Digitalization and exponential growth of data are challenges for future IT-Systems. They need to leverage accelerators of different kinds, such as GPUs, FPGAs as well as special-purpose units to meet these new requirements. This in turn will lead to more heterogeneous systems and the interconnection of these different units becomes a field of innovation. However, the integration into the system as well as programming models still offers room for improvement. Cache coherent interconnects like OpenCAPI, CCIX and CXL are examples of recent developments. We will give you an introduction to OpenCAPI and show new possible computing models.
New system architectures like edge computing place computing capabilities closer to the data-producing facilities. Open questions like the best placement of computing tasks, dependability, and others remain. Workloads will nevertheless use the cloud, but there will be data or other workloads which won't be moved to a third-party provider. Thus the interconnection of the private cloud and the public need to address those challenges. With microservices, there is a way to structure and deploy applications on these computing infrastructures.
Completely different computing paradigms, such as quantum computing become available. It becomes clear, that it will accelerate certain algorithms and requires software to be compatible with this new way of computing. There are still technical challenges to solve, but quantum computers will soon be ready to make a significant impact. IBM experts leading the field will present to you the current status of this exciting new technology.
