ISSN 2278 – 3091 International Journal of Advanced Trends in Computer Science and Engineering (IJATCSE), Vol. 4 , No.4 Pages : 16 - 21 (2015) Special Issue of ICEEC 2015 - Held on August 24, 2015 in The Dunes, Cochin, India http://warse.org/IJATCSE/static/pdf/Issue/iceec2015sp04.pdf
Secure Routing in Mobile Ad-hoc Networks Using Evidence Theory Shyma M1, Nishanth N2 1
Student T.K.M College of Engineering, India, shymaprasad89@gmail.com 2 Professor T.K.M College of Engineering, India, nishtkm@gmail.com
Abstract: In Mobile Ad-hoc NETworks (MANETs) nodes communicate directly with each other when they are both within the same communication range. Otherwise, they rely on their neighbors to relay messages. The open medium and wide distribution of nodes or the lack of centralized infrastructure make MANET vulnerable to malicious attackers. Security has become a primary concern in order to provide protected communication between mobile nodes in a hostile environment. Hence providing secure route is the most challenging task to be carried out in MANET environment. This work proposes a unified trust management scheme for MANETs to provide secure routing. In this scheme every node calculates the trust value of its one hop neighbor by both direct observation and recommendations provided by other neighbors. The calculated trust values are then used for calculating the path trust of all possible paths between any source and destination node by AODV routing protocol which may calculate shortest path in the absence of trust incorporation. From these, shortest trusted path is selected for communication.
Key words: MANETs; Security; Trust; Trust management.
INTRODUCTION Mobile Ad hoc Networks (MANETs) [1] represent complex distributed systems that consist of wireless nodes that can dynamically and freely self-organize into arbitrary and temporary ad hoc network topology. This allows people and devices to seamlessly inter networked in areas where no pre-existing communication infrastructure exist. In MANET a primary requirement for the
establishment of communication among nodes is that nodes should cooperate with each other. In the presence of malicious nodes, this requirement may lead to serious security concern. Because mobile ad hoc networks have far more vulnerabilities than the traditional wired networks, security is much more difficult to maintain in MANET [2].The unique characteristics of MANETs such as dynamic topology and resource constrained devices pose a number of nontrivial challenges for efficient and secure routing protocols [3], [4]. Therefore, establishing and quantifying behavior of nodes in the form of trust is essential for ensuring proper operation of MANET. Most of the routing protocols developed for MANETs, such as DSR, AODV and DSDV are based on the multi hop assumption and they do not incorporate any security mechanism [5]. To increase MANET performance to enforce cooperation in the network reputation and trust based schemes have been developed. This scheme utilizes the past
behavior of end-users to enable a node to decide whether other nodes are cooperative and trustworthy [6]. In this paper, our scheme is a security mechanism that mainly protects AODV against two types of misbehavior, dropping packets and modifying packets. That is, this scheme is a trust based secure routing for MANET using evidence theory [7], [8]. As compared to the trusted AODV protocol this work will use both direct observation and recommendations for trust computation. The traditional AODV only finds minimum hop path, where as this scheme consider the number of hops from source to destination as well as the path trust from source to destination and finds highly trusted shortest path for communication. The latest work in this field also considers both direct and indirect observation for trust calculation [9]. This work is as an enhancement to the existing work, by eliminating the vulnerabilities of it. This trust management scheme provides some modifications in both direct and indirect trust calculations. The added advantage of this paper is that trust satisfaction factor and penalty factor are considered in the direct trust calculation to give penalty to misbehavior. In the calculation of Bayesian trust value a penalty factor is considered to give more weights on misbehavior, and to reduce the trust value of a node when it misbehaves. With indirect observation from neighbor nodes of the observer node, the trust value is derived using either Dempster-Shafer theory (DST) or Murphy’s rule of combination [10], [11]. The advantage of this scheme is that, it will eliminate the conflict of using DST. The DST rule becomes inaccurate when the conflict becomes high. For such situation Murphy’s rule is used as an alternative rule for combining the evidences from various observers. Then over all trust value between observer and observed node is calculated as the weighted sum of trust value obtained from direct observation and trust value obtained from recommendations. Then these calculated trust values are incorporated with the AODV routing protocol to provide secure routing. The path trust is calculated as the sum of trust values between all pair of nodes between source and destination. Then the secure routing protocol will select highly trusted shortest path. RELATED WORKS In MANETs, an untrustworthy node can create considerable damage and adversely affect the quality and reliability of data. Computing the trust level of a node has a positive influence on the confidence with which an entity conducts transaction with that node. Trust based security
