Resilient Cluster-Organizing Key Management and Secure Routing Protocol for Mobile Ad Hoc Networks

Hua-Yi LIN
Yueh-Min HUANG
Tzone-I WANG

Publication
IEICE TRANSACTIONS on Communications   Vol.E88-B    No.9    pp.3598-3613
Publication Date: 2005/09/01
Online ISSN: 
DOI: 10.1093/ietcom/e88-b.9.3598
Print ISSN: 0916-8516
Type of Manuscript: Special Section PAPER (Special Section on Advances in Ad Hoc Mobile Communications and Networking)
Category: 
Keyword: 
ad hoc networks,  PKI,  CA,  CSBRP,  node-based,  cluster-organized,  

Full Text: PDF(1.4MB)>>
Buy this Article



Summary: 
Unlike traditional networks, the characteristics of mobile wireless devices that can dynamically form a network without any infrastructure and wired line mean that mobile ad hoc networks frequently display partition owing to node mobility or link failures. Consequently, an ad hoc network is difficult to provide on-line access to trusted authorities or centralized servers. Despite the existence of well-known security mechanisms, the absence of a stationary central authorization facility in an open and distributed communication environment is a major challenge. Consequently, applying traditional Public Key Infrastructure (PKI) security architecture to mobile ad hoc networks will create secure blind sides. Based on this perspective, this study proposes a novel scalable and robust cluster-organized key management scheme. Distribution of trust to an aggregation of cluster heads using a threshold scheme faculty provides mobile ad hoc networks with robust key management. Furthermore, the proposed approach provides Certificate Authority (CA) with a fault tolerance mechanism to prevent a single point of compromise or failure, and saves CA large repositories from maintaining member certificates, making the proposed approach more suitable for numerous mobile devices. Additionally, this study proposes a Cluster Secure Based Routing Protocol (CSBRP) to integrate into the key management to enhance non-repudiation of routing information and routing performances. Finally, this study introduces a mathematical model to demonstrate that the proposed cluster-based communication outperforms the node-based approach.