Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 18 Nov 2012 03:23:31 -0800
From: Colin Percival <cperciva@...snap.com>
To: Solar Designer <solar@...nwall.com>
CC: scrypt@...snap.com, crypt-dev@...ts.openwall.com
Subject: Re: scrypt time-memory tradeoff

On 11/17/12 02:53, Solar Designer wrote:
> On Sat, Nov 17, 2012 at 05:20:54AM +0400, Solar Designer wrote:
>> Thus, we have halved our memory needs (circuit area) and paid for this
>> by only a 25% increase in processing time.
>>
>> Now, if we want to reduce our memory needs a lot more, the situation is
>> a lot better (for the defender).
> 
> Actually, unless I am mistaken, the area-time product is asymptotically
> (by making extreme use of this trade-off) only 25% of what's expected in
> the scrypt paper.

You get a 2x cost reduction by trading increased time for reduced area (as
in a previous email) and another 2x reduction by ignoring the initial setup
(practically speaking) but I never intended to include the setup in my
area-time bound.

-- 
Colin Percival
Security Officer Emeritus, FreeBSD | The power to serve
Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