Whatever the size of your deployment, there are always use cases where bursting to external resources can bring benefits: covering for peak workloads can be more cost-effective, using on-demand external resources versus over-provisioning on premises; access to otherwise scarce resources, especially relevant for accelerators like GPUs or FPGAs; access to otherwise non-accessible resources such as TPUs or IPUs. 

This session focuses on how the usage of cloud native projects like Kubernetes, containerd, Prometheus, Thanos, and others can greatly simplify the integration of external resources into an existing on premises service stack. It will cover the aspects of deployment and management of all resources, the challenges in areas like networking and storage, and what the future might look like for the hybrid world.