Date of Award

9-4-2015

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Advisor

Kenji Yoshigoe

Abstract

Unattended wireless sensor networks (UWSNs) can assist a wide range of monitoring applications in hostile environment (e.g., battlefields, uncultivated fields, volcanic zones, underwater areas, etc.). However, they cannot offload captured or generated data to a sink in real time, and preservation of data at sensor nodes is important. In this work, we accomplish the availability and confidentiality of data in UWSNs. In case of failure or compromise of sensor nodes which can result in unavailability or loss of data, we propose an on-demand data replication method that can dynamically change the extent of replication to improve data survivability with minimal storage requirement. Analytical results show significantly improved data survivability with significantly less storage consumption than solutions available in literature. After thorough investigation of a well-known Rivest Cipher 4 (RC4), we improve data confidentiality in networked systems. In particular, the RC4 cipher shows some weaknesses including a correlation problem between the public known outputs of the internal state. We propose a series of RC4 derivatives to deplete the correlation problem as well as, in one case, to accelerate the key generation rate of the RC4. The recent finding suggests that the short-term and long-term known biases for RC4 can be practically exploited to capture any part of the Internet traffic relying on Transportation Layer Security (TLS) with RC4 cipher option. Similar exploitation could impact systems relying on any RC4-based cipher. We first empirically validate the feasibility of the attack. We then reconfigure the RC4’s pseudo-random number generator (PRNG) to entirely remove both of the well-known short-term and long-term biases. More importantly, the new RC4 implementation accomplishes this without adding complexity to the well-known RC4 to meet the practical cryptographic requirement for resource-constrained devices, which are increasingly populated as a part of the emerging Internet of Things (IoT) ecosystem.

Download

Share

COinS