Survey of Trust Models in Different Network Domains

Abstract

This paper introduces the security and trust concepts in wireless sensor networks and explains the difference between them, stating that even though both terms are used interchangeably when defining a secure system, they are not the same. The difference between reputation and trust is also explained, highlighting that reputation partially affects trust. A survey of trust and reputation systems in various domains is conducted, with more details given to models in ad-hoc and sensor networks as they are closely related to each other and to our research interests. The methodologies used to model trust and their references are presented. The factors affecting trust updating are summarised and some examples of the systems in which these factors have been implemented are given. The survey states that, even though researchers have started to explore the issue of trust in wireless sensor networks, they are still examining the trust associated with routing messages between nodes (binary events). However, wireless sensor networks are mainly deployed to monitor events and report data, both continuous and discrete. This leads to the development of new trust models addressing the continuous data issue and also to combine the data trust and the communication trust to infer the total trust.

Complex Adaptive WSNs for Polluted Environment Monitoring

“…Another model which can be used is to use a trust agent which acts as a broker; however when the aggregate agent tries to communicate with the sensing agent in order to collect its data it will be checked by the trust agent which will decide if it is trusted or not according to its tag and policies. However these are just examples to illustrate the trust between aggregate and sensing agents, many other models can be used [18].…”

Enhancing Wireless Sensor Network Security using Artificial Neural Network based Trust Model

“…Generally, as any wireless sensor network built in a harsh environment the proposed system is exposed to security attacks such as DOS, wormhole, Sybil attack, hello flood and nodecapturing attacks. The use of cryptographic algorithms alone cannot encounter and cope with the various types of attacks, so building trust between several nodes within the network is important to distinguish legitimate nodes from malicious ones [6].…”

“…Different trust models have been proposed to provide an ultimate mechanism for detecting malicious nodes within WSN. All of the proposed trust models based on calculating www.ijacsa.thesai.org trust rate for every node within the system depending on its behavior which is captured by either direct or indirect fashion, depending on these rates the controller or director decide to consider a node as legitimate node and in result raise its trust rate or it could be a layer and punished by reducing its trust rate, if the rate decreased below a specific threshold the node may be discarded or monitored to be treated later [6].…”

“…Several trust schemes had been used to discover the malicious nodes in WSN, all of them are based on two essential characteristics in terms of WSN resources limitations; in one hand they are lightweight and need less power, processing and communications, in the other hand they are powerful and capable of managing trust between various heterogeneous nodes [6].The proposed trust model aims to enhance the use of artificial neural network(ANN) in WSN using radial basis function benefiting from its simplicity in implementation to provide a comprehensive trust model that supports system security and rationalizes resources consumption.…”

An Efficient Billing Scheme for Trusted Nodes Using Fuzzy Logic in Wireless Sensor Networks

“…Studying the trusted node or trusted path model [5] differs from studying the security models in WSNs [6]. Security models address the preservation of a network from malicious nodes that may attack the network and strain their resources.…”