So far the terminology are defined by service vm and device manager independently. Converged terminology is needed. At first hash out those terminology and find common one.
NOTE: same terminology can be used for different meanings
- Device template
A template (i.e., a set of defining properties) from which device instances can be created.
- Hosting device (or possibly just Device)
A device instance capable of hosting service instances. These instances can be virtual (e.g., a VM) or physical(?).
- Service type
- Service instance
A logical resource instance of some Openstack (or Openstack-based) service. [Should be aligned with at least Neutron's service instance definition]
- Device management subsystem driver
Driver for system or service used by Tacker to manage (e.g., CRUD) a device. Examples of such services are Nova and Heat.
- Device driver
- Management driver
[Bob: Not sure what this driver should do. Services that use a hosting device for sure needs a driver to instantiate and manage their service instances but I don't think Tacker needs to be aware of those drivers, so what remains?]
- Management address (replace with management URI or similar?)
- Management service driver
[Bob: I'm inclined to say service driver is out of scope for Tacker. We should strive to be service agnostic]
- Management service address
[Bob: Do we have any examples where management address of service differs from management address of device? Unless we do, I suggest we remove this]
- Management call
- Management service call
- Service device binding
- RPC proxy agent
- Proxy management port
- Proxy service port