| Wireless systems have become an indispensable part of modern day communications, from living rooms to airplanes. While the latest technological advancements in wireless have given birth to new kinds of devices and access techniques, there is an accompanying increase in the system complexity and sophistication. Hence, measurements have become more crucial than ever to design and evaluate these complex emerging systems. While previous measurement approaches do measurements after fully building a system, we propose that measurements should be the tightly integrated as the building block of system design for emerging wireless systems.;In this dissertation we first describe our work on two areas of emerging wireless systems (1) Vehicular Networks and (2) Dynamic Spectrum Access. In vehicular networks, we describe how to efficiently disseminate data to multiple vehicles on a highway with road-side infrastructure. First, through extensive test-bed based evaluation and large scale simulations we design and build a reliable and scalable content distribution system. Next, we study the infrastructure requirements on a highway that can disseminate data to vehicles at minimum cost. In dynamic spectrum access, we seek to improve the performance of secondary users both with and without explicit cooperation from primary users. First, using fine-grained measurement traces we validate previous assumptions on secondary user performance and design solutions to improve it. Next, when there is explicit cooperation of primary user, we design solutions that can improve secondary access without negatively impacting primary users.;Finally, we discuss our work on AirLab, a publicly accessible global scale distributed infrastructure for wireless measurements. AirLab is an effort to provide a rich corpus of comparable and consistent wireless traces to researchers. By collecting and analyzing these traces, researchers can not only understand current wireless systems but also build better wireless systems in future based on the insights gained. In this dissertation, we discuss our contributions towards the design of measurement framework, experiment scheduler and reliability mechanisms. |
