1-- 4244 – 0731 – 1 / 06 / $ 20.00 @ 2006 IEEE A Distributed Trust Mechanism for Mobile Ad Hoc Networks # Jaydip Sen 1 , Piyali Roy Chowdhury 2 , Indranil Sengupta 3 1 Department of Computer Science and Engineering, Indian Institute of Technology, Kharagpur-721302, INDIA, email: sen_jaydip@yahoo.com 2 Department of Computer Science and Engineering, Future Institute of Engineering and Management Kolkata-700150, INDIA, email:roychowdhury.piyali@gmail.com 3 Department of Computer Science and Engineering, Indian Institute of Technology, Kharagpur-721302, INDIA, email:isg@iitkgp.ac.in Abstract In a multi-hop hop mobile ad hoc network (MANET), mobile nodes cooperate with each other without using any infrastructure such as access points or base stations. The nodes’ mobility and fundamentally limited capacity of the wireless medium, together with wireless transmission effects such as attenuation, multi-path propagation, and interference combine to pose significant challenges for security in MANETs. Traditionally, cryptographic mechanisms such as authentication and encryption are not capable of handling some kinds of attacks such as packet dropping by malicious nodes. In this paper, we propose a mechanism for detection of malicious packet dropping attack in MANET. The mechanism depends on a trust module based on the ‘reputation’ of different nodes in the network. Every node computes the reputation of its neighbors by directly monitoring their behaviors. The reputation information is gathered, stored, and exchanged between the nodes and computed under different scenarios. We have validated the scheme with some simulation work, which shows the feasibility and efficiency of performance of the protocol. 1. INTRODUCTION Although the security objectives of both ad hoc networks and traditional networks are considered the same such as availability, confidentiality, integrity, authentication, and non- repudiation, the security issues involved in ad hoc networks are quite different due to the ‘mobile’ and ‘ad hoc’ constraints, i.e. limited resources for computation and communication, dynamic network topology as well as the mobility of the hosts. In traditional networks, most trust evidences are generated via potentially lengthy assurance processes, distributed off-line, and assumed to be valid on a long term. In contrary, few of these characteristics of trust relations and trust evidences are prevalent in MANETs. Cryptographic primitives such as authentication and key distribution are the usual mechanisms used for implementing security in MANETs. However, these schemes cannot provide security against some attacks such as packet dropping attack by malicious nodes in the network. We can identify two types of uncooperative nodes in MANETs: malicious nodes and selfish nodes. The nodes belonging to the first category are either faulty and therefore cannot follow a protocol, or are intentionally malicious and try to attack the system. The problems created by these nodes need to be addressed at many layers, for example, using spread-spectrum encoding to avoid interference over the communication channel, using a reputation system to identify the malicious system, and subsequently avoid or penalize such nodes. A selfish node, on the other hand, is an economically rational node whose objective is to maximize its own welfare, which is defined as the benefit of its actions minus the cost of its actions. Since forwarding a message will incur a cost, a selfish node will need incentive for doing it. One possibility to provide incentive is to use a reputation system [1][2][3]. Such reputation systems, however, may have several issues. First, there is no formal specification and analysis of the type of incentive provided by such systems. Second, these systems have not considered the possibility that even selfish nodes can collude with each other in order to maximize their welfare. Third, some of these systems depend on broadcast nature of wireless networks for monitoring. Such monitoring, however, may not always be possible due to asymmetric links when the nodes use power control. Further, directional antennas [4], which are gaining momentum in the field of wireless networks, will also make monitoring hard. Another approach to provide incentive is to use credit or virtual currency. Buttyan and Hubaux proposed a nice solution of this type in [8], and then presented an improved result based on credit counters [9]. For both proposals, a node receives one unit of credit for forwarding a message of another node, and such credits are deducted from the sender (or the destination). However, both these proposals require a tamper-proof hardware at each node to ensure correct credit and debit of nuglets. In this paper, we have proposed a security protocol that detects malicious packet dropping attacks in MANETs. The mechanism depends on a trust model that is based on the ‘reputation’ of different nodes. MANETs are also prone to the following security threat: a node could be tempted to discard its initial identity and re-enter the network in disguise in environments where users are punished for their selfish or