ISSN 2278 – 3091 International Journal of Advanced Trends in Computer Science and Engineering (IJATCSE), Vol. 4 , No.4 Pages : 16 - 21 (2015) Special Issue of ICEEC 2015 - Held on August 24, 2015 in The Dunes, Cochin, India http://warse.org/IJATCSE/static/pdf/Issue/iceec2015sp04.pdf schemes are studied recently in [12], [13]. Trust Trust is interpreted as the degree of belief that a node in the computations consist of three components: ‘experience’, network will carry out a task that it should. Trust can also be ‘recommendation’, ‘knowledge’. The ‘experience’ defined as the expectation of a subjective probability that a component of trust for each node is directly measured by their immediate neighbors and kept updated at regular intervals in the trust table. The existing trust table is propagated to all other nodes as ‘recommendation’ part of trust. At a regular interval, the previously evaluated trust is included in the current ‘knowledge’ component of total trust. An Ad-hoc on-demand trusted path distance vector (AOTDV) is proposed for MANETs [14]. It is a trust based multi path routing using AODV protocol. AOTDV adopts a hop-by-hop routing mechanism in which the source is not expected to know which neighbor is the next hop. In this scheme a source establishes multiple trustworthy paths as candidates to a destination in single route discovery. The main problem of this method is that it considers only direct observation for trust computation. Hence the chance of Fig: 1 Mobile ad hoc network model detecting attacker node that acts genuine to some nodes and malicious to some other nodes will be very less. Hence this truster uses to decide whether or not a trustee is reliable. scheme is not suitable for MANETs having Gray hole Based on the definition and properties of trust in MANETs, [selective black hole] attackers. the proposed scheme evaluates trust by a real number T, with The authors of [15] use Bayesian inference to evaluate the a continuous value between 0 and 1. In this scheme initially direct trust and Dempster-Shafer theory (DST) to evaluate every node calculates the direct trust of each of its one hop indirect trust. Dempster’s rule for combination is a procedure neighbors periodically and keeps updating its routing table. for combining independent pieces of evidence. The major This is done by observing the packet forwarding behavior of drawback of this method is that when the conflict between the them. Then they calculate the indirect observation trust from observers is high the DST rule of combination becomes inaccurate. That is at high conflict conditions, using DST as the recommendations provided by other one hop neighbors trust combination rule will gives false alarm. So the [16]. Combining the trust value, from direct observation and the recommendation trust value using DST is incorrect or trust value from recommendations, we can get a more inaccurate. realistic and accurate trust value of a node in MANETs. Then TRUST COMPUTATION the overall trust value can be calculated as the weighted sum of direct trust and recommendation trust as A. Network Model In the network model a number of nodes are placed = ∗ + ∗ (1) randomly in the simulation area as shown in Fig: 1. There are two types of nodes in the network, normal nodes which follow the routing rules and compromised node which drop or modify the packets maliciously. The number of malicious nodes is minor compared to the total number of nodes in the network. In the network one node is set as source node and another one is set as destination. It is required to calculate a trusted minimum hop path from source to destination node. The malicious node in the path from source to destination will claim that it is having a shortest route to the destination. So the source node will always select the route through malicious node for its communication with the destination. Hence the data packets send by the source node will never reaches the destination. For secure communication malicious nodes have to be detected eliminated from the communication path. And another shortest path that does not contain malicious node as a router has to be selected for communication. For this a trust based secure routing scheme is proposed. B. Trust model
Where W1 and W2 are the weight assigned to direct observation trust and recommendation trust respectively, W1+W2=1. And TD is the trust value obtained from direct observation, TR is the trust value obtained from recommendation. Then the trust value between pair of nodes is used to calculate the path trust. In this network model there are several possible paths between source and destination node with different hops. The best routing algorithm will select a route that is having higher path trust value and minimum hop count. The routing protocol will first calculate path trust value for all possible paths and then select a highly trusted minimum hop path from these available paths.
