Printer Friendly

Enhanced triple umpiring system for security and performance improvement in wireless MANETS.

1. Introduction

A wireless mobile ad hoc network is a self-created, self-organized and self-administering set of nodes connected via wireless links without the aid of any fixed infrastructure or administrator. Each node moves and operates in a distributed peer-to-peer mode, generating independent data and acting as a router to provide multi-hop communication. Wireless mobile ad hoc network is ideally suited for potential applications in civil and military environments, such as responses to hurricane, earthquake, tsunami, terrorism and battlefield conditions. Ensuring adequate security is an important aspect in such applications.

In this paper we tackle the problem of securing the network layer operations such as routing the control messages and data packet forwarding from malicious nodes. Malicious nodes may disrupt routing algorithms by transmitting a false hop count; by dropping data packets and by routing the packets through unintended routes and so on. In order to handle network operations successfully in the presence of malicious nodes we need a security system with the following three functionalities: (i) detect (ii) quarantine the malicious nodes and (iii) salvage. In detection the misbehaving node is traced and identified. Quarantine procedure envisages marking the offending nodes so that they do not participate any further in the network activities.

[FIGURE 1 OMITTED]

A proper security system should not only identify the disruption of the network because of malicious nodes, but should also be able to salvage the current communication path, so that despite disruption there is a successful termination of the communication event. Watchdog and SCAN systems [1, 2], the self umpiring system (Self_USS) [32-33] proposed earlier and TUS [36-37] being proposed in this paper find solutions to the first two issues. The ETUS, Enhanced Triple Umpiring System being proposed now, handles all the three functionalities successfully. Let us focus our attention on each one of these functionalities.

Detection mechanism is based on promiscuous hearing. Promiscuous hearing means listening to communications that is not meant for oneself. This is made possible by the wireless nature of the medium. Thus in Figure 1 when node B sends packets to C, A along with B's neighbours observes the event. The behavioural deviation, if any, on the part of B, in the retransmission of the messages it received from A, can be readily observed by all of the listening nodes. The question arises: "To which of these nodes the role of observation and detection is to be given?" In watchdog the immediate predecessor node--in this case A--is the designated node to watch over B's actions. In SCAN, a minimal set a 'k' neighbors of B, collectively decide whether 'B' is misbehaving or not. In Self-USS [37] during data forwarding B's behaviour will be supervised by A, while during route reply process node C will be the supervisor. In TUS, three umpires decide the fate of B collectively. Two of the umpires are neighbors specifically commissioned for the role; the third umpire will be immediate predecessor node A, in its umpiring role. Again the decision is taken collectively by all the three umpires involved. Having three umpires, instead of a single umpire that we had used in Self-USS reduces false accusation probability. Further instead of getting feedbacks from the group of 'k' neighbors, as SCAN has done, the present proposal of specifically commissioning nodes to play the role of umpires has an immediate benefit in salvaging operations. This aspect is brought out latter in this section.

Having identified the offending node, the next step is to quarantine the node, so that it does not take part any more in the network activities. The watchdog uses its Pathrater mechanism for the same. The Pathrater punishes the offending node by giving it a large negative rating, while all other nodes, which have played a collaborating role in successful completion of the message transfer, have their rating incremented periodically. Once a node gets a high negative rating, it is not likely to be included in any new communication link. In SCAN, systems of tokens have been suggested. The tokens are renewed periodically by the system of 'k' neighbors. When an offending node's token is revoked, it cannot further participate in the network. In Self-USS and the proposed TUS we achieve the avoidance of malicious nodes by a system of tokens which is similar to the one used in SCAN. Token is a pass or validity certificate enabling a node to participate in the network. It contains two fields: nodeID and status bit; nodeID is considered to be immutable. Initially the status bit of all participating nodes is set as 0 indicating "green flag" with freedom to participate in all network operations. It is assumed that a node cannot change its own status bit. The protocol ensures that the status bit can be changed only by the designated set of umpires acting collectively.

Our objective is designing the security system is to keep the overhead as minimum as possible while optimizing the output. We do not use encryption or key algorithms as done by SCAN. Further issuing tokens and token renewals create very large communication overheads and also degrade energy performance. There is no token renewal feature in the proposed system. In our system, all the nodes are pre-issued with green tokens. They continue to enjoy the status, until the system of umpires finds any node misbehaving and sets its status bit to 'red'.

