Cloud Native Architecture on the SKA SDP

The nature of the workloads on the Science Data Processor of the SKA has provided an opportunity to re-imagine what scheduling and resource management could look like for HPC.  Some of the associated research has focused on Containerisation and Container Orchestration including Kubernetes where the principles of abstraction, encapsulation and efficient task packing stand to provide benefits throughout the hardware and software ecosystem.  An overview of the discoveries are discussed in the SDP Memo_ 051 – Cloud Native Applications on the SDP Architecture  .