New Algorithms for the Efficient Design of Topology-Oriented Key Agreement Protocols in Multi-Hop Ad Hoc Networks
Maria Striki, John S. Baras, Kyriakos Manousakis
Number: ISR TR 2006-13, Year: 2006.
Securing group communications in resource constrained, infrastructure-less environments such as Mobile Ad Hoc Networks (MANETs) has become one of the most challenging research directions in the areas of wireless network security. MANETs are emerging as the desired environment for an increasing number of commercial and military applications, addressing also an increasing number of users. Security on the other hand, is becoming an indispensable requirement of our modern life for all these applications. The inherent limitations of such dynamic and resource-constraint networks impose major difficulties in establishing a suitable secure group communications framework. This is even more so for the operation of Key Agreement (KA), under which all parties contribute equally to the group key. The logical design of efficient KA protocols has been the main focus of the related research to-date. Such a consideration however, gives only a partial account on the feasibility and actual performance of a KA protocol in a real multi-hop network. This is because protocols have been evaluated only in terms of the group key related messaging in isolation from the underlying network functions that interact with the logical scheme and support its correct execution (i.e. routing). In this work, we contribute towards efficiently extending a number of Diffie-Hellman (DH)-based group KA protocols in wireless multihop ad hoc networks, and measuring their actual performance over these networks. Towards this end, we introduce a number of new algorithms and techniques that aim in efficiently merging the logical design of KA protocols with the underlying routing and topology of the network, to produce protocols that substantially improve one or more metrics of interest. Indeed, the resulting protocols are significantly more efficient in some or all of the above metrics, as our analytical and simulation results indicate.