A Trusted Routing Based Service Discovery Protocol with

Backup Nodes in MANETs

Min-Hua Shao, Yi-Ping Lee, Yen-Fen Hou and Cheng-Yi Ho

Department of Management Information Systems, National Pingtung University of Science

& Technology, 1 Hseuh Fu Road, Nei Pu, Pingtung 912, Taiwan, China

Abstract. The wireless MANET is particularly vulnerable on account of its in-

trinsic characteristics of open medium, dynamic topology, absence of central

authorities, distributed cooperation and constrained capability. These vulnera-

bilities create significant challenges for routing protocols operating in the entire

network. They have inspired lot of research interests regarding node connectivi-

ty in MANETs, but very few measures exist to trust-integrated cooperation for

service discovery. In this paper, we employ cross-layer approaches to propose a

new trusted routing based service discovery protocol called TRSDP. The

TRSDP is a kind of a reactive routing protocol, improved on the Dynamic

Source Routing. Moreover, a backup node mechanism for quick reconnection

during link failures is provided in TRSDP in order to cope with the dynamism

of such networks. Case studies involving security and service discovery scena-

rios were presented to demonstrate how the proposed protocol works. As a re-

sult, this paper gives a solution for the trusted and efficient cooperation of

routing and service discovery in MANETs.

1 Introduction

A mobile ad hoc network (MANET) [5] is considered a collection of wireless mobile

nodes that are capable of communicating with each other without relying on any

infrastructure or any centralized administration. In such a network, each mobile node

operates not only as a host but also a router, forwarding packets for other mobile

nodes in the network that may be outside wireless transmission range. Therefore,

wireless MANET is particularly vulnerable due to its intrinsic characteristics, includ-

ing of open medium, dynamic topology, absence of central authorities, distributed

cooperation and constrained capability [4]. Routing plays an important role in the

security of the entire network [1]. Most of the research on MANETs has focused on

issues dealing with the connectivity between mobile nodes mainly through routing

protocols to cope with the dynamism of such networks and the arising problems the-

reof. However, solving the problem of connectivity alone is not sufficient for the

adoption of MANETs. Because their basic role is to allow mobile users to exchange

data and user other’s services, there also is a need for service discovery [2].

In academic literature, service discovery mainly was addressed in the context of

wired networks and most of their solutions are centralized, where services are stored

Shao M., Lee Y., Hou Y. and Ho C..

A Trusted Routing Based Service Discovery Protocol with Backup Nodes in MANETs.

DOI: 10.5220/0003559901150122

In Proceedings of the 8th International Workshop on Security in Information Systems (WOSIS-2011), pages 115-122

ISBN: 978-989-8425-61-4

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

in a central server directory, and relied on powerful infrastructure nodes and reliable

links. By contrast, design of mechanisms proposed for service discovery in MANETs

that must cope with node mobility, unstable channel conditions and nodes with li-

mited energy and availability significantly differs from traditional wired solutions.

There exist three basic architectures that a service discovery approach can adopt, that

is, directory-based、directory-less and hybrid architectures. In order to benefit from

information available at the routing protocol, we couple cross-layer techniques with

directory-less architectures to propose a new trusted routing based service discovery

protocol called TRSDP. The proposed scheme is based on integrating the routing

process with the service discovery process and further, redundant transmissions of

service discovery packets at the application layer are avoided, and energy is saved.

Considering efficiency and quality of route service against the dynamic nature of

MANETs, we introduce reputation-based trust model and backup routes mechanism

into the popular and widely used Dynamic Source Routing (DSR) [8]. In which, the

communicating parties can select the most reliable route based on trust relationship

among nodes and then feedback the connection experience to maintain the relation-

ship. Therefore, TRSDP is a solution which looks after both sides of secure and

efficient.

The rest of this paper is structured as follows: In section 2, we present the pro-

posed scheme in detail. In section 3, evaluating the proposed scheme through experi-

ment and analyze the result. The conclusion is in section 4.

