[3433] in bugtraq
Re: InterNIC Shenanigans (crypt-pw)
daemon@ATHENA.MIT.EDU (Igor Chudov @ home)
Sat Oct 12 03:56:48 1996
Date: Fri, 11 Oct 1996 19:47:22 -0500
Reply-To: Igor Chudov <ichudov@algebra.com>
From: "Igor Chudov @ home" <ichudov@algebra.com>
X-To: hamors@litterbox.org
To: Multiple recipients of list BUGTRAQ <BUGTRAQ@netspace.org>
In-Reply-To: <Pine.LNX.3.95.961011120728.3070A-100000@socks.litter717.net>
from "Sean B. Hamor" at Oct 11, 96 12:21:50 pm
-----BEGIN KGB SIGNED MESSAGE-----
PGP auth scheme also seems vulnerable to replay attacks.
For example, suppose alice@victim.com sends a signed message to
hostmaster@internic.net asking him/it to set the nameserver for
victim.com to box1. Mallory intercepts this message and stores it for
future attacks.
A year later, Alice decides to move and change a provider, and sends a
signed message asking to change the domain nameserver to box2. Mallory
also intercepts it and finds out that alice is moving.
A month later when Alice thinks that she has changed her nameserver
successfully, Mallory strikes and re-sends the first message. At this
time, Alice's domain is effectively disabled because it is served by a
wrong nameserver.
At best, the exchange should be done using cookie protocol when
Alice has to request a non-reusable cookie from InterNIC
prior to sending any domain change requests.
At worst, internic and its guardian should allow for encrypted
(and signed) messages. That will at least thwart some of the attackers.
igor
Sean B. Hamor wrote:
>
> -----BEGIN PGP SIGNED MESSAGE-----
>
>
> Well, the InterNIC has started protecting against fakemailed domain name and
> NIC handle changes by adding "crypt-pw" and PGP support to their databases.
> For those of you not familiar with this, you can now request that your email
> address not be used to authenticate you, but instead add a crypted string or
> PGP key id to your domain/NIC template to authenticate you. You have to
> submit your PGP public key block to the InterNIC keyserver if you've chosen
> to use the PGP option.
>
> Regardless, it seems that there may be a hiccup in the InterNIC's method of
> generating crypted strings. I admit, I'm not very knowledgeable when it
> comes to encryption schemes, but even I can see an initial problem here.
> Because of my lack of knowledge, however, I wouldn't be able to continue any
> further to see how deep this initial discovery I made runs.
>
> If you don't want to use the entire WWW domain/NIC template, you can use
> http://rs.internic.net/guardian/crypt-pw.html to generate a crypted passwd
> for you. Basically, you type in your cleartext passwd, hit submit, and it
> fires back the crypted version at you. Here are some cleartext/ciphertext
> combinations:
>
> nuke nuX9097V9o/TY
> narque naXwgSS98Q3xo
> cq cqjtFeb2JgXwg
> 222222 22Yrs645sLqh2
>
> Is it just me, or does it seem silly to you that the first two characters of
> the passwd are revealed by the first two characters in the crypted passwd?
> A quote from the crypt-pw.html page: "Please note that this option is not
> as secure as PGP. We recommend the use of PGP when possible." Go figure.
>
> Just my 00000010 sense...
>
> Finger hamors@ishiboo.com /\_/\ mailto:hamors@litterbox.org
> for PGP public key block. ( o.o ) http://www.ishiboo.com/~hamors/
> alt.litterbox, The Home of TOCA > ^ < http://www.litterbox.org/~hamors/
> Hi! I'm a .signature virus! Add me to your .signature and join in the fun!
>
> -----BEGIN PGP SIGNATURE-----
> Version: 2.6.2
>
> iQEVAwUBMl50JjU6HlxZIJ+FAQGeGgf+NOuQRbTdWz1nxovYsZ324Ij/fxAzQ643
> dVwf8yc1HNfNTw0zZlpz47EVaNQ5w6yrISXHmmSQ6UT+E2tnCOnK6dpaTamNO941
> HiADrWMQz+OvHrNM/z4BZPPQlJrWZJ3Jbak88S1fboDoNKqb4tLhS//3I7oFURiB
> 2Dnxy0W8oge4rJhoP+XEIsW+CdyFZrYxy2TpwEGfYxybm7I890TA5u43XEEA+QLQ
> Mm7AQXcNlcaYpH33Pavr964c1q68aRWvCXgnH4f9aSkRzjvYancVDpBIRZAbZM2Y
> 4XAsJ6yLhoHmrP6PaZQ7Xj7ChmIEfE8P0FNWdVqGAypRG8+/tymc+Q==
> =jyAN
> -----END PGP SIGNATURE-----
>
- Igor.
