full-disclosure - In case y'all didn't catch it yet...

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20050216161044.GB12598@sentinelchicken.org>
From: tim-security at sentinelchicken.org (Tim)
Subject: In case y'all didn't catch it yet...

Hello,

> A couple of things to note from mr schneider's blog warning...

It's "schneier", not "schneider".

> Fact..until it's published and the method handed out and it's replicated 
> independently, SHA-1 is NOT broken.

I feel that statement is misleading.  SHA-1 may be broken.  It may have
been broken years ago, for all we know, by some well-funded government.
Just because the methods aren't available, that doesn't mean it isn't
broken.

Does this mean we can all trust these researchers to be telling the
truth?  No.  No matter how trustworthy we think these guys are, the
public can't be 100% sure that SHA-1 collisions can be found, until the
methods are public and tested.  That I agree with, and I think that's
probably what you meant.

> Fact..the people posting here missed a fairly important bit of schneider's 
> blog-post, and I quote: 
> 
> >>The paper isn't generally available yet. At this point I can't tell if 
> the attack is real.
> 
> read it again:  >>AT THIS POINT I CAN'T TELL IF THE ATTACK IS REAL.

Yes, exactly.  In a way, it is all FUD until the methods are published.
Hard to ignore FUD, though.

> Fact:  If the paper and method are sound...the sky STILL is not falling 
> (although it will be raining pretty darned hard)...2^69 operations...to 
> get a collision...how many hours of current gen cpu cycles?? (some notes 
> from the blog thread postulate a 4 ghz machine would need 4000 
> years...4000 CPU@ 4 ghz ea ...1 year.)  Based on the rumors, would I 
> entrust state secrets to SHA-1? Nope.  Digital signature on a dosument 
> authorizing the use of Nukes against Utah?  Nope.  Failing that level of 
> protection requirement, for the time being, of course I would, and will.

True, if the amount of computation required is 2**69, then it is still
an attack out of reach for most applications, since it is often cheaper
just to bribe people, or to hold them at gunpoint.

However, if a specific collision in SHA-1 is released, as it was with
MD5, then there are some special case attacks that can be done, without
needing all that computational power. See:
  http://www.doxpara.com/md5_someday.pdf

In addition, if these methods are eventually made public, peer review
may yield extensions which are more efficient.  It is at least something
to keep an eye on.

tim