In the real world civilian situations, we find a couple of policemen armed with no more than a whistle and a stick is able to control substantial crowds. Gun trodden heavily jacketed commandos are required only for special situations. Economy and operational constraints dictate that such a security cannot be provided everywhere. Ninety percent of our applications correspond to the former category. Our focus is on simple, civilian situations, where we would like to achieve high throughput even in the presence of malicious activities.

Now we move on to the third functionality, that of salvaging the route after any disruption. We will consider two distinct situations: disruptions during (i) route reply and (ii) data forwarding phases. In the Enhanced Triple Umpiring System (ETUS) we have provided a mechanism to handle each of these two situations. Let us now consider disruption during route reply phase.

The loss of route reply packets causes serious impairment of performance of routing protocol. This is because route reply packets are obtained after flooding the entire network with RREQs. Mekesh Singhal et al [34] [35]have proposed and implemented the idea of salvaging route reply (SRR) for on demand routing protocols. The basic idea is illustrated in Figure. 2. Assume that, initially there exists no active path from source node S to destination node D. Node S is discovering a route to node D. Node D sends a RREP to node S, through intermediate nodes X, C, B and A. Node C cannot send the RREP to node B because B has moved away.

Node C becomes the salvor, it saves the RREP message, and then it broadcasts a [RREQ.sub.SRR]. Node V receives the [RREQ.sub.SRR] and finds a route to the source node S in its routing table, so V sends a RREPSRR to C. C receives the RREPSRR and successfully salvages the original RREP by sending it along the path discovered by new path. It can use the new alternative route to send RREP packets to node S, through intermediate nodes A, Y, T and C. Then the return path after SRR is D-X-C-T-Y-A-S.

[FIGURE 2 OMITTED]

The proposed ETUS enhances the performance of triple umpiring system (TUS) with the incorporation of SRR. With this if the nodes behave maliciously during route reply phase, say, by giving a wrong hop count, such nodes will be flagged off from the network by the umpire and salvaging route reply packet commences immediately. Salvaging also takes place when there is a cut in the communication path to the recipient node [Table 1].

We move on to the second aspect of salvaging the network after disruption in the data packet forwarding operation. The other systems proposed earlier do not have this functionality. In ETUS, once the guilty node is flagged off with a red flag, the remainder part of the message transfer is completed by umpires switching their roles. When a guilty node is identified and flagged off the communication link is cut. The corresponding neighbouring umpires switch their roles; throw off their umpiring coats and give a helping hand in continuation of the message transmission. Does that mean that umpiring is totally discarded at this stage? The answer is 'no'. Triple umpiring system is on, in all the segments except the affected segment. In the affected segment, self umpiring system (Self_USS) proposed by us [37], which is similar in concept to watchdog becomes operational. With ETUS, there is one more benefit. Even if there were no malicious nodes, the performance will be better than normal AODV, because any disruptions of communications due to mobility are readily handled by our umpires!. The rest of the paper is organized as follows: section 2 provides details about network model and assumptions; Section 3 discusses TUS / ETUS models. Section 4 presents simulation results; Section 5 discusses analysis of results; Section 6 discusses the related work and Section 7 gives the conclusions.

2. Network Model and Assumptions

In this section, we formulate the wireless mobile ad hoc network and security models.

2.1 Network Model

We consider a wireless mobile ad hoc network where there is no restriction on the number of networking nodes. For differentiation purpose, we require each node to have a unique non zero ID. Further assumptions made in the design of TUS and ETUS systems are as follows:

1. A wireless mobile ad hoc network where nodes are free to move about or remain at stand still, at their will is assumed. Each node may join to the network, or node may leave from the network at any time.

2. The source and the destination nodes are not malicious.

3. Nodes may fail at any time.

4. Every node in the network has a list of neighbors.

5. There exists a bi-directional communication link between any pair of nodes, which is a requirement for most wireless MAC layer protocols including IEEE 802.11 for reliable transmission.

