Using generally available packages - in the form of container image - from an official source or a certified provider comes with a big caveat in regards to performance-sensitive workloads. These packages may provide ABI compatibility, but they are not optimized for our specialized hardware, such as GPUs, high-performance NICs, or CPU chip architecture. The best way to address this issue is to compile packages (build your images) on your own deployment.
Orchestrating resource-specific applications or performance-sensitive applications requires real-time knowledge of the available resource inventory and dynamic hardware and software feature discovery on the cluster. Knowing where to put the application to fit its needs or enhance its behavior is becoming increasingly important among customers.
This lightning talk will feature a short update on the status of the Node-feature-discovery Operator. It will also include an update on the GO package ArchSpec-feature discovery, which will enable GO-based Operator to query NFD generated Kubernetes-node labels to better allocate resources on a system.