A Novel Multipoint Relay based Secure Routing in MANET

International Journal of Network Security & Its Applications (IJNSA), Vol.4, No.6, November 2012
DOI : 10.5121/ijnsa.2012.4610 133
A Novel Multipoint Relay based Secure Routing
in MANET
Himadri Nath Saha Dr.Debika Bhattacharyya Dr.P.K.Banerjee
IEM,Kolkata,WB IEM,KolaKata,WB JU,Kolkata,WB
him_shree_2004@yahoo.com, bdebika@yahoo.com, pkbju10@yahoo.com
Abstract— Security in routing is a challenging issue in mobile ad-hoc (MANET) network because of its
open nature, infrastructure less property, mobility and energy constraints. Messages typically roam in
multi-hopped fashion and nodes may be powered by limited energy source and with limited physical
security. So we proposed a new scheme which is significantly different from others available schemes to
provide security during routing in mobile ad hoc networks. In this paper, our proposed scheme, Secure
Multipoint Relay based Routing in MANET (SMRR) provides routing based on trust, which is an integer
value that helps to select Multipoint Relay (administrator) inside the network for routing. We have also
implemented the message confidentiality and integrity in our proposed scheme. Our simulation results
show the robustness, reliability and trustworthiness of our scheme.
Keywords- MANET, SMRR, trust, administrator, digital signature, willingness function, olsr,
secure routing
I. INTRODUCTION
Mobile Ad hoc network (MANET) can operate in a self-organized and non predefined
infrastructure. The nodes in such networks can communicate with each other through direct
wireless links or multi-hop routing. It has been used in a wide range of applications ranging
from a battlefield to the user’s living room. Many efficient routing protocols have better
network performance however; they are more vulnerable to security threats. Ad hoc network
has faced even more serious security problems as compared to traditional wireless networks.
Several security solutions require a centralized server for key distribution or a secret
understanding between communicating entities. This lack of infrastructure has posed serious
threats as far as routing security is concerned. Secondly, the vulnerability of the nodes towards
physical compromise gives rise to serious internal threats within the network which make the
issues of authentication, integrity and confidentiality even more challenging than conventional
wireless networks. Thirdly, the next important characteristic of a mobile ad hoc network is that
the topology of the network changes dynamically. Thus, any security model, based on a fixed
architecture cannot be used in such a scenario. This assumption is also coupled with pre-
configuration of nodes with encryption keys prior to joining the network. Public key
cryptography with digital signature makes the network stronger to stand up against the attackers
and secure communication is assured. However, due to the limitation of battery energy of
mobile nodes, methods of prolonging the lifetime of nodes as well as the network become the
key challenge in MANET. The performance of MANET depends on the routing scheme
employed and the traditional routing protocols do not work efficiently in MANET. Developing
routing protocols for MANET has been an extensive research area in recent years, and many

