Target Locking Implementations should be configurable

This is not a high-priority task — based on discussions with SON-OOF Team
In the current policy implementation “Target Locking” is implemented by identifying targets as a single string — e.g. VNF ID, PNF ID, Service, etc. However in practice, targets are a complex structured entities, and actions can be taken on targets at various sub-elements of a target. For example — an action taken on a cell of an eNodeB need not result into the whole eNodeB to be target locked. As a policy designer, I want to ensure that targets need to be represented with a model that identifies the various subparts of a given target or a collection of entities (such as a group of eNodeBs). Subsequently handle policy actions on sub-elements within that target or collections of these targets. For this we need the following:

• A model for defining targets
• Ability to identify subparts of a target or collections of targets
• Ability to implement target locking mechanisms over such sub-parts of a target or collection of targets.

For the SON-OOF use case, however, this is a low priority compared to queuing, and current target granularity at the PNF level is sufficient.