On 25 May 2018 the GDPR entered into force, a general regulation for data protection at European level. The aim of these rules is to put in place the necessary procedures to protect personal data. It targets public bodies and all private companies that process a large volume of data (storage, transmission, classification). The Internet of Things is one of the most important sources of personal data (medical wristband, connected watch, smartphone, etc.). According to current forecasts, in 2020 there will be 50 billion connected objects, 25 % of cyber attacks will target these connected objects and 50 % of the security of these connected objects will depend on that of the cloud.In this context, we want to explore through this RETRO project the possibilities of setting up an autonomic fog/cloud computing architecture to ensure the security of the data generated as part of an IoT network for E-Health. This architecture should be efficient (time and resources allocated) and low energy consumption (especially for objects). Retro aims to expand the work carried out under the safety axis of the ERA team. It will use the autonomic approach, the core of the team, to propose an intelligent fog/cloud-based solution that would ensure data security in an IoT network for eHealth applications