A Mobile Platform for Managing Mobile MapReduce Participatory Sensing Data

The proliferation of powerful, programmable mobile devices along with the availability of wide-area connectivity has created opportunities for participatory sensing applications where users sense and share location, motion and acoustic data in a network of mobile devices. This paper presents a mobile system, based on the MapReduce framework that is designed to manage a large number of mobile users and handle overloads. Our system uses a Voronoi-based algorithm to establish a distributed grid of servers responsible for providing Reduce tasks and final data fusion. A location- and energy-aware scheduler assigns Reduce tasks on mobile phones based on their computational capabilities and geographical position. Our experimental results, using our intelligent transportation service, illustrate the benefits of our approach.