#41: Add key wrap to the "use" member in key containers

Comment (by [email protected]):

 Recap of a conversation that I had with John @ Berlin

 It would be possible to avoid defining a new keywrap key usage if we allow
 the enc key usage to have two different meanings with the meaning being
 selected by the algorithm.

 enc could be one of either encrypt data or encrypt key but would never be
 both.  This would imply something that we probably want to be true, that
 you cannot use an RSA key to encrypt both data and keys, but have not
 currently explicitly stated.  We have gone in this direction with the
 assignment of GCM algorithms to be either data encryption or a key
 encryption algorithm.

 This approach would work, however there are three issues to be considered:
 1.  It would need to be explicitly documented
 2.  This would not match the way that the WebCrypto group is thinking of
 things
 3.  It is harder on novices to understand the difference between
 encrypting data and encrypting keys so that being explicit on the string
 would make the Joes of the world happier.

-- 
-------------------------+-------------------------------------------------
 Reporter:               |       Owner:  draft-ietf-jose-json-web-
  [email protected] |  [email protected]
     Type:  defect       |      Status:  new
 Priority:  major        |   Milestone:
Component:  json-web-    |     Version:
  key                    |  Resolution:
 Severity:  -            |
 Keywords:               |
-------------------------+-------------------------------------------------

Ticket URL: <https://grenache.tools.ietf.org/wg/jose/trac/ticket/41#comment:2>
jose <http://tools.ietf.org/jose/>

_______________________________________________
jose mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/jose

Reply via email to