6. Wireless interfaces support promiscuous mode of operation. Most of the existing IEEE 802.11 based wireless cards support such promiscuous mode of operations, to improve routing protocol performance. Such promiscuous operations are already in vogue in standard protocols like DSR.

2.2 Security Model

The unique characteristics of wireless mobile ad hoc networks make them more vulnerable to the security attacks compared to wired networks or infrastructure based wireless networks. We consider the attacks on network layer operations. The two most important operations of network layer are routing message and packet forwarding operations. Attacks on either of them can lead to the failure of the network. In this model, the malicious nodes can readily participate on both categories of attacks: misbehaviors during routing and data packet forwarding operations. In our model, the routing misbehavior is restricted to route reply phase only. While forwarding RREP packets, a malicious node may give a wrong hop count or a wrong sequence number. In the data forwarding phase, any malicious node can drop packets with out forwarding them to intended destination.

3 TUS and ETUS Models

3.1 Models description

We present the configuration of nodes for TUS/ETUS models in Figure. 3. There are m + 2 nodes in the active data forwarding path, including source and destination. There are (m + 1) umpiring nodes. For any node [N.sub.i], the immediately preceding node [N.sub.i-1] and [U.sub.i] and [U.sub.i+1] will constitute the umpires during data forwarding. They collectively decide whether Ni is misbehaving or not.

The most important thing to ensure in order that the system works properly, is that there is a communication link between [N.sub.i-1], and [U.sub.i] , and [U.sub.i] and [N.sub.i] , and [U.sub.i+1] and [N.sub.i] and finally between [U.sub.i] ,and [U.sub.i+1]. We ensure this by incorporating required features while appointing umpires, during route reply phase.

[FIGURE 3 OMITTED]

Let us assume that RREQ information has reached destination and destination is to initiate RREP phase. For this along with regular RREP information, destination sends its list of neighbors to [N.sub.m]. [N.sub.m] finds an intersection of its own neighbor list and the list sent by destination. From among the common nodes, it arbitrarily selects one as umpire and unicast a packet to it. The selected umpire [U.sub.m+1] in turn, gives an acknowledgement, and enclose its list of neighbors to [N.sub.m]. [N.sub.m] again finds an intersection of its own neighbors and neighbor list sent by umpire [U.sub.m+1], and forwards the same to [N.sub.m]-1. [N.sub.m]-1 finds an intersection of its own neighbors and the list sent by Nm and randomly appoints an umpire [U.sub.m]. This process goes on until source is reached. Thus there will be [U.sub.m+1] umpires. This procedure ensures that there exists a communication link between nodes as stipulated earlier. During RREP phase, the self umpiring system is operational and during data forwarding phase three umpires effectively monitor the behavior of the designated node. The previous procedure ensured that there exists a communication link as stipulated. We have to further ensure that these links are maintained throughout the data transmission phase. This we ensure by hearing "hello" message packets form the concerned entities. Thus [N.sub.i-1] monitors the presence of [U.sub.i] and vice versa.

3.2 Analysis of operations

3.2.1 Route Reply phase

During route reply phase Self_USS is operational. In this the immediate predecessor monitors the performance. If node Ni misbehaves, and say announces a wrong sequence number. This is immediately observed by the node [N.sub.i+1] which sets the status bit of [N.sub.i] to red and starts RREP salvaging operation. On the other hand if [N.sub.i] simply loses communication link with [N.sub.i+1]) [N.sub.i] is not booked; only route reply salvaging is undertaken.

3.2.2 Data Forwarding Phase

During data forwarding phase if node [N.sub.i] misbehaves i.e. it drops packets, it is observed by nodes [N.sub.i-1] , [U.sub.i-1], and [U.sub.i+1] and they send M-Flag messages, and convict the guilty node ; further they unicast messages among themselves to establish an alternative path via [N.sub.i-1] , [U.sub.i], [U.sub.i+1] and [N.sub.i+1]. In this segment Self_USS will operational. On the other hand if node [N.sub.i] goes out of communication link but the umpiring nodes do not receive the hello messages from [N.sub.i] and simply switch over to alternative path, thus booking of innocent node is avoided. Thirdly umpire [U.sub.i] may go out of communication link with [U.sub.i-1], [N.sub.i-U] [N.sub.i], and [U.sub.i+1] In such a case [N.sub.i-1] monitors [N.sub.i] and [N.sub.i] monitors [N.sub.i+1] under Self_USS. [Table 1].

