Our wireless usage characterization includes the two largest-ever studies of a wireless LAN in operation, examining the activity of thousands of users over several months (or years, in some papers) on over 500 access points on a complete campus-wide network. The methodology (and software) we developed for our work has been re-used elsewhere to gather and analyze data in other settings, including other campuses and corporate offices. The data we developed for our work is in use at 44 other university and corporate research labs. The result is that the wireless research community is much more deeply grounded with an understanding of the way that real users and real networks behave. Furthermore, the data and analysis should allow network providers and vendors to improve the quality of the products used by wireless computer users everywhere.
Similarly, our work on mobility prediction is unusual in its use of real data, so we were able to quantitatively compare a wide range of prediction algorithms for the first time. This work has applications in wireless network management, particularly for VoWLAN, and should allow network providers to improve service quality for voice customers.
Our survey paper on wireless network authentication protocols is the only comprehensive study and comparison of these protocols; we believe the community benefits from a careful description and cross-comparison as they develop new protocols or attempt to understand the value of each protocol.
Finally, our work in ad hoc networks includes the largest-ever outdoor experiment using real implementations of ad hoc routing algorithms and a real wireless network. We were thus able to quantitatively compare routing algorithms that are usually compared only in simulations, and we were able to compare our experimental results to those found in simulation. The results include a broad series of recommendations to the community, which desperately needs better simulators, better models, and more careful experimental methods.