phc-discussions - Re: [PHC] more woes, last woes

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]

Message-ID: <20140415191138.GA8463@bolet.org>
Date: Tue, 15 Apr 2014 21:11:38 +0200
From: Thomas Pornin <pornin@...et.org>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] more woes, last woes

On Tue, Apr 15, 2014 at 01:58:44PM -0500, Steve Thomas wrote:
> Also it loses a lot of speed if you don't have AES-NI.

Actually that one can be listed as a weakness, too. If a function uses
some operation which relies on a specific hardware support (e.g.
AES-NI), then you surely hope that the defender has access to it,
because the attacker surely will. Using an AES-enhanced password hashing
scheme when your actual hardware cannot run it efficiently equates to
giving away a huge advantage to the attacker.

It is a delicate balance: AES-NI looks like a good idea because an
AES-optimized circuits are quite expensive for an ASIC/FPGA attacker,
and also somewhat adverse to GPU, while the defender has AES-NI in his
server. However, this holds only as long as the defender indeed has
AES-NI in his server. This would not be true, for instance, on ARM-based
server farms (that's the new fashion; it apparently saves a lot on
energy and cooling).

(Although there have been some announcements about soon-to-be ARM
processors with AES opcodes. The situation is not completely clear.)

	--Thomas Pornin