MDD4IoT: Model-Driven Development Framework Proposal for Internet of Things Rafael Figueira Goncalves 1 , Andr´ e Menolli 2 1 Universidade Estadual de Londrina - UEL http://uel.br/ 2 Universidade Estadual do Norte do Paran´ a - UENP https://uenp.edu.br/ rafael.goncalves@uel.br, menolli@uenp.edu.br Abstract. The creation of Internet of Things (IoT) solutions is fairly complex, having to integrate and communicate between sensors, devices, and larger sys- tems. Specifically, Cyber-Physical devices often lack software quality, resulting in hard-to-maintain solutions, presenting deficiencies in important aspects of IoT, such as portability and security. Software quality is of great importance to a solution’s long-term success, this work describes the proposal of a framework that aims to aid in the developing of these solutions, applying a Model-Driven inspired methodology. Resumo. A criac ¸˜ ao de soluc ¸˜ oes para Internet das Coisas (IoT) se trata de um processo complexo, integrando e conectando diversos sensores e dispositivos com sistemas maiores. Especificamente dispositivos Cyber-F´ ısicos frequente- mente apresentam deficiˆ encias em aspectos importantes da IoT, como porta- bilidade e seguranc ¸a. Qualidade de software ´ e de grande importˆ ancia para o sucesso desta soluc ¸˜ oes, este trabalho descreve a proposta de um framework que tem como objetivo auxiliar no desenvolvimento destas soluc ¸˜ oes, aplicando uma metodologia Dirigida a Modelo. 1. Introduction and Problem Characterization The percentage of internet-connected individuals grew 1125% in 19 years [InternetWorldStats 2019], and according to [Howell 2017], the number of con- nected devices (not considering personal computers and smartphones) is going to reach 125 billion by 2030 (from 27 billion in 2017). Internet of Things (IoT) researches and implementations have grown exponentially, with the objective of bettering personal comfort in everyday lives with the use of an ever-growing number and types of connected devices. Software development for these new solutions are not easily approachable, they are intrinsically complexes when compared to better established paradigms such as desk- top or mobile, having higher requirements for availability, speed, as well as other chal- lenges. For example, the low computational power in the node devices, the high vul- nerability for security issues [Alaba et al. 2017]. The fundamental need for integrating with other services, and a previously unmatched concern with energy consumption only present more challenges to the systems.