Abstract
Widespread deployments of spatially distributed sensors are continuously generating data that require advanced analytical processing and interpretation by machines. Devising machine-interpretable descriptions of sensor data is a key issue in building a semantic stream processing engine. This paper proposes a semantic sensor stream processing pipeline using Apache Kafka to publish and subscribe semantic data streams in a scalable way. We use the Kafka Consumer API to annotate the sensor data using the Semantic Sensor Network ontology, then store the annotated output in an RDF triplestore for further reasoning or semantic integration with legacy information systems. We follow a Design Science approach addressing a Smart Airport scenario with geolocated audio sensors to evaluate the viability of the proposed pipeline under various Kafka-based configurations. Our experimental evaluations show that the multi-broker Kafka cluster setup supports read scalability thus facilitating the parallelization of the semantic enrichment of the sensor data.
Paper Type
Event
Marrying Big Data with Smart Data in Sensor Stream Processing
Widespread deployments of spatially distributed sensors are continuously generating data that require advanced analytical processing and interpretation by machines. Devising machine-interpretable descriptions of sensor data is a key issue in building a semantic stream processing engine. This paper proposes a semantic sensor stream processing pipeline using Apache Kafka to publish and subscribe semantic data streams in a scalable way. We use the Kafka Consumer API to annotate the sensor data using the Semantic Sensor Network ontology, then store the annotated output in an RDF triplestore for further reasoning or semantic integration with legacy information systems. We follow a Design Science approach addressing a Smart Airport scenario with geolocated audio sensors to evaluate the viability of the proposed pipeline under various Kafka-based configurations. Our experimental evaluations show that the multi-broker Kafka cluster setup supports read scalability thus facilitating the parallelization of the semantic enrichment of the sensor data.
Recommended Citation
Zalhan, P.-G., Silaghi, G. C. & Buchmann, R. A. (2019). Marrying Big Data with Smart Data in Sensor Stream Processing. In A. Siarheyeva, C. Barry, M. Lang, H. Linger, & C. Schneider (Eds.), Information Systems Development: Information Systems Beyond 2020 (ISD2019 Proceedings). Toulon, France: ISEN Yncréa Méditerranée.