[Engl. Session] Many new applications must be able to handle a high transactional kind of workload, while at the same time analytics should be possible on these data without transferring these data first in a data warehouse. This presentation shows how this problem can be solved with a new technology based on Spark.