The emergence of new architectures like the cloud open new challenges for data management. It is no longer pertinent to reason with respect a to set of computing, storage and memory resources, instead it is necessary to conceive algorithms and processes considering an unlimited set of resources usable via a "pay as U go model", energy consumption or services reputation and provenance models. Instead of designing processes and algorithms considering as threshold the resources availability, the cloud imposes to take into consideration the economic cost of the processes vs. resources use, results presentation through access subscription, and the parallel exploitation of available resources. Our research contributes to the construction of service based data management systems. The objective is to design data management services guided by SLA contracts. We proposed methodologies, algorithms and tools for querying, deploying and executing service coordinations for programming data management functions. These functions, must respect QoS properties (security, reliability, fault tolerance, dynamic evolution and adaptability) and behaviour properties (e.g., transactional execution) adapted to application requirements. Our work proposes models and mechanisms for adding these properties to new service based data management functions.