This webinar is divided into 2 sessions:

• First, FORTH-ICS will discuss HPC-Big Data convergence on Kubernetes: This part of the webinar outlines how we run MPI-based applications using Slurm directly from a web-based notebook. In EVOLVE, we deploy on-demand, per-user “virtual clusters”: virtual, containerized HPC execution environments that allow MPI workloads to run seamlessly in Kubernetes, alongside other applications and services. Virtual clusters come preinstalled with all necessary software (compilers, OpenMPI libraries, Slurm, and several MPI-based frameworks), while also exploiting the high-speed, low-latency InfiniBand network available at physical nodes.
• Secondly, NEUROCOM Luxembourg will address “Distributed Stream Processing with Kafka and Spark Streaming”: This webinar highlights the ways and advantages of a real-time stream processing application combining state-of-the-art platforms such as Apache Kafka and Apache Spark (using Spark Structured Streaming). We showcase a. the distributing processing resources of the EVOLVE platform by combining EVOLVE micro-services across workflow stages in the data pipeline and b. the EVOLVE dashboard facility for flexible deployment.

More information here