[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5513B138.3090401@gmail.com>
Date: Thu, 26 Mar 2015 08:11:52 +0100
From: Milan Broz <gmazyland@...il.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Another PHC candidates "mechanical" tests (ROUND2)
On 03/25/2015 11:40 PM, Bill Cox wrote:
...
> In the specific case of LUKS, I think we can expect the authors
> (including you) to do a good job picking default parameters for the
> password hashing algorithm. This is not a simple task, and we can
> expect users generally to mess up on this step, which is why we
> should give them an API with only one knob: runtime. However, for
> critical applications such ask LUKS, I think you guys can choose the
> parameters to match the machine.
LUKS was designed by Clemens to benchmark PBKDF2 on the machine where
it is formatted (in 2004) and to optimize iteration count to take ~1second
(default) there. I definitely want to do similar thing with the new kdf.
> Is it actually the case that there are LUKS encrypted hard drives
> that are decrypted without multi-threading capabilities? If this is
> the case, can't we choose to disable multi-threading when compiling
> for that platoform, while taking advantage of the improved protection
> multi-threading provides on most platforms? I do not like the idea
> of substantially weakening password security for the majority of use
> cases based on a rare corner case.
This is exactly why I posted benchmarks here for discussions (and not
blindly go to publish it somewhere and forget later :), I can be
wrong in some conclusions and the non-thread requirement is really
questionable.
(And I do not care for other applications running in memory. Unlocking
a disk is not an operation which is running often)
I would ask another way - there is already some clone of LUKS support
in Grub2 bootloader (but take is just an example).
Are the algorithms easily portable without major changes to such environment
including parallel attributes (no problem if it is run sequentially there)?
It is not major user case (most of Linux distributions uses non-encrypted
inittramdisk and unlock system device from fully running OS environment.
But I do not want to close this way for the future.
Or more generic question: what about embedded world?
PHC seems to be mainly about online services, where you have some big
Intel server hosting behind it but I would like to not forget about embedded
systems.
If not embedded, even low-cost Raspberry PI2 have 4 cores
(and Neon engine capable accelerate e.g. AES).
(Actually I should run tests there as well perhaps :-)
And yes, people are often using FDE on these platforms.
Moreover Android: it uses dmcrypt but do not utilize LUKS but has similar
key storage which uses PBKDF2 replaced by scrypt (since 4.4) AFAIK.
Thanks,
Milan
Powered by blists - more mailing lists