ietf-openpgp
[Top] [All Lists]

Re: New results against SHA-1

2009-05-05 14:16:16

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


On May 4, 2009, at 7:46 PM, Peter Gutmann wrote:


Daniel Kahn Gillmor <dkg(_at_)fifthhorseman(_dot_)net> writes:

What do other folks think?

Given that the MDC is a hash of plaintext that's then encrypted, and  
the hash
value is itself encrypted, I'm not losing any sleep over it.  The  
hash attacks
so far have required bit-for-bit carefully-chosen plaintext with  
known hash
values, not unknown (or even partially-known) plaintext with an  
unknown hash
value.

I'm not losing a lot of sleep over it, either.

The point of the MDC is to provide a low-level integrity check.  
There's an easy high-level integrity check, a digital signature. The  
MDC exists for people who don't want to sign, but do want more  
protection than naked CFB mode, which is completely vulnerable to  
truncation.

The construction we use is not "secure". I put scare quotes around it  
for a reason. In particular, it's vulnerable to existential forgeries.  
However, every spam in the world is an existential forgery, and if you  
wanted to send an MDC forgery to someone, it's much easier to just  
write the message and encrypt it to them than modifying an existing  
message. What that means is that while there are some protocols that  
really have to worry about existential forgeries (like IPsec), we're  
really not one of them, especially since there's always signing for us.

In 4880, we described how one might upgrade the MDC. If someone  
believes it's important, I would support anyone writing a draft for an  
upgraded MDC. (But as an implementer, I can't make a statement as to  
when or if PGP would implement it.)

        Jon



-----BEGIN PGP SIGNATURE-----
Version: PGP Universal 2.6.3
Charset: US-ASCII

wj8DBQFKAHrqsTedWZOD3gYRAo0BAJ4maMvMTEHDIiJBQ+ry3VuUt3gW7gCglCkE
0nX3EUzYQ+alsPjef8RSeE4=
=Tq6M
-----END PGP SIGNATURE-----

<Prev in Thread] Current Thread [Next in Thread>