Date of Award

Winter 2006

Document Type


Degree Name

Doctor of Philosophy (PhD)


Computer Science

Committee Director

Ravi Mukkamala

Committee Member

Mohamed Eltoweissy

Committee Member

Stephan Olariu

Committee Member

Hussein Abdel-Wahab

Committee Member

Min Song


A growing number of secure group applications in both civilian and military domains is being deployed in WAHNs. A Wireless Ad-hoc Network (WARN) is a collection of autonomous nodes or terminals that communicate with each other by forming a multi-hop radio network and maintaining connectivity in a decentralized manner. A Mobile Ad-hoc Network (MANET) is a special type of WARN with mobile users. MANET nodes have limited communication, computational capabilities, and power. Wireless Sensor Networks (WSNs) are sensor networks with massive numbers of small, inexpensive devices pervasive throughout electrical and mechanical systems and ubiquitous throughout the environment that monitor and control most aspects of our physical world.

In a WAHNs and WSNs with un-trusted nodes, nodes may falsify information, collude to disclose system keys, or even passively refuse to collaborate. Moreover, mobile adversaries might invade more than one node and try to reveal all system secret keys. Due to these special characteristics, key management is essential in securing such networks. Current protocols for secure group communications used in fixed networks tend to be inappropriate. The main objective of this research is to propose, design and evaluate a suitable key management approach for secure group communications to support WAHNs and WSNs applications.

Key management is usually divided into key analysis, key assignment, key generation and key distribution. In this thesis, we tried to introduce key management schemes to provide secure group communications in both WAHNs and WSNs.

Starting with WAHNs, we developed a key management scheme. A novel architecture for secure group communications was proposed. Our proposed scheme handles key distribution through Combinatorial Key Distribution Scheme (CKDS). We followed with key generation using Threshold-based Key Generation in WAHNs (TKGS). For key assignment, we proposed Combinatorial Key Assignment Scheme (CKAS), which assigns closer key strings to co-located nodes. We claim that our architecture can readily be populated with components to support objectives such as fault tolerance, full-distribution and scalability to mitigate WAHNs constraints. In our architecture, group management is integrated with multicast at the application layer.

For key management in WSNs, we started with DCK, a modified scheme suitable for WSNs. In summary, the DCK achieves the following: (1) cluster leader nodes carry the major part of the key management overhead; (2) DCK consumes less than 50% of the energy consumed by SHELL in key management; (3) localizing key refreshment and handling node capture enhances the security by minimizing the amount of information known by each node about other portions of the network; and (4) since DCK does not involve the use of other clusters to maintain local cluster data, it scales better from a storage point of view with the network size represented by the number of clusters.

We went further and proposed the use of key polynomials with DCK to enhance the resilience of multiple node capturing. Comparing our schemes to static and dynamic key management, our scheme was found to enhance network resilience at a smaller polynomial degree t and accordingly with less storage per node.