C. Trust Computation of One-hop Neighbors In the direct observation, it is assumed that each observer can overhear packets forwarded by an observed node and compare them with original packets so that the observer can identify the malicious behaviors of the observed node. In this proposed scheme direct observation trust is computed as the
ISSN 2278 – 3091 International Journal of Advanced Trends in Computer Science and Engineering (IJATCSE), Vol. 4 , No.4 Pages : 16 - 21 (2015) Special Issue of ICEEC 2015 - Held on August 24, 2015 in The Dunes, Cochin, India http://warse.org/IJATCSE/static/pdf/Issue/iceec2015sp04.pdf product of two components (in Algorithm 1), the first component is trust satisfaction factor [18]. At the beginning component of direct observation trust is calculated by when there is no observation history available, then the trust observer node using Bayesian inference [17]. This value of a node is taken as 0.5. That means the node is component is termed Bayesian trust. And the second seemed as neutral when no history records behaviors is established. The value trust can be revised continuously through follow-up observation. Then trust from Bayesian Algorithm 1 Trust Calculation with Direct Observation inference is taken as the expectation of beta distribution along 1: if node i, which is an observer, finds a one hop neighbor, with a penalty factor to give more weight on misbehavior. then Incorporating penalty factor can help the proposed scheme 2: set variables, total packets generated, no. of packets distinguish the malicious node quickly and avoid them forwarded 3: if node i, finds that its 1 hop neighbor, receives a packet, disrupting the normal traffic between benign nodes again because of two reasons. Firstly, this can lower the trust of an then 4: the total packets generated increases one attacker when it misbehaves. Secondly, the trust of the attack 5: if node i, finds that its 1 hop neighbor, forward the will not recover quickly even if it forwards a large number of packet successfully, then packets correctly due to the impact of the penalty factor. The 6: the no. of packets forwarded increases one penalty factor is inspired by our daily lives in human society, 7: end if where a scandal can badly affect a person who has a good 8: end if reputation. What’s more, it is hard to quickly recover a good 9: end if reputation. The factor of punishment makes the trust 10: Calculates the Bayesian trust BT, from (2) 11: Calculates Trust satisfaction factor TS, from (3) evaluation more realistic. Algorithm 1 will describes the trust 12: Calculates the Direct observation trust TD, from (4) computation with direct observation. For any source node set variable for the number packets generated by it and for the Algorithm 2 Trust Calculation with Recommendation number of packets forwarded correctly by each of its 1: if node i, which is an observer, finds no one hop neighbor, neighbors. then Yn-1- Number of packets generated by a node 2: set recommended trust to zero Xn-1-Number of packets forwarded correctly by its 3: else neighbor 4: if node i, finds only a single neighbor, then 5: set recommended trust as the direct trust of 1 hop Zn-1- Number of failed packets neighbor Zn-1=Yn-1- Xn-1 6: else 7: if node i, finds more than one neighbor, then 8: Calculates conflict factor C, compare with Let, 0=0=1 threshold, then 9: Calculate recommended trust TR, from (5) n=n-1+Xn-1 10: end if 11: end if n =n-1+Yn-1-Xn-1 12: end if Then the Bayesian trust can be calculated as
Table I: Simulation Parameters Parameter Application protocol CBR transmission time CBR transmission interval Packet size Transport protocol Network protocol Routing protocol MAC protocol Physical protocol Data rate Transmission power Radio range Propagation path loss model Simulation area Number of nodes Simulation time
=
Value CBR 1s to 500s 0.5s 512 bytes UDP IPv4 AODV IEEE 802.11 IEEE 802.11b 2Mbps 6dBm 180m Two-ray 500Χ500, 1000Χ1000 10,20,30,40,50,60 600s
(2)
Where =
2
>
−
< (
)
4 1
>
−
≥ (
)
he trust satisfaction factor =
(3)
Then the direct observation trust of Node B by Node A is calculated as the product of Bayesian trust and trust satisfaction factor. The trust satisfaction factor will lower the trust value when the failure rate is high, but it doesn’t lower 18
ISSN 2278 – 3091
PDR
Throughput
International Journal of Advanced Trends in Computer Science and Engineering (IJATCSE), Vol. 4 , No.4 Pages : 16 - 21 (2015) Special Issue of ICEEC 2015 - Held on August 24, 2015 in The Dunes, Cochin, India http://warse.org/IJATCSE/static/pdf/Issue/iceec2015sp04.pdf the trust even though the node will forward a large number of 50 packets correctly. Then the direct observation trust TD is 40 30 = ∗ (4) 20 For any source node set variable for the number packets 10 generated by it and for the number of packets forwarded 0 correctly by each of its neighbors. 0 1 2 3 4 5 1 Number of attackers 0.8 0.6
