Enhancing Coexistence, Quality of Service, and Energy Performance in Dynamic Spectrum Access Networks.

Dynamic Spectrum Access (DSA) is an upcoming wireless technology which aims to alleviate spectrum-usage inefficiency from the current static spectrum allocation model of wireless communication. However, DSA technology is still in its infancy. Despite its enormous potential for improving wireless networking performance and quality of service (QoS), contemporary DSA is highly ineffective in realizing such advantages to consumer-oriented wireless systems and networks. Therefore, existing DSA technology is unsuitable for actual deployment in its current form. The thesis explores this hypothesis in depth and identifies its important causes, which span three key dimensions—coexistence, QoS, and energy. The thesis proposes novel and practical system-oriented solutions in order to address the issues identified with traditional DSA. “Awareness-cum-adaption” is the central theme across the proposed methods. First, the thesis focuses on the generic DSA coexistence problem. It presents a dual-mode DSA operation scheme featuring joint sensing and transmission scheduling for safe and efficient time-domain incumbent-unlicensed coexistence. This solution expands the application of DSA to most of the licensed spectrum, which greatly improves the utility and effectiveness of DSA technology. A prototype, called Spectrum-Conscious WiFi (SpeCWiFi), is also developed. Second, it provides important insights into the QoS impact of DSA through the case study of a consumer DSA-based wireless network. Third, it presents Context-Aware Spectrum Agility (CASA) to address the DSA’s QoS issues from a device-centric perspective. Fourth, to tackle the QoS problems at the network level, it proposes a network service framework called DSA Synchronization (DSASync). DSASync consists of algorithms based on buffering and traffic-shaping to effectively manage end-to-end connections in DSA networks. CASA and DSASync together provide a complete solution to application QoS issues associated with existing DSA, and hence, make DSA a highly effective and performance-enhancing technology for consumer wireless networks. Finally, this thesis explores the energy cost of DSA through an empirical analysis, and proposes the Dynamic Energy Management for DSA (DEMD) scheme to reduce its energy footprint.PhDComputer Science & EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/84615/1/ashwinik_1.pd