Possibilities and Challenges for Reconfigurable Hardware and Cloud Architectures in Data-Intensive Scientific Applications

Advances in process technology and new design tools have expanded the scope of embedded systems. This ranges from the implementation in several chips on board to module groups in integrated circuits. Reconfigurable hardware and, in particular, FPGAs are used more frequently in scientific applications, where they enable the development of complex and intelligent field devices. Furthermore, this increased the use of a platform-based design approach that facilitates the development and verification of complex FPGAs through the full reuse of hardware and software modules. Especially in the area of heterogeneous accelerators, which can improve the exploitation of modern data centers. Another critical aspect in the evolution of embedded systems is the trend towards networking embedded nodes using specialized computational and networking technologies called cloud computing. In this paper, we will present our data-intensive experiments at the terahertz source at ELBE accelerator-based light source, and its integration into a reproducible data management workflow at the heterogeneous cluster in our data centre