Proposed scheme
Existing scheme
0.4 0.2 0
Fig 5: Throughput v/s Number of Attacker
20
30
40 50 60 Number of nodes Proposed scheme Existing scheme
Fig 2: PDR v/s Number of node
Routing load
10
5 4 3 2 1 10
30
40
50
60
Proposed scheme
0
1
Fig 6: Routing load v/s Number of nodes
2 3 4 5 Number of attackers
Proposed scheme
Existing scheme
Fig 3: PDR v/s Number of Attackers
Troughput
20
Number of nodes
End to end delay
PDR
0 1 0.8 0.6 0.4 0.2 0
35 30 25 20 15 10 5 0 10
49 47 45 43 41 39 37 35
20
30
40 50 60 Number of nodes
Proposed scheme Existing scheme 10
20
30
40 50 60 Number of nodes
Proposed scheme Existing scheme
Fig 4: Throughput v/s Number of nodes
Fig 7: End to end delay v/s Number of nodes
A. Indirect Trust Computation The recommendations provided by neighbor nodes are used to evaluate the trust value of the observed node. That is every node is calculating the indirect observation trust of its 1 hop neighbors from the recommendations provided by other 1 hop neighbors (in Algorithm 2). If there is no one hop neighbor to provide recommendations then the indirect
ISSN 2278 – 3091 International Journal of Advanced Trends in Computer Science and Engineering (IJATCSE), Vol. 4 , No.4 Pages : 16 - 21 (2015) Special Issue of ICEEC 2015 - Held on August 24, 2015 in The Dunes, Cochin, India http://warse.org/IJATCSE/static/pdf/Issue/iceec2015sp04.pdf observation trust is taken as zero. And if there is only single performance of the proposed scheme with that of AODV neighbor to provide recommendations then indirect without security mechanism. observation trust is taken as the direct observation trust of A. Simulation Environment Settings recommender by the observer. If there is more than one recommendation provided by one hop neighbors then some We randomly placed nodes in the defined area. combining methods are used to find out the overall trust Simulations are performed in different scenario; with each value. For combining these recommendations either DST or scenario has a pair of nodes as the source and destination. The Murphy rule of combination is used, based on the value of traffic used for simulation is constant bit rate traffic (CBR). conflict factor C [19], [20]. In the indirect trust calculation first conflict factor is The simulation parameters are listed in Table: I. In the calculated. Based on the value of conflict either DST or simulation it is assumed that there are two types of nodes in Murphy rule is selected for indirect trust calculation. DST is the network, normal node and malicious node. Normal nodes the best method for trust computation at low conflict values. are nodes that follow routing rules, where as malicious nodes When the conflict becomes high then conflicts between will drop or modify packets maliciously. As compared to the different pieces of evidence are mismanaged by DST. The application of DST leads to an undefined condition and total number of packets the number of malicious nodes is very less. In this adversary mode, proposed scheme is cannot be applied for trust computation. Hence under high conflict condition Murphy rule is applied. evaluated and compared with the original AODV protocol. The conflict factor C is the mass allocated to the empty set We have simulated the networks with different numbers of and is calculated as, nodes. Fig 2 is an example of the network set up, where node 3 is the source node, node 13 is the destination node and = 1 − ( ) = 1 − [ ( ) ( ) + ( ) ( )] (5) nodes 6 and 9 are malicious nodes. For node mobility, the R random waypoint mobility model is adopted in 60 node Then the indirect observation trust T is calculated as, MANET. The maximum velocity of each node is 0 to 20 m/s. . Four performance metrics are considered for understanding = (6) . the performance variation of MANET with and without malicious nodes. 1) Packet delivery ratio (PDR) is the ratio of Where, the number of data packets received by a destination node to = ( )⊕ ( ), = [ ( )+ ( )] (7) the number of data packets generated by the source node. 2) Throughput is the total size of data packets correctly received is the trust value obtained from Dempster’s rule and the trust by a destination node every second. 3) Routing load is the ratio of number of control packets transmitted by nodes to the value calculated by Murphy’s rule respectively. number of data packets received successfully by destination E. Trust Based Secure Routing during the simulation. 