134
proactive, reactive and hybrid protocols have been proposed from a variety of
perspectives[1]. In rest of the paper, Section II describes the Working Methodology of SMRR,
while section III explores related works in this domain. Section IV describes our proposed
algorithm. In section V, we present the proposed packet format while section VI gives us the
picture of the performance evaluation. Lastly, section VII deals with all future work and section
VIII express the conclusion in relation to this domain.
II. WORKING METHODOLOGY OF SMRR
Our proposed routing algorithm, SMRR (Secure Multipoint Relay based Routing) is a pro-
active scheme inspired by OLSR [2]. This algorithm uses signed acknowledgement (signing
based on asymmetric key cryptography) to establish trust. Key distribution is out of the scope
of this paper and any popular key distribution methodology can be followed. This scheme uses
administrator to disperse all the packets to the networks through the next hop administrator.
Admin node is a minimal subset of all nodes that can form a fully connected network. It
consists of all the administrators which can reach out to all the neighbor nodes. This admin
node selection depends on symmetric link, node coverage, willingness of that node & Trust.
SMRR scheme not only encrypts Transport layer data but also control messages sent by SMRR
itself. Application data and control traffic are handled in separate way.
III. RELATED WORKS
Till date many secure routing protocols have been developed. SOLSR [3] (based on OLSR) has
used symmetric key for encrypting all data and control packets but Trust concept has not been
implemented yet. While TAODV [4] (based on AODV [5]) does not use any encryption
technique but it uses the trust factor. Again when considering the case of SAODV [6], it uses
public key cryptography and digital signature to protect RREQ & RREP messages [7]. It also
uses hash-chain to authenticate hop-count of each message. Few secured routing protocols like
SRP [8], FTAODV [9], Ariadne [10] and others [11][12][13] have similar kind of approaches
and so are not included in this paper. So in our scheme we have blended the concepts of both
crytography and trust factor to enhance its security. We are using digital signature to verify
each acknowledgement so forged packet generation is not possible.
IV. PROPOSED ALGORITHMS
In this paper, we are proposing few new mechanism for calculating administrator based on
willingness & trust of a node by taking care of more exhaustive parameters so that numbers of
admin nodes are kept to minimum (as per basic OLSR) and routing becomes secure. Our
algorithm forces same return path as while in sending. Only when it is absolutely necessary,
admin node can switch from one node to another, offloading their job to other one, increasing
network runtime. Below are the important algorithms that each node runs individually to
maintain a secure and reliable network.
First we will discuss about Admin node selection algorithm. The value of willingness will be
derived from next algorithm given in part B and trust from algorithm given in part C of section
IV.
A. Dynamic Willingness calculation:
Our algorithm takes a weighted sum of battery power of a node, coverage area and reliability of
the node while calculating willingness value. The weighted values are experimentally tested and
optimized. The battery power in MANET is crucial because more routing will result battery
power depletion and node may stop working. So it has been assigned the highest weighted value.
Coverage is next important factor because more coverage ensures better routing and third
important factor is reliability which is very important for implementation.

135
Willingness (P, C, R) =
(0.75 * P) + (0.15 * C) + (0.1 * R) (2)
Where,
P: power available for that node (in %)
C: coverage (in %)
R: reliability of the node (in %)
Power (P) is defined as:
P = (current node power/rated capacity of the node)*100 (3)
Coverage (C) is defined as:
C = (no of 1-hop neighbors of that node / no of 2-hop neighbors of nodes that want to select this
node as its ADMIN)*100 (4)
Reliability (R) is calculated from various sensor inputs regarding outside environment condition.
R ranges from 0% to 100% depending upon the node’s position.
R = {0% … 100%} (5)
B. Admin Node selection:
This algorithm selects the administrator node which can cover most of the 2-hop neighbour of its
selector. Selection also takes care of willingness and trust value of node. In case of tie, node with
higher trust/power will be selected.
Few definitions:
• ADMIN(x): Admin set of node x which is running this algorithm.
• N1(x): One hop neighbor set of node x (symmetric neighbors)
• N2(x): Two hop neighbor set of node x [symmetric neighbors of nodes in N(x)].The
two hop neighbor set N2(x) of node x does not contain any one hop neighbor N(x) of
node x.
• D(x,y) : Degree of one hop neighbor node y (where y is a member of N1(x) -- means y
belongs to N1(x)), is defined as the number of symmetric one hop neighbors of node y
EXCLUDING the node x and all the symmetric one hop neighbors of node x, i.e.,
D(x, y) = N(y) - {x} – N1(x) (1)
• W = Current willingness value of the node. [can range from 0 to 7]
• T = Current trust value of the node. [can range from 0 to 10]
• Trust_Threshold = Implementation dependent [we choose 2]
Initialization:
1. Initialize Node_Trust table with default trust value 3 for each node.
2. Initialize PATHLIST = [].

