On-demand data integration is among the key challenges in Cloud Computing. In this paper, we present an ontology-based framework for describing and integrating data on the fly to answer transient business needs. We provide a semantic modeling for cloud’s data services. The proposed modeling makes it possible to automatically resolve the different types of data heterogeneity that would arise when data from heterogeneous and autonomous providers need to be combined together to answer the business’s data needs. We validate our approach with a prototype. The main contribution of this paper is an efficient on-demand integration system for the clouds.