Data management architectures will change radically in the next few years: The spectrum of applications that have to process large amounts of data has expanded significantly. In addition, new trends are making it possible to rethink current system architectures – both at the hardware and operating system level.
The DFG priority program “Scalable Data Management for Future Hardware” (https://www.dfg-spp2037.de) has set itself precisely this goal: For future database systems, the researchers want to loosen the tight corset implied by assumptions and available hardware and replace it with a more flexible architectural approach. To achieve this vision, ten universities and the German Research Center for Artificial Intelligence are working together for another three years in Phase II of the joint project, which was launched in 2017.
At the TU Darmstadt, Professor Carsten Binnig from the Data Management Group is heading the project “Scalable Data Management on Next-Generation Networks beyond Remote-Direct-Memory-Access (RDMA)”. His team is researching how the programmability of next-generation high-speed network cards and switches can be used to improve the performance of scalable data processing systems.