The improvement of the teaching and learning process today is possible through the integration of day-to-day technologies. Among the existing technological methods for learning, there are those focused on content, like educational games and virtual environments, and those centered on the student and the environment, like adaptive, ubiquitous and context-aware environments. In this last scenario, you will find smart classrooms smart, which must have the support of a Fog Computing infrastructure to operate effectively. The objective of this work is to demonstrate methods of identification, prevention, and failure treatment in this of cloud computing infrastructure, to achieve greater efficiency of the ubiquitous and adaptive environment of Smart Classroom. The proposal is not only limited to this application, but is applicable regardless of the type of Internet of Things (IoT) environment implemented. To verify these methods effectiveness, a Smart Classroom is simulated on the iFogSim platform. The results obtained regarding the level of availability and operability of the components were promising, evidencing proper levels of methods applicability.