SemSub: Semantic Subscriptions for the MQTT Protocol

Abstract

The Internet of Things is constantly reaching new industrial and social domains, subsequently, encompassing virtually all domains. The number of connected devices continues to rise exponentially. Deployments with particularly high and extremely heterogeneous data volumes are particularly characteristic of several IoT domains such as the Smart City domain. Efforts to create homogeneous and universally valid semantic models have so far achieved valuable approaches to maintaining the order required for automation and interoperability. Semantically descriptive and classifying solutions are currently limited to IoT devices and their management. The established IoT protocol Message Queuing Telemetry Transport (MQTT) benefits very little from these semantic-driven efforts. On the one hand, there is only a simplistic approach to address semantics for the MQTT protocol, and on the other hand, there are insufficient efforts to adopt the available semantic-driven technologies, which primarily focus on the user in his role as publisher. In this work, we propose SemSub, a prototypical concept that primarily supports subscribers with suggestions to get semantically suitable information from a MQTT topic collection. Using simulations for smart cities, we show that the hard syntactic coupling between information and topics in the standardized MQTT can be efficiently addressed through the proposed stepwise provision of four progressive semantic levels regarding the subscription.