ietf
[Top] [All Lists]

Followup on draft-mraihi-inch-thraud-06.txt (was re: Gen-ART Last Call Review of draft-mraihi-inch-thraud-05)

2008-07-03 14:43:33
I reviewed performed a Gen-ART last-call review on draft-mraihi-inch- thraud-05. Since my review, the authors published version 06 of that draft. I'm not sure what the current status of the draft is, so here is an informal followup on my original review.


Summary: This draft is almost ready for publication as a informational RFC. I have (a much smaller set of ) comments that should be considered prior to publication.

General:

This draft is a significant rewrite of the previous version, and is much easier to follow. It no longer claims to specify a protocol, and a lot of text appears to have been removed. I think this is a much better draft, but I am surprised to see this much change post IETF LC. (I'm not suggesting any action here--just pointing it out)

Otherwise, most of my previous comments are addressed, although I have a few minor new ones.

Specific:

Section 1, paragraph 4:

Were the lower case "should" and "may" in this paragraph intended to be normative?

Paragraph 7:

IODEF is referenced as a work in progress draft--is it expected to progress with this draft?

Section 5.2.1:

What does the "xs" in "xs:string" refer to? It _looks_ to me that it is merely an XML namespace prefix (i.e. xmlns:xs="..."), which is fairly ephemeral. That is, any given document might use a different prefix. It seems odd to use it in the text this way. I guess the text is intended for humans that may reasonably be expected to figure it out. In any case, this usage is pervasive through the document (also for "iodef:" and "thraud:").

5.4.1:

Is there any registration or other mechanism to promote consistent usage for new OtherEventTypes? What keeps people from picking the same type for different kinds of new events, or use inconsistent types for otherwise similar new events?

Section 7, definition of Incident.Method.URL:

Can you offer any guidance on what sort of URI is appropriate here?



Additionally, IDNITS reports the following:

  Miscellaneous warnings:
----------------------------------------------------------------------------

-- The exact meaning of the all-uppercase expression 'NOT REQUIRED' is not defined in RFC 2119. If it is intended as a requirements expression, it should be rewritten using one of the combinations defined in RFC 2119;
     otherwise it should not be all-uppercase.


  Checking references for intended status: Informational
----------------------------------------------------------------------------

  == Unused Reference: 'OATH' is defined on line 791, but no explicit
     reference was found in the text

  == Unused Reference: 'XMLSIG' is defined on line 798, but no explicit
     reference was found in the text







On Mar 18, 2008, at 2:49 PM, Ben Campbell wrote:

--General:

I'd like to see some text clarifying the relationship of this draft to the Open Authentication initiative. The draft states that the work has been endorsed by that group. Is this draft merely intended to document work done by that group? (I note that the XML name space is scoped to http://www.openauthentication.org/. ) Or is intended to specify an interoperable extension to the IODEF format? If the latter, how much consideration has been given to whether this should properly be an informational vs a standards-track RFC?



--Detailed Comments:

Abstract:

Please expand IODEF on first use.

Section 1, paragraph 4:

Please expand IODEF and XML on first use.

Paragraph 5:

"Verification procedures and the specific requirements for authorization are outside the scope of this specification."

Are these requirements specified elsewhere? They seem pretty fundamental for this mechanism to be useful.

Section 4, paragraph 3:

"The primary difference in the "inbound" and "outbound" reports is the removal in the "outbound" reports of reporting organization information in order to protect confidentiality. We elaborate on this aspect in section 7, Security Considerations."

It's a little unclear to me at this point what the outbound report contains and what it is used for. Maybe a discussion of what network element comsumes "inbound" reports and what generates "outbound reports" would help.


Section 4, 2nd paragraph on Page 7

(nit) There is an unfortunate line break in EventData.AdditionalData that has the effect of rendering the paragraph incoherent. It looks like AdditionalData starts a new sentence.

Section 5.4.1:

Has there been thought whether OtherEventType needs to be registered somewhere?

Section 6.2:

The section is titled "Optional Contents" but goes on to say these contents SHOULD be included. That's really stronger than "optional". Perhaps the section should be retitled something to the effect of "Recommended Contents".

paragraph 6:

"The IPv4 or IPv6 address or subnet mask..."

Address of what?

Section 6.3, first paragraph:

I'm having trouble parsing the last paragraph. I suspect a missing comma, along with the inconsistent verb tense ("performing", "views", "has changed") contribute to the problem.

Method.URL:

"A URL that represents the detailed definition of the fraud event signature."

Does represent mean that the URL points to the definition, or somehow encodes it? Is there a definition of the format of the data this will point to? What URL schemes can go here?

Section 8:

I think the security consideration section should talk a little more on what attacks are known possible, and how the required security features address them. That is, _why_ we need signatures, etc. Don't just assume it to be obvious.


Section 8.3, first paragraph:

"A simple mechanism MUST enable the query of any data to return a valid reponse without disclosing the unique Identifier of a specific organization. "

Is the requirement that such a mechanism must simply exist, or that the implementation must actually use it?

paragraphs 2 through end of section

"We suggest to use" should probably be recast as normative language. Also, I'd like to see a little more about how OpaqueIdentifier is to be used--I gather that the publisher needs to be able to correlate the OpaqueIdentifier with the original IncidentID Field in the future? If so, we should mention that.

Section 8.4:

I think we need a little more here--if you use a secure transport, do you still need application level crypto? Can you say a little more about what security properties are needed? Also, do you mean SSL or TLS?


_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf

<Prev in Thread] Current Thread [Next in Thread>
  • Followup on draft-mraihi-inch-thraud-06.txt (was re: Gen-ART Last Call Review of draft-mraihi-inch-thraud-05), Ben Campbell <=