2 TRSDP Routing Protocol

In this section, we propose an efficient trust-based service discovery routing protocol

called as TRSDP routing protocol for MANETs. TRSDP is based on DBR

P [7]

routing protocol, we divided it into three phase: 1.Route discovery phase 2.Backup

node setup phase 3.Route maintenance phase.

2.1 Route Discovery Phase

At first, source node will set trust threshold (TT) and service of destination node

(DS). Source node will search node which can provide service and correspond to trust

threshold from route cache and trust cache, if intermediate nodes can match trust

value between source and destination node, then connect it; otherwise, source node

can’t search destination node cause by it cannot provide service and correspond to

trust threshold from route cache and trust cache, it will send RREQ packet. Algorithm

1 deals with the source node of enable route discovery, and algorithm 2 deals with the

examining of route cache.

116

Algorithm 1:

//DS: Service of destination node

//TT: Trust threshold

For each node in cache

If the route cache exist DS Then

Check timestamp of route cache;

Else

If the trust cache is expire Then

Update the trust cache;

End If

If the node is match TT in trust cache Then

Insert the node address in Target address field of

RREQ;

Forwarding the RREQ packet by unicast send to neighbor

nodes which must meet TT;

Else

The must meet TT of node cannot be found, failed to

connect;

End If

End For

Algorithm 2:

//TT: Trust threshold

If the timestamp of route cache is expire Then

Delete the record;

Else

If the destination node has match TT Then

If trust value of all intermediate nodes in the main path

>= TT Then

Connect it;

Else

If trust value of all intermediate nodes in the backup

path >= TT Then

Connect it;

Else

Insert the destination node address in target address

field of RREQ;

Search neighbor nodes from trust cache are matching

TT;

End If

Else

Search neighbor nodes from trust cache are matching TT;

End If

To avoid trust value deviation from record change of trust cache, we will the check

the record data. If record data expires, it will update data by searching trust value to

assure accuracy of trust value.

After intermediate node receives RREQ packet, the node checks itself if it is the

destination node at first. If not, next it checks it has received the same RREQ packet

by examining #Rid (Route request ID) in RREQ packet. Suppose the node receives

the packet at the first time; the node will record <#Rid,1> corresponded with time-

stamp of #Rid in RREQ sequence and number of receive both locating in route dis-

cover cache. It will search destination node which can provide service and correspond

117

to trust threshold from route cache and trust cache, if intermediate node can match

trust value between source and destination node, then it replies RREP packet to

source node. If it cannot search destination node which can provide service and cor-

respond to trust threshold from route cache and trust cache, then it chooses node

which can match trust threshold from trust cache, and forwards RREQ packet, de-

tailed procedure shown in algorithm 3.

Algorithm 3:

//RD: Route discovery cache

//DS: Service of destination node

//<#Rid,C>: C is a counter

//TT: Trust threshold

//index: The intermediate node’s IP

While the intermediate node doesn’t meet the conditions for

If the <#Rid,1> existing in RD Then

If the Address[i] of RREQ == index Then

If n>=3 || expire in RD Then

Drop packet;

Else

C=C+1; //<#Rid,C+1>

End If

Else

Drop packet;

End If

Else

Newly added <#Rid,1> into RD;

Update the timestamp of RD;

End If

For each node in route cache

If the route cache exist DS Then

Check timestamp of route cache;

Else

If the trust cache is expire Then

Update the trust cache;

Else

If node is match TT in trust cache Then

Insert the node address in Target address field of

RREQ;

Forwarding the RREQ packet by unicast send to neigh-

bor nodes which must meet TT;

End if

End For

End While

Reply RREP

If the timestamp does not expire or the total receive times of same RREQ packet

are less than two, intermediate node receives the same #Rid in RREQ packet again;

then the node will add one to the receive times as <#Rid,n+1>. If the timestamp ex-

pire or receive times of same RREQ packet are more than two, intermediate node will

drop the packet. If destination node receives RREQ packet, it will reply RREP packet

to source node, detailed procedure shown in algorithm 4.