4) Average end to end delay. It is the Compared to the existing AODV scheme that uses the mean of end to end delay between a source node and a shortest path based on hop count, thrust based routing scheme derive the best routing path considering both trust values and destination node with CBR traffic. hop count. The Dijkstra’ algorithm is used to calculate the B. Performance Improvement best routing path. Since minimization is used in the Dijkstra’ The original AODV and our scheme are evaluated in the algorithm (e.g., to find the shortest path with the minimal hop count in traditional AODV), it is need to convert the trust simulation, where some nodes act maliciously by dropping or value to untrustworthy value. Then, we can minimize the modifying packets. In Fig: 2 we compare PDR for AODV untrustworthy value of a path using the Dijkstra’ algorithm. MANET with and without trust scheme, which includes To this end, define the untrustworthy value between node 1 nodes from 10 to 60. From the figure, we can see that the and node 2 as U12, which can be calculated as U12=1-T12. The AODV MANET with trust has higher PDR as compared to sum of untrustworthy values of a path is original AODV. This is because the original AODV protocol =∑ = ∑ (1 − ) (5) does not have any security measurements, and the chance of dropping packets by malicious nodes is high. Hence the PDR Where Tki ki+1is the trust value between node ki and its one is very low in the case of original AODV protocol. Where as hop neighbor, node ki+1. Nodes k1, k2, . . . , kn belongs to the in the proposed scheme the trust scheme will detect malicious path with n − 1 hops. The best routing path satisfies the nodes and hence the chance of reaching packets at the minimum of Upath. destination is high. So the Packet Delivery Ratio is also very high. We can also find that the PDR of both the schemes SIMULATION RESULTS AND PERFOMANCE decreases gradually when the number of nodes grows. For IMPROVEMENT The proposed scheme is simulated in NS2 simulator with small number of nodes the PDR high. As the number of nodes AODV routing protocol. The effectiveness of the scheme is increases the packet drop increases. This is because the evaluated in malicious environment. We compare the collision of sending messages becomes more frequent as the 20
ISSN 2278 – 3091 International Journal of Advanced Trends in Computer Science and Engineering (IJATCSE), Vol. 4 , No.4 Pages : 16 - 21 (2015) Special Issue of ICEEC 2015 - Held on August 24, 2015 in The Dunes, Cochin, India http://warse.org/IJATCSE/static/pdf/Issue/iceec2015sp04.pdf number of nodes increases in the MANET. The PDR ACKNOWLEDGEMENT decreases even more as the number of attacker increases as in The opinions, findings, and conclusions or Fig: 3, this is because the black hole attacker will drop the recommendations expressed in this publication are those of packets without forwarding it to the receiver. Hence as the the authors and do not necessarily reflect the views of the number of attacker increases number of dropped packets also Department of Justice. The authors are thankful to Dr. S. Suresh Babu and the increases drastically, that will reduce the PDR. staff at Networking Laboratory, were the simulation works As compared to the existing scheme throughput of the are performed. proposed scheme is very high. This is because the security mechanism in the proposed scheme will increase the number REFERENCES of correctly received packets. It is observed from Fig: 4 that [1] J. Loo, J. Lloret, and J. H. Ortiz, Mobile Ad Hoc Networks: Current the throughput also decreases with number of nodes; this is Status and Future Trends. Boca Raton, FL, USA: CRC, 2011. because the number of packets received correctly decreases [2] F. R. Yu, H. Tang, S. Bu, and D. Zheng, “Security and Quality of Service (QoS) co design in cooperative mobile ad hoc networks,” as long as the number of nodes increases. Fig: 5 reveal that EURASIP J. Wireless Commun. Netw., vol. 2013, pp. 188–190, Jul. 2013. the number of attackers has significant impact on the [3] S. Marti, T. Giuli, K. Lai, and M. Maker, “Mitigating routing throughput of the network. As the number of attackers misbehavior in mobile ad hoc networks,” in Proc. ACM MobiCom, Aug. 2000, pp. 255–265. increases the throughput also decreases to a very low value. [4] W. Lou, W. Liu, Y. Zhang, and Y. Fang, “SPREAD: Improving In Fig: 6, the result demonstrates that proposed scheme has a network security by multipath routing in mobile ad hoc networks,” lower