The OpenStack control plane
When we started designing our first iteration layout, we encountered a variety of OpenStack services, each composed of several components. One or several components can be designed to run in separate physical machines assigned with roles for logical operations and management reasons. As highlighted in Chapter 1, Revisiting OpenStack – Design Considerations, separating the roles logically will support the orchestration of the OpenStack deployment and operation. The term control plane is widely used in the network glossary. It dictates how to process and run requests. As an abstract concept, the control plane can be formed of one or many system components that centrally manage and describe the way a request can be fulfilled. When referring to the OpenStack logical layout, a cloud controller is designated to run most of the pieces of the control plane.
Running the control plane – cloud controller
The OpenStack APIs, scheduler services (including...