The durability of concrete is a major challenge for the construction, which devotes one third to one half of its annual investment to building maintenance. The lack of field data regarding concrete durability, especially in the case of exposure to sulfate ions ("sulfate attack") makes it difficult to determine the appropriate test methods and performance criteria. Additionally, the increased use of blended concretes (cement with mineral admixtures) suffers from a lack of experience regarding their long-term performance. Most results for sulfate resistance are derived from accelerated laboratory tests where performance criteria are based only on macroscopic properties, especially expansion. To fill this gap and better understand the mechanisms of sulfate attack under real conditions, a parallel study of laboratory micro-concrete and field concrete samples under sulfate exposure was undertaken, focussing on microstructural changes in addition to the typical macroscopic characterisation. Four exposure regimes were designed in the laboratory: full immersion (traditional test in "ponding"), pH-control, semi-immersion and wet/dry cycles. Pure Portland blends and slag blends with high level of slag replacement (70 wt.-%) were investigated. The exposure regime has been found to play a main role in the damage process. In ponding conditions, the damage process takes place three stages characterised by a first period of induction, followed by surface damage that finally extends to the bulk of the material. Paradoxically, the w/c-ratio does not seem to have much impact on the ionic transport phenomena but might be more decisive in the microstructure mechanical strength against local stresses. The slag blends, considered as sulfate resistant in ponding exposure, revealed bad performance under wet/dry cycles. This behaviour was attributed to poor proper physical resistance of the slag hydrates to the applied drying. Field observations tend to confirm the laboratory results and validate the test settings. It has been underlined that a direct relationship between the damage (e.g.; cracking/expansion) and the phase assemblage was not evident. However, the study highlights that sulfate combination with the hydrates of cement (e.g.; C-S-H) and to those of slag was found to play a role in the initiation of expansion, which would be initiated either by a swelling of the hydrates or by the precipitation of fine ettringite when the saturation level in sulfate of the hydrates has been reached.