phc-discussions - Re: [PHC] How important is salting really?

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: <734926844.290068.1418455622417.JavaMail.open-xchange@oxuslxltgw11.lxa.perfora.net>
Date: Sat, 13 Dec 2014 01:27:02 -0600 (CST)
From: Steve Thomas <steve@...tu.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] How important is salting really?

> On December 12, 2014 at 11:19 PM Ben Harris <ben@...rr.is> wrote:
> 
>  On 13 December 2014 at 11:00, Steve Thomas <steve@...tu.com
> <mailto:steve@...tu.com> > wrote:
>    > > 
> >    P.S. MD5("deliciously-salty-" || pw) is from an infamous article that
> > gets salt
> >    wrong :). It's near the top of Google when searching for rainbow tables.
> >  >   
>  Joking aside, if you have a unique "deliciously-salty-", truncate the MD5 to
> 24 bits, and implement a good rate limiting system you'd probably have a
> pretty secure system. Good enough against online attacks, but not so good that
> you are giving up someone's password if you leak hashes. [weak passwords are
> still weak, and leaks from multiple sources for the same UID would eventually
> give up a password]
>

Huh? Maybe I don't get your sense of humor. You do mean this as a joke, right?
Because with IPv6 it's common to get a /48 which gives you 2^16 subnet prefixes
(so 2^80 IP addresses if the limit is not per subnet prefix). Just using the
subnet prefixes you only need to try 128 passwords per subnet prefix before you
are expected to get a collision. Obviously you can buy more servers and/or
subnet prefixes. You use to be able to buy a /32 for no reason at all (this
would be 1/512 passwords per subnet prefix for a collision).

I'm not saying rate limiting is dead because of IPv6. It just severely messes
things up.