Date of Award

Fall 2009

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Electrical Engineering - (Ph.D.)

Department

Electrical and Computer Engineering

First Advisor

Nirwan Ansari

Second Advisor

Roberto Rojas-Cessa

Third Advisor

Edwin Hou

Fourth Advisor

Yanchao Zhang

Fifth Advisor

Rajarathnam Chandramouli

Abstract

Group communications has been facilitating many emerging applications which require packet delivery from one or more sender(s) to multiple receivers. Owing to the multicasting and broadcasting nature, group communications are susceptible to various kinds of attacks. Though a number of proposals have been reported to secure group communications, provisioning security in group communications remains a critical and challenging issue.

This work first presents a survey on recent advances in security requirements and services in group communications in wireless and wired networks, and discusses challenges in designing secure group communications in these networks. Effective security services to secure group communications are then proposed. This dissertation also introduces the taxonomy of security services, which can be applied to secure group communications, and evaluates existing secure group communications schemes.

This dissertation work analyzes a number of vulnerabilities against trust and reputation systems, and proposes a threat model to predict attack behaviors. This work also considers scenarios in which multiple attacking agents actively and collaboratively attack the whole network as well as a specific individual node. The behaviors may be related to both performance issues and security issues. Finally, this work extensively examines and substantiates the security of the proposed trust and reputation system.

This work next discusses the proposed trust and reputation system for an anonymous network, referred to as the Adaptive Trust-based Anonymous Network (ATAN). The distributed and decentralized network management in ATAN does not require a central authority so that ATAN alleviates the problem of a single point of failure. In ATAN, the trust and reputation system aims to enhance anonymity by establishing a trust and reputation relationship between the source and the forwarding members. The trust and reputation relationship of any two nodes is adaptive to new information learned by these two nodes or recommended from other trust nodes. Therefore, packets are anonymously routed from the 'trusted' source to the destination through 'trusted' intermediate nodes, thereby improving anonymity of communications. In the performance analysis, the ratio of the ATAN header and data payload is around 0.1, which is relatively small.

This dissertation offers analysis on security services on group communications. It illustrates that these security services are needed to incorporate with each other such that group communications can be secure. Furthermore, the adaptive trust and reputation system is proposed to integrate the concept of trust and reputation into communications. Although deploying the trust and reputation system incurs some overheads in terms of storage spaces, bandwidth and computation cycles, it shows a very promising performance that enhance users' confidence in using group communications, and concludes that the trust and reputation system should be deployed as another layer of security services to protect group communications against malicious adversaries and attacks.

Share

COinS