The present document identifies deficiencies of the NRF APIs as specified in TS 29.510
and studies potential optimization solutions. The document will focus on the following aspects:
avoid data storage overhead in the NRF due to many NFs (especially NFs from the same NF set) registering with identical information,
avoid signalling overhead (both in terms of payload size and number of HTTP requests) due to discovery responses containing lots of duplicate information and due to the number of data change notification sent when duplicated data are changed one by one.
avoid processing overhead (both at NRF and its consumers) resulting e.g. from signalling overhead.