4. SIMULATIONS AND RESULTS

We use a simulation model based on QualNet 5.0 in our evaluation[5]. Our performance evaluations [38-40] are based on the simulations of 100 wireless mobile nodes that form a wireless ad hoc network over a rectangular (1500 X 600 m) flat space. The MAC layer protocol used in the simulations was the Distributed Coordination Function (DCF) of IEEE 802.11. The performance setting parameters are given in Table 2.

Before the simulation we randomly selected a certain fraction, ranging from 0 % to 40 % of the network population as malicious nodes. We considered only two attacks modifying the hop count and dropping packets. Each flow did not change its source and destination for the lifetime of a simulation run. We had kept the simulation time as 1500s, so as to enable us to compare our results with that of SCAN.

4.1. Throughput

In the world of MANET, packet delivery ratio has been accepted as a standard measure of throughput. Packet delivery ratio is nothing but a ratio between the numbers of packets received by the destinations to the number of packets sent by the sources. We present in Tables 3, 4 and 5 the packet delivery ratios, for malicious node percentages of 0, 10, 20, 30 and 40, with node mobility varying between 0 to 20 m/s. From Tables 3, 4 and 5 the following observations can be made. For Tables 4 and 5 TUS results correspond to plain triple umpire system where as ETUS results correspond to TUS with the incorporation of salvaging.

1. In general packet delivery ratio decreases as mobility and percentage of malicious nodes increase.

2. In the case of plain AODV, with 0% malicious nodes, packet delivery ratio drops from 98.28%, when the nodes are stationary to 93.73%, when the nodes are moving at 20 m/s.

3. With plain AODV, packet delivery ratio has a steep fall from 98.28 (0% malicious nodes, mobility = 0 m/s) to 26.04 (40% malicious nodes, mobility = 20 m/s). The corresponding values for TUS are 98.98, 63.79 and 99.99, 67.98 for ETUS. Thus throughput is increased by 145 % for TUS and by 161 % for ETUS.

4. As compared to plain AODV, TUS results are superior.

5. Again comparing TUS and ETUS results from tables 4 & 5, we find throughput with ETUS is higher.

6. Throughput with ETUS is higher as compared to plain AODV, self_USS, self_USS with SRR, SCAN and TUS.

[FIGURE 4 OMITTED]

From the above results we conclude that ETUS leads to a substantial improvement over plain AODV, Self_USS, Self_USS with SRR and TUS, from the point of view of throughput. The other question to be answered is how does ETUS compare with SCAN? We present the details in Figure 4, where a comparison corresponding to 30% malicious nodes with mobility varying from 0 to 20 m/s is given. The data for SCAN corresponds to Figure. 8 of the paper [2]. TUS and SCAN are use results as shown in the Figure 4. We make no claims and offer our comments in the analysis section.

4.2 Failure to deduct (False Negatives) Probability

Figure 5 and Figure 6 present failure to deduct probability as a function of mobility and percentage malicious nodes of TUS and ETUS respectively. False Negatives Probability, which is the chance that umpires, fails to convict and isolate a malicious node.

The above definition requires some elaboration. We can think of two groups of malicious nodes that are left undetected. In the first group are those nodes, which never played a part in the network operation; they were probably traveling along the boundaries and never had a chance to participate in the network activity. The second groups of malicious nodes are those that played a role as a forwarding node, but went undetected. Clearly our umpiring system is responsible only for the second group. The first group of nodes is similar to reserve players in the sidelines and clearly any umpire cannot show red flag and march off players in the sidelines. Appropriately we have done the failure to detect probability calculation taking into consideration only those nodes, which took part in the network activity. Other researchers adopt the same approach also. The results are similar that of SCAN [2] as shown in the Figure. 7. We find that false negative probability has decreased with ETUS.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

4.3 False Accusation (False Positives) Probability

False accusation probability, which is the chance that umpires incorrectly convicts and isolates a legitimate node. Figure 8 and Figure 9 Presents false accusation probability as a function of mobility and percentage malicious nodes for TUS and ETUS respectively. We find similar decrease in false accusation probability at all other combinations of malicious node percentages and mobility values, with ETUS. This is the probability of wrongly booking innocent nodes. We find false positive probability increases with increasing percentage of malicious nodes and increased mobility.