136
Algorithm:
Step1: Start with an empty ADMIN(x) set.
Step2: Calculate D(x,y), where y is a member of N1(x), for all nodes in N1(x) (put for all
+ve sign)
Step3: First select as ADMINs those nodes in N1(x) which provide the "only path" to reach
some of the nodes in N2(x). [Trivial case]
Step 4: For each node in N1(x)
{
4.1. SELECT current node as a ADMIN as per table 1.
4.2. While if some nodes still exists in N2(x)
that is not covered by ADMIN(x):
{
For each node in N1(x), calculate the
no. of nodes in N2(x) which are not yet
covered by ADMIN(x) and are reachable
through this one hop neighbor of x.
}
4.3. Select as an ADMIN that node of N1(x) which
reaches the maximum number of uncovered nodes
in N2(x) & refer table 1.
4.4. If a tie occurs, select that node as ADMIN
whose D(x,y) is greater & refer table 1.
}
Step 5: To optimize, process each node y in ADMIN(x), one at a time, if ADMIN(x) - {y}
still covers all nodes in N2(x) then remove y from ADMIN(x).
Step 6: After that Convert the link between node x and ADMIN as SYM_LINK to
ADMIN_LINK
Step 7: Exit
NODE 1: NODE 2:
TRUST
(T1)
%
POWER
(P1)
%
TRUST
(T2)
%
POWER
(P2)
%
SELECTION
SPCL SPCL SPCL SPCL WHEN BOTH THE NODES HAVE THE SAME
VALUES THEN SOURCE NODE CAN BROADCAST
THE MESSAGE TO THE NETWORK THROUGH
EITHER OF THE NODES. EITHER NODE1 OR NODE2
L L L H NODE2
L L H L NODE2
L L H H NODE2
L H L L NODE1
L H L H
(IF P1>P2 THEN NODE1 ELSE NODE2)
ELSE
(IF P1==P2 THEN IF T1>T2 THEN NODE1 ELSE
NODE2)
L H H L (IF P1-TH_PWR>T2-TH_TR & T1-TH_TR > P2-
TH_PWR THEN NODE1)
ELSE
(IF T2-TH_TR>P1-TH_PWR & P2-TH_PWR > T1-
TH_TR THEN NODE2)

137
L H H H NODE2
H L L L NODE1
H L L H (IF P1>P2 THEN NODE1 ELSE NODE2)
H L H L (IF T1>T2 & P1-TH_PWR>P2_TH_PWR THEN NODE1)
ELSE
(IF T2>T1 & P2-TH_PWR>P1_TH_PWR THEN NODE2)
ELSE
NODE1
H L H H NODE2
H H L L NODE1
H H L H NODE1
H H H L NODE1
H H H H ( IF P1>P2 THEN NODE1 ELSE NODE2 )
Table 1: Admin Selection in case of tie
C. Digital Signature And Trust Value Calculation
I) Sender node’ job
Step 1: Encrypt the message with Public Key of destination
ENC_MSGENCRYPT (PlainText_MSG)
Step 2: Calculate HASH VALUE for ENC_MSG
HASH_VAL  HASH (ENC_MSG)
Step 3: Create a entry for PATHLIST table with following Data :
< HASH_VAL, CRNT_N_ID + MSG.PATH >
Step 4: Set a TIMER for this entry with timeout value T.
[Value of T is scheme dependent]
II) Original message (DATA)
• If the Node is Intermediate Node:
Step 1: Receive the encrypted message.
Step 2: Append next Hop ID to the variable Path.
Step 3: Update the packet size to reflect the modified Path.
Step 4 (a): Calculate
HASHVALHASH (MSG.ENC_MSG)
Step 4(b): Store the following entry in PATHLIST table:
<HASHVAL, CRNT_N_ID+MSG.PATH>
Step 5: Set TIMER with T sec Timeout for this entry.
Step 6: Forward the updated encrypted message.
• If the Node is Intended Receiver Node (DEST):
Step 1: Extract PATH from received message:
PATH  MSG.PATH
Step 2: Extract message:
MSGDECRYPT (MSG.ENC_MSG)
Step 3: Create a HASH value for ACK message generation:
HASHVAL_CHASH (MSG.ENC_MSG)

