The Internet of Things (IoT) is generally defined as a global infrastructure that enables advanced services by interconnecting physical and virtual things based on interoperable network technologies. It provides access to a wide range of applications, such as smart cities, smart agriculture, and asset tracking. The increase in the number and diversity of network technologies brings new challenges to IoT architects, who are in charge of the design and deployment of IoT solutions. Indeed, the selection and configuration of network technologies are crucial to ensure the good behavior of IoT applications. In this thesis, we address this issue through the development of a set of methods and associated tools that help IoT architects in their process of designing and deploying IoT solutions. First, we propose a generic framework for the performance evaluation of an IoT network technology given a specific application context. Second, we introduce a newalgorithm relying on multi-criteria optimization methods that enables the automatic selection of the most adapted network technology for a given application context. Then, we propose a solution for optimizing the configuration of any IoT network technology using simulation models and regression methods. For the sake of reliability, we also explore the possibility of calibrating simulation models with data collected from real deployments, which may find a potential application on network digital twins. Finally, we implemented these methods in a no-code tool, integrated in the Stackilab platform, relying onsimulation and aiming to help IoT architects easily future-proof an IoT solution.