As I mentioned the exchanged token could include a number of policies, like rate limits. But I expect there could be more sophisticated policies as well. The s

As I mentioned the exchanged token could include a number of policies, like rate limits. But I expect there could be more sophisticated policies as well. The service could send ban or lockout requests to the eId authority so that a misbehaving real life user could be locked out from the service even though the service doesn't know who they are (irl). I would guess it could even be designed so that the authority doesn't know which services a given user has been banned from either. And all the service would need to know is "This user has violated policy X at <timestamp>".