138
Step 4: Sign the ACK message:
SIGN ENCRYPT (HASHVAL_C, PVT_KEY_DEST)
Step 5: Transmit the ACK message with SIGN to Previous
Node found in PATH.
• On expiration of time out for particular entry in path list
Step 1: Extract path from Time out entry:
PATH_T  Timeout Entry. PATH
Step 2: Decrease TRUST value for last node in PATH_T by 1
Step 3: Remove the entry from PATHLIST table.
III) SMRR specific messages (CONTROL)
• Sending HELLO, MID, TC Message
Step 1: Create message as per existing OLSR scheme.
Step 2: Before sending it to network layer (such as IP), message content is encrypted with
current node’s Private Key as per packet format shown in Fig 2. Scheme & Algorithm
field are set as per chosen parameter.
• Receiving HELLO, MID, TC Message
Step 1: Receive encrypted message from network layer.
Step 2: Before processing the received packet it is decrypted using originator node’s Public
Key and original message format is restored. This message is processed and forwarded
(encrypted) as per existing scheme.
• For Acknowledge Message
Step 1: Receive the ACK packet.
Step 2: Extract the encrypted hash value.
HASHVAL_RACK.ENC_HASH
[Where ACK.ENC_HASH = ENC (HASH (ENC_MSG), PRK_DEST))]
Step 3: Find entry in PATHLIST with HASHVAL_R
Step 4 (a): If entry found
i) Extract stored path
E_PATH Entry.PATH
ii) If
the last node in E_PATH is Sender Node of this ACK packet
then
increase TRUST of Sender Node by 1.
Else
decrease TRUST of Sender Node by 1 and discard the packet.
Remove this entry from PATHLIST.
Round off TRUST to within 0 to 10.
GOTO Step 5
iii) Update the E_PATH of ACK packet by removing the Sender Node ID.
iv) Forward the ACK message to the previous hop in E_PATH.

139
Step 4 (b): If
entry not found decrease TRUST of Sender node
by 1(round off within 0 to 10) and discard the packet.
GOTO Step 5
Step 5: Done.
• On expiration of time out for particular entry in path list
Step 1: Extract path from Time out entry:
PATH_T  Timeout_Entry.PATH
Step 2: Decrease TRUST value for last node in PATH_T by 1
Step 3: Remove the entry from PATHLIST table.
V. PACKET FORMAT
Our scheme only works to create and maintain a valid route table in host Operating System. It
is host Operating System’s responsibility to correctly route the packet to actual destination (as
in OLSR). SMRR specific messages are handled by SMRR Network layer module.
Application data are directly sent by IP layer (or any underlying Network layer
protocol). But to calculate Trust value, we need to handle Application layer data coming from
Transport layer separately by SMRR module. Transport layer data are encrypted in network
layer with destination node’s public key. In each hop, Message Path field is updated to add the
current hop address. In receiver side, before sending the data to Transport layer, it is decrypted
by destination node’s private key and an ACK message is created from the received packet as
per the above algorithm. Intermediate nodes can’t decrypt the data packet as they don’t know
the private key of intended receiver node. Actual Transport layer data are packed as per
following packet format:
Figure 1: Data Packet Format
SMRR’s native messages (i.e. ACK, TC, MID, HELLO) are sent/received & processed
directly by SMRR module. As multiple SMRR messages are piggybacked (as in OLSR) into a
single packet, each message part will contain separate encrypted message content. All message
type except ACK_MESSAGE will be encrypted with originator node’s private key. Scheme &
Algorithm field is used to send SMRR specific data. Accordingly, Message Size & Packet
Length is updated.

140
Figure 2: SMRR Message Packet Format
For ACK packet (Message Type = ACK) , instead of Encrypted Message part, signature is send:
Figure 3: ACK packet format
Figure 4: Flow of packets
VI. PERFORMANCE EVALUATION
A. Simulation Environment
We used OLSR protocol implementation from Niigata University for Glomosim [14] [15]
.
Parameter Value
Terrain Dimension (600x500) sq. meter
Simulation Time 500 minutes
Channel Noisy
Noise Figure 10 dB
Radio Frequency 2.4 Ghz
Radio Receive Threshold -65.046 dBm
Radio Transmit Power 22.5 dBm
Intermediate Node:
1) Update msg_path
only & PATHLIST
Actual Receiver
Node:
1)Decrypt the data
using destination
Node’s Private Key
2)Send the decrypted
data up into transport
layer
3)ACK(control)
creates and send
back to previous
node.
Pac
ket
Typ
CONTRO
L
APP
DATA
1) Generally, Control
messages are for next
hop only, only TC
messages are
forwarded to next hop.
2) Decrypt using
originator node’s
public key.
3) Process Control
message as per OLSR
standard, except ACK
message.
4) For ACK message,
follow SMRR
,calculatCALC
TRUST!!!