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

The values vary between 0 to 10% and are similar to the patterns obtained for SCAN [2] as shown in the Figure 10. It is seen that with ETUS False Positive Probabilities slightly decrease.

[FIGURE 9 OMITTED]

4.4 Communication Overhead

Communication overhead can be evaluated based on the number of transmissions of control messages like RREQ, RREP, and RERR in the case of plain AODV and in addition M_ERROR, M-Flag, Umpire, Neighbor list messages in the TUS and ETUS. In addition salvaging concept introduced in ETUS, it uses special control messages like [RREQ.sub.SRR], [RREP.sub.SRR] and [RERR.sub.SRR]. RREQ are to be decimated to the entire network, where as RREP messages are unicasts. Table 7 shows that with communication overhead slightly decrease because salvaging concept in ETUS.

[FIGURE 10 OMITTED]

From Table 6 and Table 7 following inferences can be drawn:

1. The communication overhead increases with increasing percentage of malicious nodes and mobility for both TUS and ETUS.

2. For TUS control overhead, the increase's from 14244 (0% malicious nodes; mobility = 0) to 30678 (40% malicious nodes and mobility = 20 m/s). The corresponding variation for ETUS is from 13111 to 26521. It is seen that with communication overhead slightly decrease because to play the role of umpires has an immediate benefit in salvaging operations with ETUS proposed by us. We find that maximum decrease in communication overhead is 15.67 %.

5. RELATED WORKS

The Key Distribution Center (KDC) architecture is the main stream in wired network because KDC has so many merits: efficient key management, including key generation, storage, and distribution and updating. The lack of Trusted Third Party (TTPs) key management scheme is a big problem in ad hoc network [6-30]. Yong et al. [9][12] propose a novel cryptography for ad hoc network security, where they present a new digital signature algorithm for identity authentication and key agreement scheme. Their scheme has no central administrator. They have shown that their scheme can withstand man-in-middle and Byzantine mode conspiracy attacks. Zhang and Lee [15] were among the first to study the problem of intrusion and detection in wireless ad hoc networks. They present two algorithms in this connection. Rendong Bai and Mukesh Singhal [31] proposed the SRR mechanism, which uses to avoid unnecessary route discovery to establish the route because of the loss of the RREP messages. All the above schemes only try to protect the system from the attacker, but not bother about quarantining attackers. The twin systems of watchdog and pathrater [1] not only detect the mischievous nodes but also prevent their further participation in the network. SCAN [2] also has similar action, but is more comprehensive, in the sense not only packet dropping but also other misbehaviors like giving wrong hop count are covered.

6. Conclusions

We have conducted simulation studies using QualNet 5.0 to evaluate the performance of TUS and ETUS. The results show that ETUS significantly improves the performance of TUS in all metrics, namely, packet delivery ratio, false positives, false negatives and control overhead. In the presence of 40 % malicious node, ETUS yields a packet delivery ratio of 67.98 %, which is an improvement of 6.5 % over TUS. Further, false positives and false negatives are reduced by 47.28 % and 27.39 %, .Further there is a decrease in control overhead of 15.67 % as compared to TUS. Our proposed ETUS system is unique in that it not only detects and quarantines malicious nodes but also restores the communication between the source and destination. Our future work will focus on modeling generic network layer attacks.

References

[1] Sergio Marti, T.J. Giuli, Kevin Lai and Mary Baker, "Mitigating routing misbehavior in mobile ad hoc networks", in proc. ACM MobiCom, 2000, pp- 255-265.

[1] Hao Yang, James Shu, Xiaoqiao Meng and Songwu Lu, "SCAN: Self-Organized Network-Layer Security in Mobile ad hoc networks", IEEE Journals on selected areas in communications, vol. 24, No. 2, February 2006.

