.
So far we have been simulating the first aspect of
our algorithm. Using the neighborhood table, if a
node has a neighbor, it is authenticated, if it has not,
it is not. We conclude that only neighboring nodes
can authenticate an existent node. We tested the
neighbor authentication method to determine under
which parameters it works (the region, the
transmission range etc.). We verified it for a
different range of nodes mobility and transmission
rage.
WINSYS 2008 - International Conference on Wireless Information Networks and Systems
50
4 CONCLUSIONS
We simulated an authentication mechanism for
establishing trust in MANET. Our aim is to
investigate and describe the fundamentals behind
nodes interaction while considering their mobility
range with different transmission ranges. As
common adjacent nodes increase, more nodes are
authenticated under the proposed mechanism.
Through our simulation study we show in which
cases the authentication is achievable so that we
could in a next level introduce the trust control
messages (accusation message and suspicion
message) in order to establish a trust relationship
among the neighboring nodes.
This level will be explored in our upcoming
paper. We will try to simulate the impact of these
messages on the behavior of the MANET. We also,
ought to mention that the interference problems
associated with the large number of nodes as well as
hidden terminal problems with enlarged
transmission ranges have to be reviewed in our
future study.
REFERENCES
Lazos, L., Poovendran, R., 2005. SeRLoc: Robust
localization for wireless sensor networks, ACM
Transactions on Sensor Networks, vol.1, no.1, pp.73-
100.
Khalil, I., Bagchi, S., Shroff, N.B., 2005. LITEWORP: A
lightweight countermeasure for the wormhole attack in
multi-hop wireless networks, In the International
Conference on Dependable Systems and Networks
(DSN), pp.612-621.
Azzabi, D., Uchihara, N., Tanaka, K., Onozato, Y., 2007.
Simulation of authentication mechanism in MANET,
In The Second Asia-Pacific Symposium on Queueing
Theory and Network Applications (QTNA), pp.289-
296.
Bettstetter, C., Hartenstein, H., Perez-costa, X., 2004.
Stochastic properties of the random waypoint mobility
model, Wireless Networks, vol.10, pp.555-567.
Poturalski, M., Papadimtratos, P., Hubaux, J.P., 2008.
Secure neighbor discovery in wireless networks:
Formal investigation of possibility, In ASIACCS’08,
pp.189-200.
Tanaka, K., Onozato, Y., 2008., Node management
method using authentication process with neighbor
table in ad hoc network, IPSJ SIG Technical Report,
2008-MBL-44, pp.237-244.
Figure 3: Percentage of authenticated nodes for various
values of mobility range when r=9.
Figure 4: Percentage of isolated nodes for various values
of mobility range when r=9.
Figure 5: Number of isolated nodes versus transmission
range for various values of mobility range with NN = 100.
AUTHENTICATION EMPLOYING NEIGHBOR TABLES IN MOBILE AD HOC NETWORKS
51
numberofis olated nodes
(NN= 500)
0
50
100
150
200
250
300
350
400
450
500
123456 789
transmiss ion ra ng e
numb erof is orated nodes
Mobilit y r a n g e = 1
Mobilit y r a n g e = 2
Mobilit y r a n g e = 3
Mobilit y r a n g e = 4
Mobilit y r a n g e = 5
Mobilit y r a n g e = 6
Mobilit y r a n g e = 7
Mobilit y r a n g e = 8
Mobilit y r a n g e = 9
Figure 6: Number of isolated nodes versus transmission
range for various values of mobility range with NN = 500.
Figure 7: Number of unapproved nodes versus
transmission range for various values of mobility range
with NN = 100.
numberofunapprovedno des
(NN=50 0)
0
50
100
150
200
250
300
350
400
450
500
123456 789
transmissionra n g e
numberof unapprovednodes
Mobilit y ra n ge = 1
Mobilit y ra n ge = 2
Mobilit y ra n ge = 3
Mobilit y ra n ge = 4
Mobilit y ra n ge = 5
Mobilit y ra n ge = 6
Mobilit y ra n ge = 7
Mobilit y ra n ge = 8
Mobilit y ra n ge = 9
Figure 8: Number of unapproved nodes versus
transmission range for various values of mobility range
with NN = 500.
Figure 9: Isolated, authenticated and unapproved nodes for
increasing authenticated nodes are depicted in the four
sections where LEFTUP shows position of nodes,
RIGHTUP shows one-step authentication process,
RIGHTDOWN shows two-step authentication process,
and LEFTDOWN shows augmented relations because of
increased authentication process(NN=500, r=10, M=9).
WINSYS 2008 - International Conference on Wireless Information Networks and Systems
52