On Thu, May 1, 2014 at 7:17 AM, Daniel Kahn Gillmor
<[email protected]> wrote:

>> Hopefully, when
>> somebody publishes a research paper showing an improved attack on
>> SHA-256 — say, preimage at a cost of 2**128 (yikes!) — we should
>> immediately begin the deprecation process in anticipation of the
>> attack improving to 2**80 soon. Conscientious site operators will
>> observe the new spec, and update their headers, and will be ready for
>> the time a year later when we say "...UAs MUST NOT honor SHA-256...".
>
> But in the meantime, clients which have already implemented SHA-3 (or
> whatever) should be able to be protected against breaks in SHA-256, no?

2**128 is so expensive, and would be such an incredible improvement
over the current cost of second-preimage attack (2**256) that frankly
I'm willing to take the risk to keep the I-D from getting any more
complicated. Second-preimage attack against SHA-256 is so much not our
biggest problem right now — many other things are much more likely to
go wrong.

> It takes us a long long while to get around to deprecating things that
> we know are broken.  If both peers in a negotiation *can* handle
> stronger mechanisms, they shouldn't have to wait for an official
> deprecation of the weaker mechanism in order to get the benefit of the
> stronger mechanism.

First, we have to get any peers at all using HPKP. And we've been
sluggish long enough (blame me first and foremost, of course).

> I agree that we're talking about pretty intense hypotheticals here, but
> i see no reason that we shouldn't specify the right thing in this draft
> while we have the chance to provide this guidance.

I'm a Worse Is Better believer, rather than a Right Thingist;
especially when as far as anyone publicly knows, SHA-256 is very
strong against second-preimage. We already moved to get rid of SHA-1
as a concession to the Right Thing (and, code simplification).

> PS orthographic nitpick: in -12, you've got "interpretting" where i
> think you mean "interpreting"

Fixed, thanks.

> PPS your link to [pin-break-codes] appears to be 404:
>               Perrin, T., "Self-Asserted Key Pinning", September 2011,
>               <http://trevp.net/SAKP/>.

Oh yes, I believe Trevor mentioned this too. Got rid of it.

https://code.google.com/p/key-pinning-draft/source/detail?r=9c118381b2b6e49384f2ea22e13247ea1c992fb9

_______________________________________________
websec mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/websec

Reply via email to