Anthony J. Nicholson, Brian D. Noble
2008 Proceedings of the 14th ACM international conference on Mobile computing and networking - MobiCom '08  
As mobile devices continue to shrink, users are no longer merely nomads, but truly mobile, employing devices on the move. At the same time, these users no longer rely on a single managed network, but exploit a wide variety of connectivity options as they spend their day. Together, these trends argue that systems must consider the derivative of connectivitythe changes inherent in movement between separately managed networks, with widely varying capabilities. To manage the derivative of
more » ... vative of connectivity, we exploit the fact that people are creatures of habit; they take similar paths every day. Our system, BreadCrumbs, tracks the movement of the device's owner, and customizes a predictive mobility model for that specific user. Rather than rely on a synthetic model or aggregate observations, this custom-tailored model can be used together with past observations of wireless network capabilities to generate connectivity forecasts. Applications can in turn use these forecasts to plan future network use with confidence. We have built a BreadCrumbs prototype, and evaluated it with several weeks of real-world usage. Our results show that these forecasts are sufficiently accurate, even with as little as one week of training, to provide improved performance with reduced power consumption for several applications.
doi:10.1145/1409944.1409952 dblp:conf/mobicom/NicholsonN08 fatcat:rriur2ldszemxpcej6gkbuzzr4