routing load. This is due to the fact that the security ACM Wireless Netw., vol. 15, no. 3, pp. 279–294, Apr. 2009. mechanism imposed by the proposed scheme will increase [5] Asad Amir Pirzada, Chris McDonald, and Amitava Datta, “Performance Comparison of Trust-Based Reactive Routing the number of packets correctly received by the destination. Protocols,” IEEE Trans. Mob. Computing., VOL. 5, NO. 6, JUNE 2006 The cost of adding security mechanism is increase in [6] C. Perkins, E. Belding-Royer, and S. Das, Ad hoc On-Demand Distance Vector (AODV) routing, Jul. 2003, IETF RFC 3561. average end to end delay as compared to original AODV [7] S. Bu, F. R. Yu, P. Liu, P. Manson, and H. Tang, “Distributed combined authentication and intrusion detection with data fusion in protocol. Fig: 7 show that the proposed scheme has a slightly high-security mobile ad hoc networks,” IEEE Trans. Veh. Technol., higher average end-to-end delay than the existing scheme. vol. 60, no. 3, pp. 1025–1036, Mar. 2011. This is because the trust computation and update time are [8] Lef`evre E., Colot O., Vannoorenberghe P., Belief functions combination and conflict management, Information Fusion Journal, added along with route discovery time of original AODV Elsevier Publisher, Vol. 3, No. 2, pp. 149-162, 2002. protocol. Also the trusted path is always a longer one as [9] Zhexiong Wei, Helen Tang, F. Richard Yu, Maoyu Wang, and Peter Mason, “Security Enhancements for Mobile Ad Hoc Networks with compared to the path provided by the original routing Trust Management Using Uncertain Reasoning,” IEEE Trans. Veh. protocol. There is a trivial delay introduced by this scheme as Tech., 2014. compared to the existing scheme, but high security is [10] T. M. Chen and V. Venkataramanan, “Dempster–Shafer theory for intrusion detection in ad hoc networks,” IEEE Internet Comput., vol. 9, guaranteed. no. 6, pp. 35–41, Nov./Dec. 2005. [11] H. Yu, Z. Shen, C. Miao, C. Leung, and D. Niyato, “A survey on trust and reputation management systems in wireless communications,” Proc. IEEE, vol. 98, no. 10, pp. 1755–1772, Oct. 2010. [12] C. Zouridaki, B. L. Mark, M. Hejmo, and R. K. Thomas, “A quantitative trust establishment framework for reliable data packet delivery in MANETs,” in Proc. 3rd ACM Workshop SASN, Nov. 2005, pp. 1–10. [13] H. Wu, M. Siegel, R. Stiefelhagen, and J. Yang, “Sensor fusion using Dempster–Shafer theory,” in Proc. IEEE Instrumen. Meas. Technol. Conf., May 2002, pp. 7–12. [14] N. Marchang, R. Datta, “Light-weight trust-based routing protocol for mobile ad hoc networks,” IET Inf. Sec., Vol. 6, lss. 2,pp. 77-83 May 2011. [15] T. M. Chen and V. Venkataramanan, “Dempster–Shafer theory for intrusion detection in ad hoc networks,” IEEE Internet Comput., vol. 9, no. 6, pp. 35–41, Nov./Dec. 2005. [16] H. Yu, Z. Shen, C. Miao, C. Leung, and D. Niyato, “A survey on trust and reputation management systems in wireless communications,” Proc. IEEE, vol. 98, no. 10, pp. 1755–1772, Oct. 2010. [17] Audun Josang, Roslan Ismail, “The Beta reputation system,” 15th Bled Electronic Commerce Conf., 2002. [18] K. Rostamzadeh, H. Nicanfar, N. Torabi, S. Gopalakrishnan, V. C. M. Leung, “A context-aware trust-based information dissemination framework for Vehicular networks,” Internet of Things Journal, IEEE vol.2, pp. 121-132, 2013. [19] Voorbraak F., On the justification of Dempster’s rule of combination, Artificial Intelligence, 48, pp. 171-197, 1991. [20] Murphy C.K., Combining belief functions when evidence conflicts, Decision Support Systems, Elsevier Publisher, vol. 29, pp. 1-9, 2000.
CONCLUSIONS Owing to multi hop routing and absence of centralized administration in open environment MANETs are vulnerable to various security attacks. Hence providing secure route is the most challenging task to be carried out in MANET environment. This paper proposes a trust management scheme for MANETs to provide secure routing. In this scheme every node calculates the trust value of its one hop neighbor by both direct observation and recommendations provided by other neighbors that is indirect trust. We use packet forwarding ratio to evaluate the one hop neighbor trust. The calculated trust values are then used for calculating the path trust of all possible paths between any source and destination node by the routing scheme. AODV routing protocol may calculate shortest path in the absence of trust incorporation. By incorporating trust the AODV protocol will select shortest trusted path for communication. Dijkstra’s algorithm normally used for calculating shortest path between any pair of nodes can be used for trusted path computation.
21