118

Algorithm 4:

//RD: Route discovery cache

//<#Rid,C>: C is a counter

Set D = 0; // Initialization

If C!=0 Then

If the timer is expire in RD Then

Select backup nodes;

Else

C=C+1;

Save <#Rid,C> into Address[i];

End if

Else

<#Rid,1> into Address[i];

Start timer of RD;

If D==1 Then

Forward the RREP packet according to RREQ Address[i];

End If

Insert trust value in RREP packet;

Reply RREP packet by Address[i] of RREQ;

The intermediate node according to RREP updates route

cache and trust cache;

Update timestamp of trust cache;

D=1;

While(not source node)

Connect it;

End If

In order to avoid the course of RREQ delivery being a loop, the node will check if

its own IP address exists in address[i] in RREQ packet when intermediate node rece-

ives RREQ packet. If not, it will drop the RREQ packet.

2.2 Backup Node Setup Phase

When destination node receives RREQ packet at the first time, it will reply RREP

packet. Forward through intermediate nodes, along the reverse paths back to the

source node, this can create a two-way link. In our scheme, when destination node

receives RREQ packet, it will record information in the route table and start a timer,

then keep receiving all RREQ packets in certain time until the time finishes then reply

RREP packet. The amount of time can allow user to adjust follow according to net-

work environment. After timer expires, the destination node goes to build backup

path phase.

The destination node will pick out some backup nodes from address[i] which was

recorded at route discovery cache, and meanwhile sent backup set message to notify

each backup node. At the same time, the backup path will be saved in backup route

cache.

2.3 Route Maintenance Phase

When the information is transmitting, the node cannot receive the reply from

downstream node which means the path is invalid, and then enter route maintenance

119

phase. First of all, detecting invalid path node will check itself is a backup node or

not. If it is, it will directly replace the main route with a backup route from its own

backup route cache. If it is not, the node will send a route break packet to the

upstream node. Then similarly, the receive route break packet node checks if it is

destination node. If not, the route break packet be forward to upstream node

continually, until backup node receives. When the source node receives the packet,

finding out there is no substitutive path to destination node. The source node will

restart route discovery phase.

3 Analysis

We have implemented TRSDP routing protocol in Java environment. The essence of

a perfect information security can be classified into three parts: 1.Confidentiality

2.Integrity 3.Availability. In the paper, we focus on the data integrity and the path

availability.

1) Data integrity: Each node only recognizes part of the network topology in

MANET, it has to cooperate with other trusty nodes to gain correct route information

and deliver information completely. Based on many secure route protocol references,

link meeting the requirement of trust is connected through a relationship like interper-

sonal trust, excluding malicious nodes from the linking path to ensure data integrality.

2) Path availability: Because of each node has high mobility in MANET; network

topology will change quickly and be easy to disconnect. As soon as the link discon-

nects, original DSR route protocol will restart route discovery. However, it wastes

much time and influences the real-time of data transfer. The important issue in

MANET is how to build backup path and improve the path availability.

3.1 Security Scenario

In [6], scholars defined trust rate into three security levels. We refer to their research,

setting the trust threshold (TT) previously to divide the trust in three levels which are

ranges from zero to one. If the value approaches one is on behalf of the node has

higher reliability. On the other hand, low reliability signifies that the node does not

accumulate enough of trust value. To make trust degree clearer in representation, we

separate the range into three parts. We simulate the scenario by 23 nodes; an example

of the randomly generated topology is given in Fig. 1.

Fig. 1. The randomly generated topology.

120

3.1.1 Case 1: Low Security Level

Using TRSDP routing protocol at low security level, nodes can send RREQ packet to

establish paths when trust values are more than zero. Consequently, they build up 25

paths. Node S、2、5、12、17 and 18 are backup nodes.

3.1.2 Case 2: High Security Level

Using TRSDP routing protocol at high level, nodes can send RREQ packet to

establish paths if the trust values are more than 0.66. There are two paths connected.

