Transferring Remote Ontologies to the Edge of Internet of Things Systems

Abstract

Edge computing paradigm allows computation to be moved from the central high powered Cloud or data center to the edge of the network. This paradigm often enables more efficient data processing near its source and sends only the data and knowledge that have value over the network. Our study focuses on performing semantic reasoning at the edge computing devices, which requires transferring ontologies to the edge devices. This paper presents different representations for transferring Web Ontology language (OWL) version 2 ontologies to the edge. We evaluate different representations in an experimental IoT system with edge nodes and compare lengths of different syntaxes and their computation effort of building models in Cloud and edge computing devices in terms of processing time.

Distribution of Semantic Reasoning on the Edge of Internet of Things

“…The IoT system should support popular RDF syntaxes such as RDF/XML, JSON-LD, and N3. Moreover, a mechanism is required to access the ontologies and rule sets from Cloud to enable mobile devices to preform various reasoning tasks [51].…”

“…In a basic mode, the edge nodes store an ontology locally, execute semantic reasoning, and send the data to the Cloud server simultaneously. In a second mode, the edge nodes fetch an ontology from a remote ontology repository server [51]. RDF data is stored with Android Jena framework and a rule-based hybrid rule engine is developed with Android Jena.…”