[2] Sung-Ju Lee and Mati Gerla, "AODV-BR: Backup routing in ad hoc networks", in proc. IEEE wireless communication and Networking conference (WCNC), vol.3, 2000, pp. 1311-1316.

[3] Perumal Sambasivam, Ashwin Murthy and Elizabeth M. Belding-Royer, "Dynamically adaptive multipath routing based on AODV", in proc. MedHocNet, Turkey, June 2004, pp. 106-117.

[4] Scalable Networks Technologies: QualNet simulator version 5.0 http ://www. scalable-networks.com

[5] Marianne A. Azer, Sherif M. El-Kassas, and Magdy S. El-Soudani, "Certification and revocation schemes in ad hoc networks survey and challenges, in proc. IEEE ICSNC 2007.

[6] J. Kong, P. Zerfos, H. Luo, S. Lu, and L. Zhang, "Providing robust and ubiquitous security support for MANET", in Proc. IEEE ICNP, 2001, pp. 251-260.

[7] Lei Feng-Yu, Cui Guo-Hua, and Liao Xiao-Ding, "Ad hoc Networks security mechanism based on CPK", in proc. IEEE ICCISW, 2007, pp. 522-525.

[8] Pi Jian Yong, Liu Xin Song, Wu Ai, Liu Dan, "A Novel Cryptography for Ad Hoc Network Security", in Proc. IEEE 2006, pp. 1448-1451.

[9] Michael Hauspie, and Isabelle Simplot-Ryl, "Enhancing nodes cooperation in ad hoc networks", in proc. IEEE 2007, pp. 130-137.

[10] S. Capkun, L. Buttyan and J. Hubaux, "Self-organized public-key management for mobile ad hoc networks", IEEE Trans. Mobile Computing, vol. 2, No. 1, pp. 52-64, January, 2003.

[11] Pi Jian Yong, Liu Xin Song, Wu Ai, Liu Dan, "A Novel Cryptography for Ad Hoc Network Security", in Proc. IEEE 2006, pp. 1448-1451.

[12] J. Hubaux, L. Buttyan, and S. Capkun, "The quest for security in Mobile ad hoc networks", in Proc. ACM MobiHoc, 2001, pp. 146-155.

[13] William Stallings, "Cryptography and network Security principles and Practices", Pearson Education, First edition, 2007.

[14] Y. Zhang and W. Lee, "Intrusion detection in wireless ad hoc networks", in Proc. ACM MobiCom, 2000, pp. 275-283.

[15] S. Capkun, J.Hubaux, and L. Buttyan, "Mobility helps security in ad hoc networks", in Proc. ACM MobiCom, 2003, pp 46-56.

[16] J. Hubaux, I. Buttyan and S. Capkun, "The quest for security in mobile ad hoc networks", in Proc. ACM MobiHoc 2001, pp. 251-260.

[17] Sergio Marti, T.J. Giuli, Kevin Lai and Mary Baker, "Mitigating routing misbehavior in mobile ad hoc networks", in proc. ACM MobiCom, 2000, pp- 255-265.

[18] Y. Hu, D. Johnson, and A. Perrig, "SEAD: Secure efficient distance vector routing for mobile wireless ad hoc networks", in Proc. IEEE WMCSA, June 2002, pp. 3-13.

[19] Y. Hu, A. Perrig, and D. Johnson, "Ariadne: A Secure on-demand routing for ad hoc networks", in Proc. ACM MobiCom, 2002, pp. 12-23.

[20] P. Papadimitratos and Z. Haas, "Secure routing for mobile ad hoc networks", in Proc. CNDS, 2002, pp. 193-204.

[21] K. Sanzgiri, B. Dahill, B. Levine, C. Shields, and E. Royer, "A secure protocol for ad hoc networks," in Proc. IEEEICNP, 2002, pp. 78-89.

[22] M. Zapata and N. Asokan, "Securing ad hoc routing protocols", in Proc. ACM Wise, 2002, pp.1-10.

[23] Azeddine Attir, Farid Nait Abdesselem, Brahim Bensaou, and Jalel Ben-Othman, "Logical Wormhole Prevention in Optimized Link State Routing Protocol", in proc IEEE GLOBECOM 2007, pp. 1011-1016.

