The final review of the ATMOSPHERE H2020 European project in collaboration with Brazil was held in Campinas (Sao Paolo), last December 12, 2019. ATMOSPHERE designed and developed a software framework to provide trustworthy large-scale data processing services on top of a federation of cloud providers. The framework architecture consists of several layers that provide trustworthiness on various dimensions, such as data confidentiality, integrity and privacy, offering also service dependability and performance guarantees. The DEIB group participating to ATMOSPHERE includes Prof. Danilo Ardagna, Prof. Giacomo Verticale, Prof. Marco Lattuada, Ing. Andrea Marchini, and Prof. Antonio Capone. The group designed solutions for deploying reliable virtual networks interconnecting multiple federated data centers over a wide areaand implemented machine-learning-based performance models supporting the capacity planning of Big Data applications and the performance prediction and scheduling of Deep Learning applications. The research results are available as open source in the official project web site (https://www.atmosphere-eubrazil.eu) and on GitHub (https://github.com/eubr-atmosphere)