This is useful for an upstream system that wants to inexpensively determine which parts of a configuration has changed.
In the simplest implementation of the draft, the root element of a datastore must be updated with some unique transaction ID on each change, which an upstream system can store and use to determine if changes have been made.
In a more complete implementation, each changed node could be annotated with the transaction ID, which must then propagate to all ancestor nodes, so that the exact changes can be identified.
The draft is here.
This is useful for an upstream system that wants to inexpensively determine which parts of a configuration has changed.
In the simplest implementation of the draft, the root element of a datastore must be updated with some unique transaction ID on each change, which an upstream system can store and use to determine if changes have been made.
In a more complete implementation, each changed node could be annotated with the transaction ID, which must then propagate to all ancestor nodes, so that the exact changes can be identified.