I have reviewed draft-ietf-mpls-sfl-control version 4 on behalf of the ops directorate.

This draft is ready to proceed.

one note on the privacy considerations section

   This privacy threat may
   be mitigated by encrypting the control protocol packets, regularly
   changing the synonymous labels and by concurrently using a number of
   such labels.

So sure, you could do the suggestions beyond first one (securing communications between devices in a command admin domain if appropriate to circumstnaces seems prudent), it doesn't seem realistic or likely that operators would go out of their way to do more than the minimum amount of signaling merely for the purpose of obfuscation, that's kinda of like adding noise to your IGP on theory that it's harder to parse out the topology as a result. LSPs are similarly privacy identifying in terms of describing a path between two end points but you wouldn't create more of them as a result.