Towards In-network Aggregation for People-Centric Sensing

Abstract

Abstract. Technological advances in the smartphone sector give rise to people-centric sensing that uses the sensing capabilities of mobile devices and the movement of their human carriers to satisfy the ever increasing demand for context information. The quick adoption of such pervasive and mobile services, however, increases the number of contributors, strains the device-to-server connections, and challenges the system's scalability. Strategies that postpone load balancing to fixed infrastructure nodes miss the potential of mobile devices interconnecting to preprocess sensor data. This paper explores opportunistic service composition to coordinate in-network aggregation among autonomous mobile data providers. The composition protocol defers interaction with peers to the latest possible moment to accommodate for the dynamics in the operating environment. In simulations such an approach achieves a higher composition success ratio at similar or less delay and communication effort than an existing conventional composition solution.