bugtraq - Re: Thoughts and a possible solution on homograph attacks

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <20050316100202.GY28919@egrid-14.egrid.it>
Date: Wed, 16 Mar 2005 11:02:03 +0100
From: Riccardo Murri <riccardo.murri@...p.it>
To: khockenb <khockenb@...vens.edu>
Cc: bugtraq@...urityfocus.com
Subject: Re: Thoughts and a possible solution on homograph attacks


[khockenb, Tue, Mar 15, 2005 at 07:10:16PM -0500]
> On Tue, 15 Mar 2005, Riccardo Murri wrote:
> 
> > I would rather suggest that the string comparison function used in IDN
> > takes "homograph caracters"[1] into account: just like the current DNS
> > considers 'a' == 'A', the IDN DNS should consider "LATIN SMALL LETTER
> > a" == "CYRILLIC SMALL LETTER a" == "CYRILLIC CAPITAL LETTER A" ==
> > "GREEK CAPITAL LETTER A"[2], and similarly for the other homograph chars.
> 
> But that breaks case insensitivity for Greek, for instance (other
> languages, too, I am sure).
> 
> Consider Greek letters eta and nu.
> 
> A upper case eta looks like an upper case Latin "H", but a lower
> case eta looks like a lower case Latin "n".
> 
> Similarly, an uppercase nu looks like a upper case Latin "N", but a lower
> case nu looks like a lower case Latin "v".
> 
> If such a system as you suggest is in place, and someone in Greece wants
> the site (Greek nu).gr, they would have to have control of both N.gr and
> v.gr, otherwise people who type in the wrong case would go to the wrong
> site.  Now let's say a competitor comes along, and wants (Greek eta).gr.
> They can get H.gr, but n.gr is already take, since N=n.
> 
> I suppose we could get around that by making H=n=N=v(=V=H), but that would
> get cohfusivg.
> 

You're perfectly right - this equivalence relation would backfire on
ASCII-only domains too. 

Riccardo

-- 
Riccardo Murri
EGRID Project
The Abdus Salam ICTP

Strada Costiera, 11
34016 Trieste
Italy

email: riccardo.murri@...p.it
phone: +39 040-2240-542
fax:   +39 040-224531