141
Node Placement Random
Mobility Speed 0-10 m/s
MAC Protocol 802.11
MAC Propagation Delay 1000 ns
Bandwidth 11 Mbps
Routing Protocol OLSR, SMRR, SAODV
Number of Interface per node 2
Rated Battery Power
(each node)
1500 mAh
Data Packet Type FTP, CBR
Data Packet Size 2044 byte
Cryptographic algorithm RSA (512 bit)
Table 2: Simulation parameters
To simulate the proposed algorithm we used Glomosim 2.03 network simulator [14].
Glomosim can simulate both wired and wireless network with layered TCP/IP stack with model
based on noisy & noiseless channel with MAC protocol 802.11/CSMA/MACA/TSMA and
various network, transport & application layer protocols. Glomosim is written using PARSEC
language [16], a C derivative for large scale parallel simulation.
B. Energy Consumption Model
We are using IEEE 802.11b (DSSS modulation) as MAC protocol. The transceiver uses
energy both to transmit and to listen for incoming packet. It also consumes energy in idle state.
Let, the energy needed to transmit a packet Et for duration tt and to receive a packet Er for
duration tr .Also assume it waits for ti consuming energy Ei . Then total energy consumed by that
node will be approximately:
Ec =Et * tt + Er * tr + Ei * ti (6)
We assumed each node will use 5V DC battery with rated capacity of 1500 mAh.
Transmission energy consumed will depend on radio signal strength of transmission; here we
assumed 22.5 dBm; which approximately translate into 177.83 mW.
A =
V
W
(7)
From equation (7) we get, A= 35.57 mA for V=5V DC. If we draw the same amount of current,
using 1500 mAh battery, we’ll get approximately 42 hour of runtime before the battery dies.
Adding Idle and receiver power we’ll get less than that.
C. Simulation Results
We have made a comparative study between OLSR, SAODV and our scheme SMRR. We
carried out the result is based on the simulated data, the ACK being sent and frequency of data
transfer. First we evaluate number of admin in the network by both protocol variant as a
function of number of nodes. Maximum numbers of nodes were set to 50. Also to simulate
attack vector, we configured Glomosim in such a way that 20% of those nodes will randomly
drop packet or delay the delivery to next hop.
Simulation results are illustrated in following figures:

142
Figure 5: ADMIN Count
Here we can’t see much difference in average ADMIN count over basic OLSR protocol. We
can also see that number of ADMIN count has increased slightly when numbers of nodes were
20, 40 & 50. This increase in ADMIN count is due to shift in responsibility as the node’s
willingness & trust changes with time. Significant increase in ADMIN count adversely decreases
network performance. But here the count has increased only slightly. SAODV does not use
ADMIN concept. Increase in ADMIN count will affect radio layer packet collision as depicted
in above figure5.
Figure 6: Average Collision
We can see the collision in fact has increased, but only slightly as the increase of ADMIN
count was not so drastic. This increase was due to reselection of ADMIN and subsequent
topology message being broadcasted internally. It also increases due to sending and receiving of
acknowledgement packets. For SAODV, the increase in collision is due to frequent route
request-reply in each transmission. Collision increases with network density as more and more
nodes are trying to compete for radio frequency. Using 802.11b reduced collision due to
deliberate use of collision avoidance scheme (such as RTS/CTS) built into radio layer protocol
itself. Also we saw a slight change in throughput in the scheme. SAODV’s performance was
poor as compared to OLSR & SMRR as depicted in the above figure6.

143
Figure 7: Average Throughput
With 11 Mbps network bandwidth and multiple FTP and CBR data transfer, we saw average
throughput stayed around 27 kbps. Actually the average throughput increases in the case of
successful data transfer. Implementation of security helps us to avoid retransmission of packets
as well as data packet flooding. We also found that end-to-end delay also increased with our
proposed scheme compared to stock OLSR as depicted in the above figure 7.
Figure 8: Average End-To-End Delay
Compared to SMRR, SAODV has increased end-to-end delay; we suspect it is due to
transmission through suboptimal path. End-To-End delay increases for encrypting each message
though the transmission of every packet is secured. Then ACK transmission and encryption of
messages also increases the end to end delay considerably as depicted in the figure8.
VII. FUTURE WORK
We have already implemented trust factor in SMRR using signed acknowledgement which
has enhanced the security of the routing scheme. We have also been able to mitigate black hole,
gray hole, forged ACK, snooping attacks using this scheme. We have also implemented
parameterized willingness function in SMRR. Now our next future goal is to mitigate as many
routing attacks as possible by simulating each of those attacks individually.
VIII. CONCLUSION
Our secure SMRR which is inspired from OLSR may not be energy efficient but is quite
secure for end to end communication as compared to other routing scheme. In this paper
Administrator and trust based routing has been proposed. This novel feature allows us to forward
the data packets to the destination and by receiving the acknowledgement it verifies the validity

