Last Call Review of draft-ietf-hokey-erp-aak-
review-ietf-hokey-erp-aak-genart-lc-garcia-2012-02-03-00

I have been selected as the General Area Review Team (Gen-ART)
reviewer for this draft. For background on Gen-ART, please see the FAQ at
<

http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>

Please resolve these comments along with any other comments you may receive.

Document: draft-ietf-hokey-erp-aak-07
Reviewer: Miguel Garcia <miguel.a.garcia at ericsson.com>
Review Date: 2011-01-02
IETF LC End Date: 2012-02-07



Summary: This draft is on the right track but has open issues, described 


in the review.




Major issues:

- None

Minor issues:



- The main problem I have with this draft is the lack of normative text 


(RFC 2119 reserved words) in relevant paragraphs. If interoperability is 


to be granted, an effort should be taken in adding quite a few more 


normative statements.






However, having said that, the section where I find more that there 


should be more normative text, is Section 3, which is an "Overview" 


section. In general, an overview section should use descriptive, but not 


normative text.






For example, take the last paragraph in Page 5 (that continues to Page 


6). One possible change is to make normative the text and move it outside 


a section whose title is "Overview".




   Upon receiving the message, the ERP/AAK server MUST first use the
   keyName indicated in the keyName-NAI to look up the rIK and MUST
   check the integrity and freshness of the message. Then the ERP/AAK
   server MUST verify the identity of the peer by checking the username
   portion of the KeyName-NAI.  If any of the checks fail, the server
   MUST send an early- authentication finish message (EAP-Finish/Re-auth
   with E-flag set) with the Result flag set to '1'.  Next, the server
   MUST authorize the CAP specified in the CAP-Identifier TLV.  In
   success case, the server MUST derive a pMSK from the pRK for each CAP
   carried in the the CAP-Identifier field using the sequence number
   associated with CAP-Identifier as an input to the key derivation.
   (see d. in the figure 1).

   Then the ERP/AAK server MUST transport the pMSK to the authorized CAP
   via AAA Section 7 as described in figure 2 (see e.1,e.2 in the figure
   2). Note that key distribution in the figure 2 is one part of step d.
   in the figure 1.



The the last paragraph in Section 3 also contains an "Optionally", which 


I believe should be replaced with a capitalized "OPTIONAL"




Another instance: towards the end of Section 5.2, the text reads:

   HMAC-SHA256-128 is mandatory to implement and should be enabled in
   the default configuration.

and should probably be:

   HMAC-SHA256-128 is REQUIRED to be implemented and SHOULD be enabled in
   the default configuration.

Similarly, the last paragraph in Section 5.2 reads:

   If the EAP-Initiate/Re-auth packet is not supported by the SAP, it is
   discarded silently.

and should probably be:

   If the EAP-Initiate/Re-auth packet is not supported by the SAP, it
   SHOULD be discarded silently.



- Another topic, Section 9 (IANA Considerations) reads:

   Further, this document registers a Early authentication usage label
   from the "USRK Key Labels" name space with a value:

      EAP Early-Authentication Root Key at ietf.org




I am missing the sentence to name the master registry where the USRK Key 


Labels subregistry is stored. This is the Extended Master Session Key 


(EMSK) Parameters registry (I guess). And probably this comment is also 


valid for the rest of the IANA actions: the main registry is not named, 


and it is hard to find it.





/Miguel
--
Miguel A. Garcia
+34-91-339-3608
Ericsson Spain