OpenStack is great for delivering small and large amounts of resources, such as Virual Machines or Containers, via APIs "as a service“. However, when it comes to enrolling and managing physical resources themselves, e.g. Switches or PDUs, or even more generally the entire data center, OpenStack itself reaches its limits. Is the IPAM the source of truth or OpenStack itself? How is the reconciliation done? Who approves changes there? How can anomalies be found?
An intent-based holistic data center management approach allows a high abstraction of the underlying configurations and control and solves the previously listed problems.