Towards reliable self-clustering Mobile Ad Hoc Networks

Abstract

With the introduction of Mobile Ad Hoc Networks (MANETs), nodes are able to participate in a dynamic network which lacks an underlying infrastructure. In this paper, we present a novel approach to improve the search efficiency and scalability of MANETs by clustering nodes based on trust mechanism. In our method, the trust relationship is formed by evaluating the level of trust using Bayesian statistic analysis, and clusters can be formed and maintained autonomously by nodes with only partial knowledge. Simulation results show that each node can form and join proper clusters based on their trust degree, and the cluster-based search algorithm with trust mechanism outperforms over those in current popular clustering models.

Introduction

A Mobile Ad Hoc Network (MANET) is a self-configuring network in which nodes rely on other nodes for communications. Security poses some fundamental challenges in such networks as they are not conducive to centralized trusted authorities. Although several secure routing protocols [1], [2] have been proposed to defend against predefined attacks, they are vulnerable to new and dynamically changing attacks. This has led to the development of several trust models [3], [4], [5], [6], [7], [8], in which mobile nodes capture evidence of trustworthiness of other nodes to quantify and represent their behavior, and then to establish trust relationships with them.

However, there are two crucial issues of current trust models, one is the inefficient query routing, which makes these models not highly scalable. For example, in reference [3], queries will be flooded throughout the whole network and the search traffic increases dramatically with the network size. The other is how to select appropriate nodes to cooperate with, since the candidate nodes are autonomous and may be unreliable or dishonest. This raises the question of how much credence to give each resource, and we cannot expect each user to know the trustworthiness of each resource. Without a good solution to this problem, MANET systems are not likely to be deployed for serous applications. Essentially, nodes need to manage the risk of communicating or cooperating with each other without prior experience and knowledge about each other.

In this paper, we present on how our novel Bayesian method based trust model facilitates mobile nodes to explicitly represent and manage behavior evidence in their trust relationships with other nodes. In addition, we propose a trust-based self-clustering algorithm to improve search performance and scalability of MANETs with trust assurance. Query messages are limited within nodes’ own or nearby clusters, which can make a query hit quickly with low search overhead. This paper is based on our previous work [4], and larger expansion is carried out in this paper with the main contributions listed below:

• A comprehensive Bayesian trust model is proposed to evaluate the trust degree of nodes;

• A novel approach to improve the search efficiency and scalability of MANETs by clustering nodes based on trust mechanism is proposed;

• Simulation results show that each node can form and join proper clusters based on their trust degree and the cluster-based search with trust assurance performance outperforms those in current popular trust models.

The rest of this paper is organized as follows. We review some related work in Section 2. In Section 3, we introduce the proposed clustering algorithm. The evaluation of our approach by simulations is given in Section 4 and finally we conclude this paper in Section 5.