Towards Quality Guided Data Integration on Multi-cloud Settings

This PhD project addresses data integration considering data quality (freshness, provenance, cost, availability) properties in a multi-cloud context. In fact, in a multi-cloud context, data is made available through a huge offer of services deployed on different clouds with heterogeneous quality of service features. By users who thank to their contracts with the clouds expressed by traditional SLA according to their rights. Consequently, data integration in this context needs to take into account these new constraints. The aim of our work is to revisit previously proposed data integration solutions in order to adapt them to the multi-cloud context. Our solution consists in defining over the clouds a layer that provides a reasoning on the best services combination that meets services and user constraints and willings, the best way to deploy the integration process. This layer should let further data integration easier thank to the definition of a new kind of SLA called Integration SLA. This paper gives a model-oriented vision of our proposal.