Institution: University of Connecticut Abstract Date: 08/05/08 This project focuses on the development of a new class of secret key generation and renewal algorithms for securing wireless networks by taking advantage of physical layer characteristics. The basis of the approach is the identification of measurable quantities of the wireless channel between a pair of nodes that are highly correlated exclusively between them (albeit not identical). Capitalizing on this correlation, the project develops algorithms for obtaining a shared key under various adversarial models. This addresses a critical problem area as the broadcast nature of a wireless link provides a natural eavesdropping and intervention capability to an adversary and thus it follows that securing a wireless link is essential to the security of any wireless network. This project undertakes the design, implementation, analysis and testing of several protocols and clarifies the relationship between the channel characteristics and the cryptographic protocols that use them. It also investigates the trade-offs between computational and communication costs, and quantifies the level of security achieved by the key generation and renewal protocols assuming various adversarial models. The broader impact of the project is in the potential of making wireless communications more secure while avoiding costs (both computational and organizational) of cryptography. Given the increased wireless internetworking and the multitude of wireless enabled devices presently carried by an average user, the security of the wireless link becomes one of the most critical aspects of securing computer systems. The reduction in the computation also will translate to a reduction in energy use, important to battery or alternative powered systems. Educationally, the project will present an attractive combination of aspects of signal processing, computer security, algorithms and cryptography and will offer a wealth of problems for student engagement
