web clustering

[Frank Heckenbach]

Jeff Davies wrote:

> Karin Kosina wrote:
> 
> > Jeff Davies [Wed, Jul 11, 2001 at 05:39:36PM +0100]
> >
> > > I also have an idea for much stronger enryption than one time pad,
> > > but I need to finish writing it up.
> >
> > This is not a crypto list, but let me state that one thing: No
> > encryption technique can possibly be stronger than OTP. If you have no
> > indications as to what the plaintext might be, OTP is unbreakable.
> >
> > M + K = C
> >
> > if M,K == unknown (and C known), and all keys K are possible, then
> > therefore all "corresponding" plain texts C-K are equally possible.
> >
> 
> "If you have no
> indications as to what the plaintext might be, OTP is unbreakable."
> Big "if".

Nope. You don't learn anything new about the plaintext except its
length. The real problem with OTP is key distribution (key size =
plaintext size), and it must be used correctly (really one-time;
real random), otherwise it becomes very weak very fast.

> My technique is: random table for character combinations perhaps
> words even.
> But then the Key maps message bits into a much larger random packet (perhaps
> 100 times the size).
> (the bits are also randomly inverted, and there are randomly assigned
> verification integers to avoid corruption
> in the packet - checksums weaken the technique of course). Also, the packet
> can then be buried in
> multiple dummy packets.
> The spraying of the bits randomly into a larger random packet is a bit like
> spread spectrum radio
> transmission.

Sounds a bit like http://theory.lcs.mit.edu/~rivest/chaffing.txt
But serious discussions of such issues really don't belong here.

> As you probably know, Echelon can decode 128 bit public private keys at the
> rate of 5000 per second
> per ASIC, and USA probably have many, many ASICs per board in a huge set of
> racks somewhere.

But 128 bit is an appropriate size for a symmetric cipher and
ridicolously small for an asymmetric one (public key). Breaking a
1024 bit PK is probably much harder than breaking 8 128 bit keys.

Frank

-- 
Frank Heckenbach, frank at g-n-u.de
http://fjf.gnu.de/
PGP and GPG keys: http://fjf.gnu.de/plan