Developing IoT Spark-Streaming Applications by Means of streamLoader

Nowadays there is a great interest in the development of cross-domain IoT applications that are able to gather an high number of observations taken from sensors belong to heterogeneous IoT platforms and efficiently combine them in flows that can be exploited for conducting analysis. However, their realization poses several concerns. From one side, graphical interfaces are needed for facilitating the specification and composition of the data acquisition operations. From the other side, facilities are required to manage big data streams, to check the application correctness, and to set up and configure clusters of machines. Starting from the graphical facilities of the StreamLoader system for the development of Data Acquisition Plans (DAPs) from sensors belonging to heterogeneous IoT platforms, in this paper we show the approach developed for translating DAPs in a Spark-Streaming application. Experiments are shown for assessing the scalability of the developed applications.