iService: A Cloud-based Scheduling Service for Efficient Usage of IoT Resources

With the proliferation of the Internet of Things (IoT), numerous IoT devices have been deployed in various living spaces such as university campuses, community centers, offices and homes. Users of IoT services are more and more willing to explore and directly interact with interested devices over mobile or web applications. However, there is a lack of online service that can efficiently schedule requests issued by various IoT users to distributed IoT devices. Efficient, cost-effective scheduling is critical to the user experience in a smart living space. In this paper, we propose iService, a cloud-based scheduling service for efficient usage of IoT resources given IoT-specific factors, such as device location, sensing capability, and energy cost. In our solution, the essential scheduling problem has been formulated as a mixed integer nonlinear programming (MINLP) problem. To efficiently address this problem in an online fashion, iService has been implemented using multiple AWS technologies. Via extensive experiments, the effectiveness of iService has been demonstrated under different request patterns and system parameters. From these experiments, we conclude that iService can improve the quality of service to the user and thus their quality of life within a pervasive living space.