Challenges in data intensive science at synchrotron based 3D x-ray imaging facilities

Advances in detector technology for x-ray imaging allow to obtain micrometer-resolution transmission images of specimens as large as a few millimeters at unprecedented frame rates. For a high x-ray flux density synchrotron facility like the Advanced Photon Source (APS), the detector exposure time ranges from hundreds of milliseconds to 150 picoseconds, and synchronization of the camera with the rotation stage allows a full 3D dataset to be acquired in less than one second. The micro tomography systems available at the x-ray imaging beamlines of the APS are routinely used in material science applications where high-resolution and fast 3D imaging are instrumental in extracting in situ four-dimensional dynamic information. In this paper we will describe the computational challenges associated with two scientific problems involving the study of deformation of SiC-particle-reinforced Al alloy matrix composites, and the formation of secondary porosity in natural rock samples. Both examples are in situ studies generating hundreds of TB of data per sample, requiring multi scale and multi modality measurement integration together with a robust computing infrastructure for data management, distribution and processing.