Opportunistic wireless network architectures

Today, significant portions of the radio spectrum are under-utilized. Hence, this dissertation develops the principles and techniques necessary to build opportunistic wireless networks, which work by continually seeking and using portions of the radio spectrum currently unused by the original licensees (incumbents), thus enhancing efficiency of spectrum use and network capacity. The primary challenge with this approach is to utilize unused spectrum efficiently without interfering with incumbents. A prominent emerging instance of opportunistic wireless networking are the so-called television white spaces. These are channels that, over a small window time, are not used by the incumbents: television stations or wireless microphones. Using these white spaces, this dissertation makes the following four contributions. WhiteFi. The first data communications network to function over the white spaces. WhiteFi characterizes the white spaces spectrum and consequently, proposes MCham—a new spectrum assignment metric and algorithm that utilizes variable channel widths and SIFT—a general purpose technique to extract hints from the physical layer to optimize access point discovery. SenseLess. The key challenge here is in ensuring safe operation of incumbents without explicit spectrum sensing while also efficiently utilizing the white spaces. We avoid sensing by relying on a combination of radio signal propagation models, up-to-date database of incumbents, and efficient mechanisms to propagate updates on incumbent status to all nodes. Comparisons to ground truth measurements reveal SenseLess does not interfere with incumbents while also efficiently extracting up to 85% of white spaces. Dyson. We extend the SenseLess architecture to enable a programmable wireless network that permits site-specific customization of wireless network deployments. Using a central controller, network administrators can customize the behaviour of the wireless network by implementing simple python based policies. DenseAP. We leverage the Dyson architecture to opportunistically exploit client density via jointly managing client-AP associations and channel assignments to increase the network throughput. Experimental evaluations of DenseAP reveal an increase of up to 800% in the total network throughput. We evaluate these systems using simulations, implementations, and measurements, and demonstrate their ability in improving wireless network efficiency and throughput using opportunistic spectrum access.