Re: Collision resistant encryption scheme
- From: Paul Rubin <http://phr.cx@xxxxxxxxxxxxxx>
- Date: 06 Feb 2006 15:20:13 -0800
Mike Amling <nospam@xxxxxxxxxx> writes:
Actually all shift cyphers (Caesar cyphers) have this feature. Ie,
C(R,M)=M+R mod(2^blocklength) have the property that for every R
(the key) C(R,M)!= C(R',M)
Same for XOR, C(R,M)=M^R, a standard implementation of a One-Time Pad.
Neither of these schemes is IND-CPA secure. There are trivial known
plaintext attacks against both.
Real-world ciphers tend not to be designed against related key
attacks. They also sometimes have properties like the DES key
complementation property. Can the application be broken only be a
real, full collision, or is it bad enough if some relationship between
keys can be pushed through to a relationship between ciphertexts?
Anyway, the usual remedy is to use hash function outputs as keys.
.
- Follow-Ups:
- Re: Collision resistant encryption scheme
- From: Sergei
- Re: Collision resistant encryption scheme
- References:
- Collision resistant encryption scheme
- From: silentser@xxxxxxxxx
- Re: Collision resistant encryption scheme
- From: Unruh
- Re: Collision resistant encryption scheme
- From: Paul Rubin
- Re: Collision resistant encryption scheme
- From: Unruh
- Re: Collision resistant encryption scheme
- From: Mike Amling
- Collision resistant encryption scheme
- Prev by Date: Re: Collision resistant encryption scheme
- Next by Date: Re: Collision resistant encryption scheme
- Previous by thread: Re: Collision resistant encryption scheme
- Next by thread: Re: Collision resistant encryption scheme
- Index(es):