Combining the Concepts of Semantic Data Integration and Edge Computing

Abstract

The Internet of Things (IoT) is growing rapidly. Therefore, there are more and more vendors, which led to IoT being a heterogeneous collection of different IoT platforms, isolated solutions and several protocols. It has been proposed to use Data Integration to overcome this heterogeneity. In addition, costs are on the raise due to increasing volume of data which increases demands on bandwidth and cloud computing capabilities. Again a solution has already been proposed by reducing the amount of data to forward by processing data at the edge of an IoT-System, e. g. filtering or aggregation. This concept is called Edge Computing. In this article the Semantic Edge Computing Runtime (SECR) is introduced, combining both concepts. The application of Data Integration enables Edge Computing to be performed on a higher level of abstraction. In addition, the developed Driver-approach allows SECR’s Data Integration algorithm to be applied to a wide range of data sources without imposing requirements on them. The Data Integration itself is based on technologies of Semantic Web, applying metadata to raw data giving it context for interpretation. Furthermore, SECR’s REST-API enables applications to alternate Data Integration and Edge Computing at runtime. The tests of SECR’s prototype implementation have shown its suitability for deployment on an edge device and its scalability, being able to handle 128 data sources and Edge Computing Tasks.