144
of the nodes in the route. The performance of this routing algorithm in comparison to OLSR has
improved. The security implementation has also protected the network from internal and
external threats.
REFERENCES
[1] Saoucene mahfoudh, pascale minet, "an energy efficient routing based on olsr in wireless adhoc and
sensor networks", 22nd international conference on advanced information networking and
applications, 2008
[2] Workshops, 2008.t. Clausen, p. Jacquet, "optimized link state routing protocol (olsr)", rfc 3626,
http://http://tools.ietf.org/html/rfc3626
[3] Fan hong; liang hong; cai fu; "secure olsr", 19th international conference on advanced information
networking and applications, 2005. Aina 2005. Page(s): 713 - 718 vol.1
[4] Xiaoqi li; lyu, m.r.; jiangchuan liu; "a trust model based routing protocol for secure ad hoc
networks",aerospace conference, 2004. Proceedings. 2004 ieee,volume: 2, page(s): 1286 - 1295
[5] C. Perkins; e. Belding-royer; s. Das, "ad hoc on-demand distance vector (aodv) routing." ietf. Rfc
3561, july 2003
[6] Songbai lu, longxuan li, kwok-yan lam, lingyan jia; "saodv: a manet routing protocol that can
withstand black hole attack",conference on computational intelligence and security, 2009. Cis '09,
page(s): 421 – 425
[7] Juwad, m.f.; al-raweshidy, h.s.; "experimental performance comparisons between saodv & aodv",
second asia international conference on modeling & simulation, 2008. Aicms 08, page(s): 247 - 252
[8] Papadimitratos p., haas z. J., samar p., “the secure routing protocol (srp) for ad hoc networks”,
draftsecure-routing-protocol-srp-00.txt, september 2002.
[9] J. Martin leo manickam, s.shanmugavel, "fuzzy based trusted ad hoc on-demand distance vector
routing protocol for manet", 15th international conference on advanced computing and
communications, 2007
[10] Y.c. hu, a. Perrig, and d.b. johnson, “ariadne: a secure on- demand routing protocol for ad hoc
networks,” proceedings on eighth annual int’l conf. Mobile computing and networking (mobicom),
2002, pp. 12-23.
[11] Yang ya-tao, yuan zheng, fang yong and zeng ping, "a novel authentication scheme based on trust-
value updated model in adhoc network", 31st annual international computer software and
applications conference(compsac 2007), 2007
[12] Raquel lacuesta gilaberte, lourdes peñalver herrero, "a secure routing protocol for ad hoc networks
based on trust", third international conference on networking and services (icns'07), 2007
[13] Jesus m. Gonzalez, mohd anwar, james b.d. joshi, "trust-based approaches to solve routing issues in
ad-hoc wireless networks: a survey", 2011 international joint conference of ieee, trustcom-11/ieee
cess-11/fcst-11
[14] Xiang zeng, rajive bagrodia, mario gerla,"glomosim: a library for parallel simulation of large-scale
wireless networks", workshop on parallel and distributed simulation, may 1998
[15] Niigata university, information & communication networks laboratory, “olsr_niigata”,
http://www2.net.ie.niigata-u.ac.jp/olsr-e.php.
[16] Rajive bagrodia, richard meyer, mineo takai, yu-an chen, xiang zeng, jay martin, ha yoon, "parsec: a
parallel simulation environment for complex systems", october 1998

A Novel Multipoint Relay based Secure Routing in MANET

More Related Content

What's hot (19)

Similar to A Novel Multipoint Relay based Secure Routing in MANET (20)

Recently uploaded (20)

A Novel Multipoint Relay based Secure Routing in MANET