IoT technologies have significantly changed the way we interact with the environments we live in, ensuring higher levels of comfort, security and efficiency. However, given the complexity and diversity of these systems, of the environment where they are applied and of the ever-diversifying requirements from stakeholders, there are significant challenges that arise w.r.t. the design, management and interoperability of IoT based home systems. In this paper, we propose a domain-specific modeling method tailored to oversee the integration and communication logic of the smart home systems. The language is deployed in a tool demonstrator developed using the ADOxx meta modeling platform. The models allow the representation of smart home devices, their interactions and how daily routines can be automatized. In addition, these models are exported in RDF format following the Linked Open Models principle enabled by ADOxx. By leveraging the strengths of both visual conceptual modeling and semantic graphs, our knowledge-driven approach aims to provide a Design Science treatment for managing the complexity of smart home systems.