phc-discussions - Re: [PHC] "Predictable" vs "pseudorandom" KDFs

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 for Android: free password hash cracker in your pocket

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

Date: Fri, 28 Feb 2014 00:11:39 +0400
From: Solar Designer <solar@...nwall.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] "Predictable" vs "pseudorandom" KDFs

On Thu, Feb 27, 2014 at 12:30:50PM -0700, Taylor Hornby wrote:
> On 02/27/2014 06:45 AM, Bill Cox wrote:
> > I need a short name for cache-timing-attack resistant KDFs which do no
> > password dependent memory addressing.  I use the phrase 20 times in my
> > paper, and "KDFs which do no password dependent memory addressing" is
> > a mouthful.  I also use "KDFs which do password dependent addressing
> > from the start" about 20 times, and I discuss my "hybrid" version,
> > which like Scrypt, does the no-password-addressing thing in the first
> > loop, and does the password-dependent thing in the second.
> > 
> > Could we call the first type "predictable" KDFs, since they do only
> > predictable address lookups?  The second type could be "pseudorandom"
> > KDFs.  Are these good names?  I find it is hard for a group to discuss
> > ideas that have no concise name.
> 
> I feel like "predictable" without any other qualifications would
> conflict too much with other uses

Definitely.

> How about "blinded" and "unblinded" (or "nonblinded")?

I dislike these because they emphasize way of achieving something rather
than that end result itself, and in fact blinding might not always be
the way (and in Bill's example is not).

How about "cache timing safe" and "cache timing unsafe"?

Alexander