[24] Nidal Nasser and Yunfeng Chen, "Enhanced Intrusion Detection System for discovering malicious nodes in mobile ad hoc networks", in proc. IEEE ICC, 2007, pp. 1154-1159.

[25] S. Basagni and et al, "A Distance Routing Effect Algorithm for Mobility (DREAM)", in Proc. MOBICOM, October 1998.

[26] V. Bharghavan, A.Demers, S.Shenkar and L.Zhang, "MACAW: A Medium Access Protocol for Wireless LANs", in Proc. of ACM SIGCOMM, August 1994.

[27] J. Broach, D. A. Maltz, D.B. Johnson, Y.C. Hu and J. Jetcheva, "A Performance Comparison of Multi-hop Wireless Adhoc Network Routing Protocols " in Proc. MOBICOM, October 1998.

[28] R. Castaneda and S R Das, "Query Localization Techniques for on demand Routing Protocols in Ad hoc networks", in Proc. MOBICOM, August, 1999.

[29] A. Perrig, R.Canetti, D.Song, and J.Tygar, "Efficient and Secure source authentication for Multicast", in Proc. NDSS, 2001, pp. 35 46.

[30] Rendong Bai and Mukesh Singhal, "Salvaging Route Reply for On-Demand Routing Protocols in Mobile Ad Hoc Networks", in Proc. ACM MSWiM, 2005, pp. 53-62.

[31] Kathirvel A, and Srinivasan R, "Enhanced Self Umpiring System for Security using Salvaging Route Reply", International Journal of Computer Theory and Engineering, Vol. 2, No. 1, 2010.

[32] Kathirvel A, and Srinivasan R, "SelfUSS: A Self Umpiring System for Security in Mobile Ad-Hoc Network", International Journal of Engineering and Technology. Singapore. (Accepted and yet to be published)

[33] Kathirvel A, and Srinivasan R, "A System of Umpires for Security of Wireless Mobile Ad Hoc Network", International Arab Journal of e Technology. (Accepted and yet to be published)

[34] Kathirvel A, and Srinivasan R, "A Study on Salvaging Route Reply for AODV Protocol in the Presence of Malicious Nodes", International Journal of Engineering and Technology, Vol. 1, No. 2, 2009, pgno. 151-155.

[35] Kathirvel A, and Srinivasan R, "Single Umpiring System for Security of Mobile Ad Hoc Networks", Journal of Advances in Wireless Mobile Communication, Vol. 2, No. 2, pp 141-152, 2009.

[36] Kathirvel A, and Srinivasan R, "Triple Umpiring System for Security of Mobile Ad Hoc Networks", International Journal of Engineering and Information technology, Vol. 1, No. 2, pp 95-100, 2009.

[37] Kathirvel A, and Srinivasan R, "Global Mobile Information System Simulator in Fedora Linux", ACM online Computer Commucation Review, 2009.

[38] Kathirvel A., Subburam S. and Srinivasan R, "Performance Enhancement of On-Demand Routing Protocols in Mobile Ad-Hoc Networks", Second National Conference on Innovations in Information and Communication Technology (NCIICT--2006), pp. 169-174.

[39] Kathirvel A and Srinivasan R, "Reactive Route Recovery for Link Failure in MANET", Proceedings of IEEE National Conference on Information and Communication Convergence (IEEE ICC--2006), pp. 42-49.

Ayyaswamy Kathirvel (1), and Rengaramanujam Srinivasan (2)

(1) B S Abdur Rahman University, School of Computer and Information Sciences, GST Road, Vandalur, Chennai, India

kathir@crescentcollege.org

(2) B S Abdur Rahman University, School of Computer and Information Sciences, GST Road, Vandalur, Chennai, India

rsvasan@crescentcollege.org
Table 1. Response of ETUS for various situations

Data forwarding phase

2 c.                2b.                 2a.

Umpire [U.sub.i]    [N.sub.i] out of    [N.sub.i]
out of              communication       misbehaves
communication       range
range

[U.sub.i] loses     [N.sub.i] loses     [N.sub.i] drops
power of goes out   power or goes out   packets
of communication    of communication
link                zone.

