Re: [Cfrg] I-D Action: draft-irtf-cfrg-dragonfly-03.txt

[Watson Ladd <[e-mail]>]

On Feb 3, 2014 1:49 PM, "David McGrew" <[e-mail]> wrote:
>
> On 02/03/2014 02:47 PM, Watson Ladd wrote:
>>
>>
>> On Feb 3, 2014 11:27 AM, "Dan Harkins" <[e-mail]> wrote:
>> >
>> >
>> >   Hello,
>> >
>> >   I updated the dragonfly draft to incorporate the comments received
>> > from Rene and Scott. Please take a look.
>> >
>>
>> It still doesn't compare favorably to JPAKE or SPAKE2. TLS has shown
less then zero interest in it. No reduction or evidence for claims made is
forthcoming. The draft excludes curves with uniform hashing to points.
>>
>> Why is this specific PAKE a work item and not the other alternatives?
>
>
> You mean like draft-irtf-cfrg-augpake-00?
>
> Drafts are taken up when someone is willing to write one, and there are
sufficiently many other people that are interested.
>
>> Was this a joke for groundhog day?
>
>
> The sarcasm is not helpful.  Let's stick to a technical discussion.
>
> This draft most certainly should be reviewed, since security concerns
were raised regarding earlier versions of the draft, especially regarding
implementation guidance and timing channels.
>
> The process by which CFRG drafts can become RFCs is described in
http://wiki.tools.ietf.org/html/rfc5743#section-2.1   Note that there is a
paragraph in the RFC that describes the relationship of that work to the
research group.    This mechanism enables the sentiment of the RG to be
captured in the RFC.
>
> Let me ask: can you suggest text for the security considerations section
of this draft that captures your concerns regarding the lack of reduction
and uniform hashing?

Sure: "Despite significant efforts, no variant of this protocol has been
proved secure even in the random oracle model with nonstandard assumptions.
None of the security claims are sensible in any accepted formalization of
security protocols. Significant dissent in the WG existed with publication
due to the lack of any more than superficial analysis. No Internet protocol
should use this PAKE when alternatives exist." Just about sums it up.
>
> David