A Scalable Sensor Application Framework Based on Hierarchical Load-balancing Architecture

ABSTRACT Many integrated services with sensors and its appliances become common in our daily life. In most of those services, the application and sensors are tightly-coupled. This causes the implementation of the application becomes more complicated. Moreover, network load between the application and the sensor and processing load to evaluate the sensor data increase severely. As the number of sensors to connect increases, the problems become more serious. In this paper, we propose a scalable sensor application framework. Using a standardized API, applications can access any sensors without implementing any sensor-specific procedure. Furthermore, the application delegates a part of evaluation process of the trigger conditions to the sensor. Due to the delegation, network load is minimized because the application communicates with the sensor only when the status of the registered condition is changed. Using these methods, we evaluate this framework qualitatively using sequence diagrams of sensor applications.