[N.sub.i-i],        Alternative path    [N.sub.i] is
[N.sub.i] find      is established      marked with red
out.                via [N.sub.i-1],    flag by
                    [U.sub.i],          [N.sub.i-1],
                    [U.sub.i+1] and     [U.sub.i],
                    [N.sub.i+1].        [U.sub.i+1].
                                        Alternative path
                                        is established
                                        via [N.sub.i-1],
                                        [U.sub.i],
                                        [U.sub.i+1] and
                                        [N.sub.i+1].

Loss of umpire.     [N.sub.i] is not    [N.sub.i] is
[N.sub.i-1]         booked. Current     booked; current
monitors            segment works       segment works
[N.sub.i] and       under self          under self
[N.sub.i]           umpiring mode.      umpiring nodes
monitors
[N.sub.i+1]

RREP phase                              Phase

lb.                 la.                 S.No

[N.sub.i] out of    [N.sub.i]           Situation
communication       misbehaves
range

Communication       [N.sub.i] gives     Event
lost with           a wrong hop count
[N.sub.i]           or sequence
                    number

[N.sub.i+1]         [N.sub.i+1] marks   Action
initiates           [N.sub.i] with
Route reply         red flag.
salvaging           Initiates Route
operations          reply salvaging
as salvor           operations as
                    salvor

Self_USS mode throughout RREP phase     Comments

Table 2 Parameter Settings

Simulation Time      1500 seconds
Propagation model    Two-ray Ground Reflection
Transmission range   230 m
Bandwidth            2 Mbps
Movement model       Random way point
Maximum speed        0-20 m/s
Pause time           0 seconds
Traffic type         CBR (UDP)
Payload size         512 bytes
Number of flows      10/20

Table 3 Packet delivery ratios for plain AODV

                          Plain AODV

Mobility           Percentage of Malicious nodes
(M/s)
               0       10       20       30       40

0          98.28    82.56    76.89    70.44    64.34
5          96.28    70.59    57.85    45.18    37.24
10         95.11    65.61    51.28    37.89    31.89
15         94.12    62.45    47.51    32.55    26.17
20         93.73    61.22    45.57    32.07    26.04

Table 4 Packet delivery ratio for TUS

                              TUS

Mobility            Percentage of Malicious nodes
(M/s)
               0       10       20       30       40

0           98.98    96.35   94.36    92.95     87.84
5           97.98   94.65     91.67    88.45    81.84
10          96.62   92.56     88.56    84.90    76.82
15         95.42    90.45     85.44    83.57    70.74
20          94.93    88.93    84.76    81.59    63.79

Table 5 Packet delivery ratio for ETUS

                              ETUS

Mobility            Percentage of Malicious nodes
(M/s)
               0       10       20       30       40

0           99.99   99.98    99.12    95.71    92.10
5           99.97   99.43     98.12   93.65     84.12
10          99.72    98.91    95.91   91.42     78.65
15          98.79    95.89    94.79    88.61    74.12
20          98.49    94.92   93.12     87.07    67.98

Table 6 Communication overhead for TUS

Mobility                      TUS
(M/s)
                    Percentage of Malicious nodes

               0%      10%      20%      30%      40%

0          14244     17869    21485    24104    27089
5          14934     18697    22734    25045    27678
10          15367    19658    23579    25937    29367
15          15915    20375    24265    26798    30561
20          17434    21278    24934    27899    30678

Table 7 Communication overhead for ETUS

Mobility                      ETUS
(M/s)
                    Percentage of Malicious nodes

               0%      10%      20%      30%      40%

0           13111    16177    19774    21844    23184
5           13327    16944    19949    22009    23849
10          13564    18885    21041    22937    24868
15          13907    19052    22112    23178    25869
20          14504    19039    22769    24624    26521
COPYRIGHT 2010 Kohat University of Science and Technology
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Kathirvel, Ayyaswamy; Srinivasan, Rengaramanujam
Publication:International Journal of Communication Networks and Information Security (IJCNIS)
Geographic Code:1USA
Date:Aug 1, 2010
Words:5680
Previous Article:Tested evaluation of fast and secure handover in FMIPv6.
Next Article:QoS provisioning in three layer MIPv6 architecture using RSVP.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters