QUOTE(Mudmaniac @ Jun 16 2006, 01:42 AM)
I dunno. I have a cousin whose really into EC and RSA type public key cryptosystems who always tells me that having enough plaintext/cyphertext pair will be the undoing of any cypher by a cracker. My own knowledge in this is 6 years old and I dont have detailed knowledge of the workings of the DS authentication system.
I hope you know tho, that virtually impossible means "very long time" and that kinda time function decreases with each new generation of computer hardware.
I know zit about how DS authentication works either. But I'm pretty sure it's easier to alter codes that does the authentication than to do the crazy maths that's involved to break a cipher.
A 1024-bit keysize will take 300,000,000,000 MIPS-year to crack. That is, it takes a CPU running at 1,000,000 million instructions per year 300,000,000,000 years to crack. To put things in perspective, the current lean-and-mean 1billion transistors Itanium would probably have a best estimate of 125,000 million instructions per second. You'll need a super-computer to get any result at all. That is a very long time to deduce a private key.
But the hash function (MD5/SHA-1) required for text signing does seems easier to break. In fact, I've heard of one successful attempt. Using collision detection, one could spoof your signature. Still this is pretty hard, and it gets harder as the hash digest size increases. There's 4 billion chance you could get a collision with a 32-bit hash digest and SHA-1 uses 160-bit hash!
And the beauty of this is that as time passes and new hardware surfaces, we'll just be seeing bigger crypto key size and bigger hash digest that would exponentially increase the time that a cracker would take to compromise them.
You can get more info on public key crypto attacks
here. PGP is a popular public key encryption program.
This post has been edited by ray_: Jun 16 2006, 09:48 AM