Configuration and customization choices arise due to the heterogeneous and scalable aspect of the cloud computing paradigm. To avoid being restricted to a given cloud and ensure application requirements, using several clouds to deploy a multi-cloud configuration is recommended but introduces several challenges due to the amount of providers and their intrinsic variability. In this paper, we present a model-driven approach based on Feature Models originating from Software Product Lines to handle cloud variability and then manage and create cloud configurations. We combine it with ontologies, used to model the various semantics of cloud systems. The approach takes into consideration application technical requirements as well as non-functional ones to provide a set of valid cloud or multi-cloud configurations and is implemented in a framework named Saloon.