Nowadays we are witnessing a significant increase in the amount of IoT devices and the volume of data generated by these devices. Associated with this, more IoT applications have ultra low latency and response time requirements, such that cloud-centric IoT platforms are unable to address these requirements and to adequately handle the increasing volume of data. To support these new applications and handle the increasing volume of IoT data, Fog Computing platforms have been proposed. The Fog Computing paradigm is characterized by a horizontal, system-level architecture where devices that are in the path of IoT data, from its origin in IoT devices to the cloud, are used for processing distribution, storage, and data control. Fog Computing platforms aim to abstract the highly dynamic and heterogeneous environment where they are inserted, in order to facilitate the development of applications by users. This chapter has the following objectives: (i) to present the Fog Computing paradigm; (ii) to survey and discuss the main requirements that guide the development of Fog Computing platforms; (iii) to discuss two proposals of reference architecture for Fog Computing; (iv) to describe the main Fog Computing platforms and contrast them with the requirements previously raised; (v) to demonstrate in a practical way the use of a Fog Computing platform for the context of smart cities. Resumo Atualmente presenciamos um aumento significativo na quantidade de dispositivos IoT e no volume de dados gerados por estes dispositivos. Associado a isto, cada vez mais aplicações IoT possuem requisitos de latência e tempo de resposta ultra baixo, de tal forma que as plataformas IoT centradas na nuvem não conseguem endereçar estes requisitos e tão pouco tratar de maneira adequada o volume crescente de dados. Para suportar estas novas aplicações e tratar o volume crescente de dados IoT, plataformas que empregam