Node S is a backup node.

3.1.3 Discussion

In the preceding sections, number of paths at high security level is less than that at

low security level under the same network topology. The reason is that users require

higher security, so the node will not send RREQ packet to neighbor nodes whose

thresholds are lower than the trust. Therefore, it will minimize the risk of that mali-

cious nodes exist in paths.

A low security level has 25 paths that include malicious node induced by low se-

curity threshold. Suppose trust value of malicious node is less than 0.5 [3]. In case 1,

trust value of node 14 to node 17 is 0.28, which is lower than 0.5. So node 17 is rec-

ognized as a malicious node in routes even though its trust threshold is qualified. The

useful paths passing through node 17 will break in all probability. Therefore, in our

scheme we combined backup path to avoid time waste because of new route discov-

ery. Suppose that node 14 does not receive ACK packet from node 17. It will go to

build backup path phase. Node 14 sends route break packet to the upstream node, as

shown in fig. 2.

Fig. 2. Route breaks and repair. Fig. 3. Backup routing table.

When node 12 receives the route break packet, if it is the backup node, it will

search backup path from backup path table. There are two substitutive paths be found,

node 12 will choose one of them as main path to deliver data (As the show in fig. 3.).

Thus, it avoids data transfer delay induced by restarting route discovery in original

DSR.

121

4 Conclusions

Wireless market today is overcrowded with large number of different devices and

mobile devices are inherently scarce in resources, necessitating the need to cooperate

among them for performing tasks that cannot be done alone. This cooperation is in

the form of services that are offered by other device in MANETs. In this paper, we

use the idea of providing routing-layer support for service discovery to propose a

trusted routing based service discovery protocol with backup nodes. The integrated

protocol, TRSDP, can take advantage of the ability of obtaining service information

along with routing information from the same message by piggybacking service in-

formation onto routing messages. This way, redundant transmissions of two message-

producing processes are lightened, and energy is saved. Specifically, the establish-

ment trust-aware routes with backup nodes can be conducted at the same time, that is,

only trustworthy intermediate nodes can participate in activities of data forwarding

and backup restoration. Therefore, the proposed scheme can provide a better tradeoff

between reliability and efficiency.

Acknowledgements

The author would like to thank the anonymous reviewers for their valuable comments

and suggestions. This paper was supported in part by the National Science Council,

Taiwan, under contract NSC 98-2410-H-020-007-MY2.

References

1. Ververidis, C. N., Polyzos, G. C.:Service Discovery for Mobile Ad Hoc Networks: A Sur-

vey of Issues and Techniques. IEEE Communications Surveys & Tutorials, 3rd Quqrter

(2008) 30-45

2. Deng, H., Li, W., Agrawal, D. P.:Routing security in wireless ad hoc networks. IEEE

Communications Magazine, Vol.40(10). (2003) 70-75

3. Poonam, K. G., Misra, M.:Trust Enhanced Secure Multi-Path DSR Routing. International

Journal of Computer Applications, Vol.2(2). (2010) 63-69

4. Shao, M. H., Huang, S. J.:Lightweight Anonymous Routing for reliability in Mobile

Ad-Hoc Networks. Journal of Research and Practice in Information Technology, Vol.41(2).

(2009) 171-180

5. Corson, S., Macker, J.:Mobile Ad Hoc Networking (MANET): Routing Protocol Per-

formance Issues and Evaluation Consideration. IETF WG Charter

6. Wang, B., Wei, R.:Zero-knowledge trust negotiation. In Proceedings of CSCWD 2009.

(2009) 390-395

7. Wang, Y. H., Chao, C. F.:Dynamic backup routes routing protocol for mobile ad hoc net-

works. Information Sciences, Vol.176 (2). (2006) 61-185

8. Johnson, D. B., Maltz, D. A.: Dynamic source routing in ad hoc wireless networks. Mobile

Computing. (1996)

122