Document: draft-arkko-pppext-eap-aka-14.txt
Review: Joel M. Halpern

 This document is ready for publication as an Information RFC.

  In section 4.1.1.1 there is a description of an IMSI. It
 describes various fields and their sizes. It then says "In other
 words, the IMSI is a string of not more than 15 digits." While this
 is a true statement, it is an addition to the previous wording (which
 would otherwise lead the reader to determine that an IMSI could be 16
 digits.) So the use of "In other words" is misleading.

  With regard to 4.1.1.7 apparently pseudonym user names do not
 include a realm, but fast re-authentication names do. It would be
 helpful if this were mentioned in 4.1.1.7, with some indication of
 the reason for the difference. (If the reason is complex, just a
 mention of the difference would still help.)

  Shouldn't the literal 16384 in 4.1.7 (second paragraph) be a named
 notification code? (There are multiple occurrences of this literal
 and 32768.)

  I presume that the last paragraph of 6.2 relates to existing EAP
 mechanisms, and that it is those mechanisms which mandate the
 response which needs to be ignored and the EAP-Success? Presumably
 due to a lack of EAP knowledge, I could not parse this paragraph.

  Is the notation of section 9.1 for indicating attribute presence
 in messages (1, 0-1, 0*, 0) used in other EAP specifications? It is
 understandable, but different from what I have seen used for other
 protocols.

 INT Extensible Authentication Protocol Method for 3rd Generation Authentication and Key Agreement (EAP-AKA) (Informational) - 1 of 4
  draft-arkko-pppext-eap-aka-13.txt
  [reviewers note: found draft 14, as cited in the note.]

 Token: Thomas Narten
 Review: Joel Halpern