Optimizing Data Processing Service Compositions Using SLA's

International audienceThis paper proposes an approach for optimally accessing data by coordinating services according to Service Level Agreements (SLA) for answering queries. We assume that services produce spatio-temporal data through Application Programming Interfaces (API's). Services produce data periodically and in batch. Assuming that there is no full-fledged DBMS providing data management functions, query evaluation (continuous, recurrent or batch) is done through reliable service coordinations guided by SLAs. Service coordinations are optimized for reducing economic, energy and time costs