Internet of Things (IoT) applications rely on networks composed of set of heterogeneous sensors and smart devices, which have the capability to constantly, observe the surroundings and gather data. This heterogeneity is reflected on raw data gathered by such type of systems. Consequently, the task of high-level IoT applications to interpret such data and detect events in the real world is more complex. Moreover, data heterogeneity leads to the lack of interoperability between IoT applications. Semantic Web (SW) technologies have been widely adopted to model and integrate data from different sources on the web; extending them to the IoT domain can be used to mitigate the aforementioned challenges. Semantically annotating IoT data is a fundamental step toward developing smarter and interoperable IoT applications. However, this type of process requires a large amount of computing resources, especially in scenarios where a large number of sensors is expected to be involved such as smart city. To address these challenges, we propose a lightweight semantic annotation approach that can be implemented on resource-constrained IoT gateways connected to a limited number of sensors. To evaluate the feasibility of the proposed approach, we have carried out a set of experimentations using a middleware prototype implementation. Several benchmarks are considered such as: Data size, response time, and resource utilization.
