Thanks for the detailed review, comments are in-lined...
From: Roni Even [mailto:ron(_dot_)even(_dot_)tlv(_at_)gmail(_dot_)com]
Sent: Monday, June 04, 2012 2:20 PM
To: draft-ietf-nea-pt-tls(_dot_)all(_at_)tools(_dot_)ietf(_dot_)org
Cc: 'IETF'; gen-art(_at_)ietf(_dot_)org
Subject: GenART LC review of draft-ietf-nea-pt-tls-05
I am the assigned 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 Last Call comments you may
receive.
Document: draft-ietf-nea-pt-tls-05
Reviewer: Roni Even
Review Date:2012-6-4
IETF LC End Date: 2012-6-13
IESG Telechat date:
Summary: This draft is almost ready for publication as a standard track RFC.
Major issues:
Minor issues:
1. In section 3.2 "Therefore, this specification requests the IANA
reserve a TCP port number for use with the PT-TLS protocol upon publication of
this specification as an Internet standard RFC." I think it will be better to
have here the assigned port number and instruct the RFC editor to put the
correct value.
[PS:] Ok, we can reword this in hopes of getting a particular value (race
condition with other upcoming RFCs).
2. In section 3.4.2.2 last paragraph you summarize the text from section
3.8 while in the paragraph above you provide the reference. Why do you need the
last paragraph if 3.8 is referenced.
[PS:] The goal of this section is to introduce and summarize the different
phases of PT-TLS. We felt a brief discussion of the general message flow was
helpful to the reader to understand what occurs during this phase (similar to
what we did in the other sub-sections). Your correct that this information is
covered later in more detail.
3. In various places you refer to SMI 0 as IETF SMI number while
according to the table it is IANA SMI number.
[PS:] I presume this is about the PEN 0 being for the IETF. Correct, it's the
IETF's name space that administered by the IANA. What text would you like to
see to make this more clear? Can we do it in one place, for example stating
that the IETF name space is administered by the IANA?
4. I assume that all implementations MUST support message type vendor ID
0. Is this mentioned?
[PS:] The purpose of this section was just to summarize and enumerate the
message types for vendor id 0. I don't think it's a general rule that any
message type defined in the IETF (IANA :)) name space must (or should be)
supported by all implementations. It will vary depending on the purpose of the
message so that normative language is included in the descriptions of the
message.
5. In section 3.5 and 6.1 you propose a policy of "Expert Review with
Specification Required ". I think that according to RFC5226 expert review is
implied if you select a specification required policy.
[PS:] I agree, it says "Specification Required also implies use of a Designated
Expert". The policy is just "Specification Required" so we could remove the
"Expert Review with" and make it clear it's the Specification Required IANA
policy.
6. In section 3.6 on 9+ "Recipients of messages of type 9 or higher
that do not support the PT-TLS Message Type Vendor ID and PT-TLS Message Type
of a received PT-TLS message MUST respond with a Type Not Supported PT-TLS
error code in a PT-TLS Error message." I think this is true only for Message
Type Vendor ID 0.
[PS:] Thanks will reword this section to make it more clear.
7. In 3.7.1 for Max vers and prefs ver you say that they MUST be set to
1. I think it will be more correct here to say SHOULD since you explain
afterwards that they may have other values.
[PS:] I think this is a MUST. The next sentence just points out that this
normative text might change in a future revision (which is not currently
planned).
8. In section 3.7.2 "the recipient SHOULD send". Why not make it a MUST
here.
[PS:] I ok with making this change, let's see what others think ...
9. In section 3.7.2 "The version selected MUST be within the Min Vers to
Max Vers inclusive range sent in the Version Request Message" I was expecting
to see pref ver here.
[PS:] Perf is just an informational (hint) preference.
10. In section 3.8.3 " The SASL client authentication starts when the NEA
Server enters the PT-TLS Negotiation phase and its policy indicates that an
authentication of the NEA Client is necessary but was not performed during the
TLS handshake protocol " my read of section 3.8 second paragraph is that it
can be done even if was done in the TLS handshake so the last part of the
sentence is not correct, if there is a policy you do it anyhow. This comment is
also for the third paragraph.
[PS:] Thanks, this was supposed to be an example. Will fix these.
11. In section 3.9 I noticed that you propose to send the entire original
message. Isn't it enough to send only the message identifier. This is based on
the last sentence of this section.
[PS:] Not "the entire original message" as its at most the first 1024 bytes of
the offending message. This allows the recipient to either caches recently
sent messages and/or message identifiers when determining what caused the
error. We thought this flexibility was useful and had very little cost.
12. Most of the text in section 6.1 repeats RFC5226 but in your words. Are
you trying to change some of RFC5226 text if not why write it in different
words?
[PS:] We were hoping to emphasize the aspects of 5226 that are most important
to this specification. We weren't trying to change how the IANA policy was
interpreted. Did you think we did so? Is there a portion of this text that is
most troubling or was this just a question?
Nits/editorial comments: