Data management architectures are key elements to support the improvement of the availability and the maintainability of fleets of transportation systems such as trains, cars, planes and boats during their use. In this context, this chapter proposes the foundations of a specific data management architecture, named “Surfer”. The design of this architecture follows a set of specifications, named “the Surfer way” that translates an original way to consider the issue of big data, aiming to transform raw data into high level knowledge usable by engineers and managers. An application of the Surfer architecture to train transportation is presented. First results are encouraging, the train constructor expects a gain up to 2% of the availability of a fleet of trains.